From the collection of the

Prelinger v JLJibrary

t

San Francisco, California 2007

JOURNAL

OF THE SOCIETY OF

MOTION PICTURE ENGINEERS

SYLVAN HARRIS, EDITOR Volume XIX JULY, 1932 Number 1

CONTENTS

Page

Standardization of Projection Lamps E. W. BEGGS 817

Biplane Filament Construction — A High Intensity Incandescent

Lamp Light Source for Motion Picture Projection . . J. T. MILI 829

Illumination in Projection Printing of Motion Pictures

C. TUTTLE AND D. A. YOUNG 842

A New Light Control for Printing Machines K. SCHNEIDER 865

Sound Motion Picture Equipment for the U. S. Navy

S. W. COCHRAN 872

The Duplication of Motion Picture Negatives ,

J. I. CRABTREE AND C. H. SCHWINGEL 891

The Screen — A Projectionist's Problem F. M. FALGE 909

Photographic Emulsions L. W. PHYSIOC 913

Book Reviews 921

Officers 923

Society Announcements 924

JOURNAL

OF THE SOCIETY OF

MOTION PICTURE ENGINEERS

SYLVAN HARRIS, EDITOR

Board of Editors

J. I. CRABTREE, Chairman

L. DE FOREST A. C. HARDY F. F. RENWICK

O. M. GLUNT E. LEHMANN P. E. SABINE

Published monthly at Easton, Pa., by the Society of Motion Picture Engineers.

Publication Office, 20th & Northampton Sts., Easton, Pa. General and Editorial Office, 33 West 42nd St., New York, N. Y.

Copyrighted, 1932, by the Society of Motion Picture Engineers, Inc.

Subscription to non-members, $12.00 per annum; to members, $9.00 per annum, included in their annual membership dues; single copies, $1.50. A discount on subscriptions or single copies of 15 per cent is allowed to accredited agencies. Order from the Society of Motion Picture Engineers, Inc., 20th and Northampton Sts., Easton. Pa., or 33 W. 42nd St., New York, N. Y.

Papers appearing in this Journal may be reprinted, abstracted, or abridged provided credit is given to the Journal of the Society of Motion Picture Engineers and to the author, or authors, of the papers in question.

The Society is not responsible for statements made by authors.

Entered as second class matter January 15, 1930, at the Post Office at Easton, Pa., under the Act of March 3, 1879.

STANDARDIZATION OF PROJECTION LAMPS* E. W. BEGGS**

Summary. — One hundred and thirty types of picture projection lamps are today required to fill the demands placed on the lamp manufacturers. This large number of types, each being available in several different voltages and in most cases with either one of two kinds of bases, brings the total to over 500. The total annual demand for the entire country is only 240,000 units for lamps of this type. The result is that the lamps are expensive and that the amount of engineering that can be devoted to each type is entirely inadequate.

This situation is partly because small improvements in lamps have been made almost annually in the past, and partly because of wide diversity of opinion among projector designers. By assembling and coordinating the ideas of equipment de- signers, an ideal set of light source requirements can be laid down. Based on these requirements the lamp manufacturers can then establish the characteristics of light sources to be introduced during the next few years.

Early in 1925, Mr. Hoover, then United States Secretary of Com- merce, urged standardization and simplification of articles manufac- tured in this country. The advantages of such a program as was then begun were that the articles could be made in larger quantities and consequently at lower cost, that reduction in the number of types to be supervised would make possible higher quality, and that with greater interchangeability between specific articles of various brands, availability of any product would be increased. These advantages naturally applied to Mazda lamps, which are highly specialized manu- factured articles of broad distribution. All of the factors involved in the program are of great importance to the far-flung but powerless ultimate consumers, and lie within the control of organized equipment manufacturers. Standardization of projection lamps, however, has not been widely discussed before this Society and now the need for it is great.

In 1931 the Mazda lamp manufacturers listed in their price sched- ules 134 types of lamps specifically for projection service. Even during the first few months of 1932, quite a large number of new pro- jection lamp types have been added to the list.

* Presented at the Spring, 1932, Meeting at Washington, D. C. ** Commercial Engineering Dept., Westinghouse Lamp Co., Bloomfield, N. J.

817

818 E. W. BEGGS [J. S. M. P. E.

Many of these types are available with several different bases and for more than one voltage so that actually the lamp manufacturers were forced to be ready to supply more than 570 different projection lamps to fill the national demand. The production schedule of the past year showed that almost all these were ordered. The total number of units manufactured was only 238,000, a relatively small demand to require as many different designs as are now listed. Be- sides this, it is occasionally found expedient to use lamps designed specifically for other types of service in order to meet requirements of peculiar projection problems. The types of lighting service for which these other lamps were actually designed are as follows:

Automotive Headlight Floodlight

Airway Beacon Airport Floodlight

Theatrical Spotlight Radio Transmission

Photocell Exciter Television

Airplane Headlight Locomotive Headlight

Searchlight Traffic Signal

This complicates the situation considerably beyond even that indi- cated by the data available on projection lamps alone.

An analysis of the present requirements reveals the sources of di- versification. By separating the effect of voltage, lamp base, fila- ment structure, bulbs, wattages, etc., the importance of each factor can be discerned. From this, certain standardization procedure may appear practicable and the suitable steps toward the desired objective can then be taken.

TABLE I

Voltages Now Listed for Projection Lamps

A. Power Line Voltages,

105, *110, *115, *120, 125, 130

B. "High Voltage,"

220, 230, 240, 250, 260

C. Country Home Voltage,

*28-32

D. Resistor Control Voltages,

*28-32, *50, *52, *75, 90, 95, *100

E. Transformer Voltages,

3, 5, 6, *6.5, 7.5, 8, 9.5, 12, 12.5, 14, 14.5. *20, 30. 32, 33

F. Battery Voltages,

5-6, 6-8, 11-12, 28-32

* In common use.

July, 1932] STANDARDIZATION OF LAMPS 819

VOLTAGE

Lamps are now available for picture projection service designed to operate on any one of 35 different voltages. These are shown in Table I.

Available data show that most of the lamps of the voltages listed are applied as indicated. Obviously there is some overlap in this ap- plication.

VOLTAGE APPLICATION

Somewhat more than half the projection lamps used in motion pic- ture equipment today are operated directly from the power line. These are indicated by group A of Table I. An almost equal number of projection lamps are operated through a voltage reducing resis- tance unit, the latter being fixed by the manufacturer or controlled by the operator. These are indicated by group D of Table I. The remainder of the lamps used fall in one or more of the other four less important classifications.

VOLTAGE CONTROL

In a projector that the user is apt to carry from place to place so that he encounters varying voltages, some sort of voltage control should be provided. Also, for projectors which may on occasion (such as to combat daylight) require that the lamp brightness be stepped up to the maximum regardless of the consequent shortening of lamp life, this same control is needed. Such equipment is now commonly provided with variable resistance units arranged with a voltmeter or an ammeter so that the operator will know when the lamp is operating at the proper voltage or amperage. The lack of agreement as to the relative merits of the voltmeter vs. the ammeter has made it very difficult for the lamp designer to produce a product that will give the same performance in all apparatus.

The most common types of control are :

(1) Resistor and voltmeter.

(2) Resistor and ammeter.

(3) Variable resistor.

(4) Resistor and selector switch.

(5) Transformer with taps, and with voltmeter or ammeter.

In Fig. 1 is shown the effect of the voltage on the lamp performance. In Fig. 2 is shown the effect of the amperage on the lamp performance. From these two curves it is evident that the same accuracy in voltage

820

E. W. BEGGS

[J. S. M. p. E

setting produces almost double the uniformity in lamp performance as an equal accuracy in ampere setting. In addition to this, the use of a voltmeter makes it possible to use more than one wattage of lamp in a projector so equipped. Also, since lamp makers life test their product normally and most effectively at constant voltage, and since the published life of lamps is usually "life at volts," the voltage method is generally the proper control. For these reasons the volt- meter may be more desirable than the ammeter.

\
\										370 340 310 260 250 220 190 140 130 100 70 40 10
\
\									/
		FC							/
	\							/
	'	y			1	UMEN	>^y
		\					/
		^	\			/		S"	•^
					1	^	^
				\	>^-	^^	WATTS
			^*s	^	Xs
		gXx'	' /				^^,	-».^.
^X	^		/						"~~~~
		/
	/
/
9O 92 94 98 96 IOO 1	2 104- 106 108 1

PERCENT VOLTS

FIG. 1. Relation between voltage impressed on the lamp and its performance.

Of course, where very high amperages and low voltages are involved the ammeter may be preferable because it eliminates the effect of contact resistance. It should be borne in mind that operation of a lamp at constant amperage results in a burn-out life about half that of a lamp burned at constant voltage. The lamp run at constant amperage will maintain its light output somewhat better throughout life, whereas one run at constant voltage will fall off approximately 10 to 20 per cent. Nevertheless, the basic fact that lamp design is founded on voltage operation makes the voltmeter the preferred means of gauging lamp operating conditions.

To clarify the relative effect of voltage and amperage operation a

July, 1932]

STANDARDIZATION OF LAMPS

821

specific instance will aid. The 250 watt, 50 volt lamp is operated at 50 volts in some machines and at 5 amperes in others. Any new lamp taken at random will produce practically the same illumination in either case. As the filament wears away during life, however, it becomes necessary gradually to increase the voltage so as to force 5 amperes through the filament wire. Actually the temperature of the filament will rise slightly under this forcing unless ttte tungsten wire sags, which does not commonly happen. This results in increasingly

\										325 300
\
\
	\								/
	Vi	.FE						^	/	225
	\							/
		\			t	UMEN!	V			175 150 125 100 75 50 25 0
		V	s			^	/	-^	^
			\			/	^
				\	x^-	^	•VTTS
			^	-y	\
		^	/			^
^	^	/	s					-^	~— -^
	y	/
	X
s

99 IOO IOI 102 IO3 IO4 105

PERCENT AMPERES

FIG. 2. Relation between the current forced through the lamp and its performance.

rapid evaporation of tungsten, which hastens the ultimate end of the lamp's life.

Some projection equipment provides voltage reduction by means of a resistance unit permitting changes in this resistance by the operator at will, and at the same time provides no meter to give the operator knowledge of the actual lamp operating conditions. Such equipment is apt to subject the lamp to extreme overload without the operator's being aware of the situation, a practice that evidently should be dis- couraged.

Certain motion picture projection equipment, particularly that in which the 900 watt Mazda motion picture lamp is used, utilizes a

822 E. W. BEGGS [j. s. M. p. E.

tapped transformer to reduce the voltage, and a voltmeter is provided for lamp control.

Another combination involves the use of a small fixed resistor which is sufficient to cut the voltage down from the minimum nor- mally encountered to the rated voltage of the lamp. Other small re- sistors controlled by a selector switch are arranged to be introduced when higher voltages are encountered. In this manner a control de- vice without a meter is provided, which makes it possible for the operator to run the lamp at approximately normal voltage at all times if he knows what the line voltage is. He can also operate the lamp below or above the normal voltage if he wishes. This compromise system is quite effective, and is similar to one used in certain radio equipment. It functions properly only when the lamp consumes the amperes for which the system was designed.

All these methods of voltage control affect lamp performance and standardization. One method should be adopted, if possible, so that the lamp performance will be approximately the same in all apparatus. The use of a resistor and voltmeter is proposed.

VOLTAGES ENCOUNTERED THROUGHOUT UNITED STATES

Table II shows the distribution of voltage within the 100-130 volt range for the years 1929 to 1931, inclusive, as indicated by sales of large Mazda lamps.

TABLE II

Voltages jrom Power Lines in U. S. within 100-130 Range

Per Cent Volts 1929 1930 1931

100-105 0.1 0.1 0.1

110 6.5 5.2 3.7

115 49.5 48.9 48.0

120 39.4 41.3 43.9

125 3.4 3.5 3.3

130 1.1 1.0 1.1

Note that the 120 volt group has been growing, chiefly at the ex- pense of the 110 volt group. Reports for 1932 show a further trend toward 120 volts. It is important also to appreciate a new factor, which is the tendency of the power companies to raise the voltage approximately two volts at the power panel of the building they are supplying with current. This has been found necessary to compen- sate for service voltage drop in the wiring of buildings throughout the country, resulting from the gradual increase in the use of electric

July, 1932]

STANDARDIZATION OF LAMPS

823

apparatus during the past ten years. This increase in voltage at the power panel in part compensates for the voltage drop throughout the wiring of the building, and consequently should be borne in mind par- ticularly where the new 25 hour higher intensity lamps are applied. Where no voltage control is provided in the projector, the voltage of the lamp should probably be the same as the voltage published by the power company supplying the current (thus agreeing with the socket voltage) unless, of course, the customer, for his own reasons, wishes the lamps to be otherwise.

1913 '15 '17 '19 21 23 25

FIG. 3. Distribution of lamps by voltage within

27 29 31 33

the 100-130 volt range.

Fig. 3 shows, in somewhat more detail, the growth of the 110, 115, and 120 volt groups since 1915. As far as distribution is concerned these three voltages represent practically 100 per cent of the voltage in the 100-130 group. Very few cases exist today where other volt- ages will be encountered.

A-C. VS. D-C. AND ALTERNATING CURRENT FREQUENCIES

No discussion of voltage is complete without mention of alternat- ing vs. direct current, and of the varying frequencies of the former that may be encountered throughout the country. There is still an ap- preciable amount of direct current encountered, particularly in the centers of such cities as New York, Detroit, Chicago, and Boston. This is being gradually reduced but still can not be ignored.

Sixty-cycle alternating current represents 90 per cent of the electric power produced today in this country. Of the remaining 10 per cent

824 E. W. BEGGS [J. S. M. P. E.

there is some direct current in restricted areas of large cities, there is a great deal of 50 cycle alternating current on the Pacific Coast, and there are odd frequencies such as 25, 30, 33, and 40 cycles now to be met in the Eastern Great Lakes district and Canada. More than 93 per cent of the residential customers of electric power receive 60 cycle alternating current and the trend everywhere is definitely to- ward that form of current.

THE LAMP BASE

There are now seven bases used in the majority of picture projection lamps. They are as follows:

(1) S. C. Bayonet Candelabra.

(2) D. C. Bayonet Candelabra.

(3) Medium Prefocus.

(4) Medium Screw.

(5) Mogul Prefocus.

(6) Mogul Screw.

(7) Focusing Ring.

The most important of these in percentage of the total demand is the medium prefocus base. The next in order is the single contact bayo- net candelabra (automotive) base, which is also effectively a pre- focus device. Next in order comes the medium screw base with the special focusing ring attached and, following this, the remainder of the demand is shared by the mogul prefocus, mogul screw, medium screw, and the double contact bayonet candelabra base.

The use of prefocus lamps is definitely growing. Such devices greatly increase the satisfaction of the user and any piece of apparatus requiring the operator to focus the lamp should undoubtedly be discouraged. This is not only because the user prefers prefocused equipment from the standpoint of convenience, but also because by the elimination of the screw bases standardization will be prompted.

FILAMENT CONSTRUCTION

For projection service many, far too many, types of filament con- struction have come into use. In fact, there are probably more than twenty different filament designs used for this type of service. Eight important constructions are shown in Fig. 4. The first four, the C-13, C-13D, C-13 A, and CC-13, are most appropriate for the lamps to be operated directly from the house lighting circuit or for voltages above 30. The lower group of four constructions, the C-2, 2-C-2, C-6, and

July, 1932] STANDARDIZATION OF LAMPS 825

C-8, are mostly suitable for lamps of 20 volts or less. The C-13D filament is of the biplane construction reported by Mr. Mili in the succeeding paper in this issue of the JOURNAL.

THE LAMP BULB

During 1931, the T-10 bulb represented about 75 per cent of the demand for lamps used in 16 millimeter picture projectors. The T-20 bulb represented almost 100 per cent of the demand for 35 mil- limeter machines. The T-8, T-8l/z, and T-10 bulb lamps were used extensively in film slide equipment.

Until last year, practically all lamp filaments were mounted at or near the axial center of the bulb. Lens designers increased the light

TOP VIEW

C-13

FRONT VIEW

M* JIL -*- __JV "V TTt

FRONT VIEW

FIG. 4. Representative filament construction in use for projection service.

acceptance of their lenses, requiring larger filaments arranged in the lamp in such a way that they could be mounted closer to the condenser lens.

In 1931, this movement reached the point where centered filaments were no longer capable of fulfilling all the lens requirements, and lamps with offset filaments were created. Working in another direc- tion, there were developed lamps having biplane filament light sources mounted in the center of the T-10 bulb. Mr. Mill's research indi- cates that the new lenses have been designed to accommodate rela- tively large filaments of high wattage close to the condenser lens, re- sulting in more severe conditions and greater danger of softening of the glass of the bulb.

These changing requirements of the projectors have made it neces-

826 E. W. BEGGS [J. S. M. P. E.

sary to study more critically the bulb glass, the bulb capacities and shapes, as well as the cooling mechanism provided by the projector. The new requirements seem to throw last year's standardization ef- forts into chaos. For this reason the lamp manufacturers make a plea for caution, and urge careful consideration of each new lamp and housing design before its adoption.

LAMP WATTAGE

The biplane filament construction will ordinarily require double the wattage of the monoplane construction for any given lens system. The wattage consumed by the lamp and resistor will ordinarily be unchanged. Nevertheless, the biplane structure will greatly increase the lamp wattages in common use. Also, changes in the concentra- tion in the monoplane type of lamp has, in general, caused a step-up of wattage. This has further added to the seeming confusion with the result that, in 1932, projector lamp standardization is farther away than in 1931. Lamp manufacturers are, therefore, faced with a seri- ous situation and are compelled to make efforts to improve matters.

A PLEA FOR COOPERATION

The details given above indicate many of the various factors in- volved. Basically, the conditions being met today result from the fact that picture projection places complex requirements on the light source; that the apparatus, particularly in the 16 millimeter field, is relatively new and the requirements are changing; and that there is a general lack of appreciation of the advantage of standardization.

The complexity of the requirements demand a high degree of engineering skill in designing projectors. The changes which are oc- curring should not be stifled, but should be more uniformly coordi- nated and guided. An appreciation of the advantage of standardiza- tion, however, can be gained by analysis. If the results of standardi- zation already achieved in general lighting lamps might be possible in projection lamps, the advantages become obvious. In Mazda lamps for general illumination of the home, for example, which have many of the characteristics of projection lamps, the list prices range as low as 20 cents. This is partly because the six types of lamps listed for this service in 1931 covered a total demand of over 200 million units for the entire United States for that year. Contrast this with the 238,000 projection lamps divided among 134 types. The chart in Fig. 5 shows the reduction in lamp prices since 1914. This was largely made possible by standardization.

July, 1932]

STANDARDIZATION OF LAMPS

827

Consider the possibilities of similar cost reduction of projection lamps as it would affect the practicability of overvoltage operation. By referring to the curves in Fig. 2, the enormous rise in lumen output is readily apparent. Suppose that the lamp price could be reduced to one-third the present figure ; lamps could then be operated at 1 1 0 per cent of rated voltage without increasing the lamp operating cost. This would increase the screen illumination by about 40 per cent. At 125 per cent rated voltage, the screen brightness would be more than doubled.

Several specific recommendations can be made in order to improve standardization: Adopt 100 volt lamps operated with resistors and voltmeters, as the standard for the industry. When the cost of the

90 LJ y E

s*

**• ^

V

\

-^~,

— — ,

^

X

s

\

\

S> '° fe % CO

\

X,

~^».

\

& M g i

^**

•—-,

— — .

19 FIG.

14 '16 '18 '20 '22 '24 '26 '28 30 '3, 5. Per cent price reduction of large Mazda lamps since befor

the World War.

apparatus prohibits this refinement, the selector switch, or its equiva- lent, should be used and in other cases the lamps should be operated direct from the power line. Use prefocus bases in all equipment.

Standardization may entirely change the trend of lamp design and application, but when the changes follow standardization they come gradually and systematically. The result is to the benefit of the public and all those interested in the motion picture industry. The Mazda lamp manufacturers are committed to do their part of the work. They solicit suggestions and cooperation from the designers and builders of projection equipment. They urge that a careful survey of the lamp situation be made, and trust that in the future they will be called upon to provide lamps and lamp data to the de- signer early in his work rather than after the projector has been laid

828 E. W. BEGGS

out or built, as has been the case in the majority of instances in the past.

DISCUSSION

MR. FARNHAM: Mr. Beggs has brought out in considerable detail the large number of types of projector lamps that have been produced through the ramifica- tions of bulb sizes, voltages, filament constructions, lives, and bases. It so happens that whenever a projector has been developed and a new lamp is designed for it, the lamp seems to be forever with us, whether or not the projector continues to be made or has been discarded; the reason being that a few people continue to use the projector and so long as there is a demand for the lamp, no matter how small, it is continued on the list. So, without meaning to detract from Mr. Beggs' plea for standardization of projection lamps, I might mention that about ninety per cent of the projection lamp business is confined to about twenty distinct types.

Furthermore, the projection field, particularly the 16 millimeter part, is going through a very active stage of development. New types of projectors are being produced at frequent intervals, and such a situation makes standardization difficult. I know of developments now in progress in the lamp laboratories which in a short time may make many of our present types obsolete.

Nevertheless, standardization is necessary, and its importance should be kept in mind by every one who uses projection lamps, and by close cooperation between the lamp manufacturers and the producers of projection equipment, a far better degree of standardization can be achieved.

Mr. Beggs did not mention one important cause of confusion. Many types of projection lamps are created unnecessarily, merely because projector manufac- turer "A" asks for a lamp that will give his projector more light than "B's" pro- jector. Then manufacturer "B" immediately returns and asks for a lamp to produce a brighter screen than "AV equipment, and so on. In such cases, the the lamp is sometimes only a partial answer to the problem. Some manufacturers seem to place on the lamp manufacturer the entire burden of providing the addi- tional light, to avoid the expense of making changes in their projectors. During the past year the optical people have made marked improvements in both con- densing lenses and objectives, which make possible the obtaining of nearly twice the illumination now available from any projector on the market. Of course, these improvements will require modification of the projector, but sooner or later some manufacturer will adopt these new lenses and the rest will have to follow. Why not make use now of these new developments and obtain the opti- mum screen results, even though this would require that the projector be re- designed? Lamp and projector manufacturers would be spared the troubles at- tending a succession of small changes which, after all, would lead to this result.

MR. BEGGS: Variations in voltage, base, and design life are almost as serious as variations in filament form, bulb, and rating. Let us standardize on one volt- age for all high intensity lamps — perhaps 100 volts. For cheap lamps we may adopt the line voltage. Let us standardize exclusively on prefocused lamps. Let us use either voltmeters or ammeters, but not both. Such steps will reduce the actual number of types required through future years to only a fraction of the present number.

BIPLANE FILAMENT CONSTRUCTION— A HIGH INTENSITY

INCANDESCENT LAMP LIGHT SOURCE FOR MOTION

PICTURE PROJECTION*

J. T. MILI**

Summary. — Manufacturers of motion picture projectors have been continuously demanding increasingly brighter and more uniform light sources. The biplane staggered filament construction, consisting of two parallel rows of coils, so placed that the coils in one plane fill the spaces between coils in the other, is now being pre- sented as the lamp manufacturer's answer to this demand of the industry. A report is presented on comparative tests made with monoplane and biplane filament lamps ranging in wattage from 200 to 1000, with several types of optical systems. A description of the methods pursued in taking screen lumen readings and making bulb temperature measurements is also included.

The desirability of a Mazda high intensity incandescent lamp for motion picture projection which could be operated on either alter- nating or direct current at or near the prevailing line voltage in the United States, and having a greater average source brightness and uniformity than the types now in common use, has led to a re-investi- gation of the biplane filament construction. The relative merits of this filament construction, which consists of two parallel rows of coils so placed that the coils of one row fill the spaces between the coils of the other, as shown in Fig. 1, were outlined as early as 1918 in a paper1 presented before this Society by Mr. A. R. Dennington, of the West- inghouse Lamp Company.

Since then some attempts, both here and abroad, to develop a prac- tical design of the biplane filament have met with little success, chiefly, perhaps, because at that time only low voltage lamps were being considered, whereas the advantages of the biplane filament seem greatest with the 100 volt lamp. About a year ago this laboratory, in conjunction with optical and projector designers, undertook to carry on a series of researches with the idea of determining the most

* Presented at the Spring, 1932, Meeting at Washington, D. C. ** Commercial Engineering Laboratory, Westinghouse Lamp Co., Bloomfield, N.J.

829

830 J. T. MILI [j. s. M. P. E.

suitable filament construction for motion picture projection lamps. The biplane filament construction which is now being discussed is the outcome of these investigations.

LIGHT SOURCE CHARACTERISTICS

By using the double row of filament coils the average brightness of the light source is, as might be expected, almost doubled. This is shown graphically in Fig. 2, which indicates the extent of this increase at various wattages for 115 volt biplane and monoplane light sources suitable for direct operation from the lighting circuit, and the 50 volt

Front View Side View

FIG. 1. 400 watt 100 volt T-10 bulb biplane filament projection lamp.

monoplane, which has been extensively used in recent years where in- creased picture brightness has been demanded. The measurements were in each case made without a reflector, and in a direction normal to the filament planes.

The photographs in Fig. 3 exhibit the actual reduction in source size and increase in brightness and uniformity produced by the biplane (C-13D) construction as compared with the monoplane (C-13), with filaments of identical electrical rating. The light sources photo- graphed were 1000 watt, 115 volt sizes, and are shown with and without the use of a spherical silvered glass reflector.

July, 1932]

BIPLANE FILAMENT CONSTRUCTION

831

For a comparison of the two filament constructions as applied to projection purposes, vertical and horizontal candlepower distribution curves, shown in Figs. 4 and 5, were taken, with sources of approxi- mately equal dimensions and with a spherical mirrored reflector. Under such conditions, the amount of light in the usual acceptance angle of a condenser system, i. e., 70 to 110 degrees, is about 60 per cent greater with the biplane than with the monoplane. This is because a biplane lamp of equal source dimensions permits the use of twice the wattage of the monoplane.

SCREEN ILLUMINATION TESTS

The comparison was further extended by making tests to deter- mine the screen performance of the two types of filament construc- tion in some of the existing commercial optical systems. Table I gives relative values of initial screen illumination for lamps with G-13

RELATIVE BRIGHTNESS OF INCANDESCENT PROJECTION LAMP LIGHT SOURCES.'

-40'

A-II5 VOLT BIPLANE FILAMENT CONSTRUCTION B- 50 VOLT MONOPLANE FILAMENT CONSTRUCTION. C-II5 VOLT MONOPLANE FILAMENT CONSTRUCTION

FIG. 2.

Relative brightness of 115 volt biplane, and 50 and 115 volt monoplane, light sources.

and C-13D filament constructions of approximately the same source area, when used in four widely divergent optical systems designed for 16 millimeter projection. It may be seen from these figures that with any system an average increase of approximately 60 to 70 per cent in screen illumination may be expected from the use of the 100

832 J. T. MILI [J. S. M. P. E.

volt biplane filament over the 100 volt monoplane of the same source size, and about 45 per cent over the 50 volt monoplane. This is achieved by doubling the lamp wattage in the biplane types.

TABLE I

Relative Initial Screen Illumination Values with Biplane and Monoplane Projection Lamps in 16 Mm. Optical Systems

Lamp Data Screen Illumination Data*

Condenser

Spheric Aspheric Aspheric Aspheric

1 in. l»/4 in. 1 in. l»/4 in.

Objective

//2.0 //2.0 //1.65 //1.65

Fil. Distance from Filament to Condenser

Watts Volts Constr.** 18 mm. 23 mm. 12.5 mm. 16 mm.

Illumination

200 100 CC-13 100.0 100.0 100.0

200 50 C-13 123.5 122.0 135.0

400 100 C-13D 187.0 184.0 178.4

300 100 C-13 100.0 100.0 100.0 100.0

300 50 C-13 116.3 116.0 116.0 115.0

600 100 C-13D 172.0 162.0 173.0 158.7

* Includes the use of a spherical mirror.

** C-13 indicates the monoplane, C-13D the biplane, and CC-13 the coiled coil monoplane type of filament construction.

Similar increases in screen illumination may be obtained, as shown in Table II, by means of the biplane in two widely used 35 milli- meter film projection systems. The data given are based on equal life to burnout when operated with a mirror reflector.

TABLE II

Relative Initial Screen Illumination Values with Biplane and Monoplane Projection Lamps in 35 Mm. Optical Systems

Lamp Data Screen Illumination Data*

Condenser Prismatic Aspheric

Objective

No. 2; 5.5 in. focus No. 2; 5.5 in. focus Fil. Distance from Filament to Condenser

Watts Volts Constr.** 2 in. l»/8 in.

Illumination

500 100 C-13 100.0 100.0

500 50 C-13 117.2 112.5

1000 100 C-13D 159.0 164.6

* Includes the use of a mirror.

** C-13 designates the monoplane type of filament construction, C-13D the biplane.

July, 1932]

BIPLANE FILAMENT CONSTRUCTION

$33

All the measurements were made with the filament at the designed distance from the condenser and do not take into account screen ap- pearance, which in some cases was equally satisfactory with both filament constructions, but in others was less satisfactory with the monoplane.

1000 watt 115 volt biplane filament. (A ) with reflector (5) without reflector.

1000 watt 115 volt monoplane filament. (Q with reflector. (D) without reflector.

FIG. 3.

Ratio of light brightness.

Source area Ratio

A/C .512 1.63

B/D .547 1.84

Relative light source area and uniformity of monoplane and biplane filament lamps of the same electrical rating.

Figs. 6 and 7 illustrate the method of mounting lenses on the optical bench, and the integrator used for obtaining total screen lumen read- ings. For measurements of uniformity, use was made of a Weston Photronic cell illumination meter as shown in Fig. 8.

The increase in total screen illumination, due to the use of a spherical

834

J. T. MILI

[J. S. M. P. E.

mirror with the biplane filament, has been found to be on the average approximately 30 to 35 per cent. This increase is due partly to the reflected light which niters through the coils between filament turns, partly to the light which passes back between the filament coils, especially at horizontal angles from 35 to 55 degrees on either side of the optical axis, and finally to the rise in filament temperature caused by the mirror image, which in this case is superimposed on the filament itself.

90* 60*

/ r

RELATIVE

VeRTICAuCAMOUCPOWER DISTRIBUTION Biplane and Monoplane Light" Sources of ELqual Area wiH» Silver Mirrc Spherical Reflector

50°

FIG. 4. Candlepower distribution in the vertical plane passing through the optical

INFLUENCE ON OPTICAL DESIGN

It has often been stated that maximum efficiency in a motion pic- ture projector would be obtained by focusing the image of a solid light source at the film gate. However, since the available light sources are not absolutely uniform, projectors must be designed with the filament focused in a plane beyond, rather than at the film aperture, in order to obtain an evenly illuminated screen. With the biplane construction the light source is quite uniform in brightness, so it be- comes possible to focus the filament in a plane nearer the aperture than is now practical with the monoplane, thereby gaining in ef- ficiency without entailing a loss in screen quality. This brightness

July, 1932]

BIPLANE FILAMENT CONSTRUCTION

835

uniformity is shown photographically in Fig. 3. Full utilization of this factor will come only with lens systems designed specifically for the new filament construction.

Furthermore, even with a well-designed projection system, owing to the existence of aberrations in the condenser, it is quite possible to get color fringes or banded appearances on the screen. These are "out-of -focus" images of the filament, and, if produced, are much less troublesome with the biplane since the periodic variation in in- tensity across the source is less with it than with the monoplane.

90* 60*

RELATIVE yf

HORIZON TAU CANDLEPOWCR DISTRIBUTION

Biplane onj Monoplane Light" Sources ^ or Equal Area wiH\ 5ilv«r flirroreJ Spherical Reflector

^^-^— Biplane Filament — — — — Monoplane Filament"

\/ X

30*

FIG. 5. Catidlepower distribution in the horizontal plane.

UNIFORMITY OF PERFORMANCE

In order to render as simple as possible the operation of placing new lamps in the projectors, most manufacturers use prefocus base lamps. This, of course, assumes that the tolerances allowed in the making and setting of prefocus sockets and bases will not appreciably change the performance of the projector with each lamp renewal. The biplane filament reduces the need for extremely accurate adjust- ment of lamp and mirror, as demonstrated by a test run to determine the actual extent of this effect.

In the test, a group of nine 250 watt, 50 volt projection lamps with

836 J. T. MILI [j. s. M. P. E.

the monoplane filament, and nine 400 watt, 100 volt projection lamps with the biplane filament, both types with medium prefocus bases, were picked at random from the factory product. A lamp of each type was set in a representative commercial 16 millimeter projector designed for prefocus base lamps, and properly adjusted for the best screen illumination. Then the rest of the lamps in each group were placed successively in the projector without changing the socket or mirror setting, and relative readings of the screen illumination were obtained. These results are embodied in Table III.

TABLE III

Effect of Adjustment on Lamp Performance

Relative Variation in Screen Illumination as Lamps with Medium Prefocus Bases

Are Interchanged in Commercial 16 Mm. Projector. Test Run with Representative

Lot of Commercial Lamps

Per Cent Screen Lumens

250 Watt,

50 Volt

Monoplane

100.0

91.0 85.6 91.0 82.0 84.4 93.3 96.7 93.7 18.0 9.2

* Lamp No. 1 adjusted for best possible focus.

The monoplane lamps show a maximum drop in screen lumens of 18 per cent, for Lamp No. 5, and an average drop of 9.2 per cent. With the biplane lamps, the maximum screen lumen drop is only 7.8 per cent (Lamp No. 6), while the average drop for the whole group is 5.0 per cent.

With the biplane the mirror increase is only slightly more than half that obtained with the monoplane. Hence less loss in illumination on account of faulty mirror setting or other similar causes would be expected. A series of readings taken with the mirror at the proper focal position and then at several other settings behind and in front of this point confirms this. The drop in illumination was but half

Lamp No.	400 Watt, 100 Volt Biplane
1*	100.0
2	95.2
3	95.2
4	94.8
5	92.7
6	92.2
7	92.4
8	99.3
9	93.7
Maximum Deviation	7.8
Average Deviation	5.0

July, 1932]

BIPLANE FILAMENT CONSTRUCTION

837

as great with the biplane as it was with the monoplane, and, of course, there was less change in uniformity.

THE COOLING PROBLEM

The use of biplane staggered filament lamps for 16 millimeter pro- jectors, on account of the relatively large wattages involved and the

FIG. 6. Optical bench set-up.

small distances between the filament and the first condenser surface, for which optical systems are now being designed, introduces an acute lamp cooling problem. For instance, some systems have recently

FIG. 7. Integrating photometer for total screen lumen measurements.

been designed for a separation of only 12.5 millimeters between the condenser and the filament. This would require placing the filament

838

J. T. MILI

[J. S. M. P. E.

in the center of a bulb with a diameter of less than 1 inch or else in an offset position in a larger bulb. Under such conditions an ex- tremely active draught of air will be required in order to prevent the bulb from bulging. The lamp manufacturers are willing to co- operate with designers in this matter, although the determination of the amount of ventilation needed for cooling lamps in the various types of projectors and the method of applying it are naturally prob- lems for the equipment manufacturer. Temperature tests indicate that no bulging or devitrification of the hard glass now used for pro- jection lamp bulbs will occur if the temperature at the hottest point

FIG. 8. Measurements of screen brightness uniformity with a Weston Photronic cell illumination meter.

FIG. 9. Equipment for measur- ing bulb glass temperature.

on the bulb is less than 500°C. The hottest spot usually lies exactly opposite the filament on the condenser side.

Fig. 9 illustrates the method of determining bulb glass temperature. As shown in the inset, a thermocouple is cemented to the bulb wall directly in front of the filament. Readings are taken with a cali- brated millivoltmeter or a potentiometer after temperature equi- librium has been reached. The photograph shows readings being taken on a lamp outside the projector. Naturally each projector design must be tested with each type of lamp in question, and with the lamp and machine in actual operation.

Hitherto, low voltage lamps requiring large rheostats and trans-

July, 1932] BIPLANE FILAMENT CONSTRUCTION 839

formers for their operation have been the only means of obtaining reduced source size and increased source brightness with a corre- sponding increase in screen illumination. Biplane projection lamps of the types here reported are being made for voltages ranging be- tween 100 and 130. This makes possible direct connection of the lamp to the lighting circuit, whether a-c. or d-c. or, as is preferable, the use of a small rheostat in series with the 100 volt lamp for close regulation and more uniform performance.

REFERENCE

1 DENNINGTON, A. R.: "Incandescent Lamps for Motion Picture Service," Trans. Soc. Mot. Pict. Eng. (April, 1918), No. 5, p. 36.

DISCUSSION

MR. GAGE: In the construction of the biplane filament as described in this article, two monoplane filaments are mounted side by side in such a way that if one observes perpendicularly to the plane of the filaments, the filaments of the rear plane fill in the gaps between the filaments of the front plane ; and if observed from this direction, the entire area of the filament appears to be covered with tungsten coils, giving the greatest brightness that can be obtained by this construction — that is, the projected area is nearly filled with hot tungsten. If, however, the lamp is slightly rotated, the individual filaments in the rear plane will be shaded by those in the front plane, and the appearance will be exactly the same as though only the front filaments were observed, and blank spaces will appear between the individual coils of the monoplane filament. It is true that these spaces can be filled with images of the filament by using a suitable concave mirror, but the efficiency of such reflection does not exceed 75 per cent when the lamp is new, and a very slight blackening of the bulb results in a drop in brightness to a low value. The result is that the maximum average brightness of the filament image is greatest along the axis of the condenser and the optical system of the projector; whereas the beams, proceeding from the filament in a diagonal direction and striking the edge of the condenser, have a low average brightness, owing to the spaces left vacant between the coils of the filament. As it is these diagonal rays that serve to illuminate the edges of the picture, a general tendency of the pro- jected screen image to be less intense toward the edge of the picture than in the center, is aggravated. While there are many arguments to the effect that it might be advantageous to illuminate the edge of the picture slightly less than its center, partly on the ground that the main interest of the picture generally occurs near the middle of the screen, careful observation of both professional and amateur motion pictures shows that the center of interest is often to be found at one extreme edge or corner of the picture, and apologies for ununiform screen illumi- nation can hardly be tolerated when optical systems allowing a high uniformity of screen illumination may be devised. To set the tolerances for what might be classed as satisfactory uniformity of screen illumination is, however, not my pur- pose in this discussion. In the case of biplane filaments, I suggest that in order to compensate for the natural tendency of optical systems to be more efficient in the

840 J. T. MELI [j. s. M. P. E.

axial than in the marginal direction, owing to the increased reflection of the diagonal rays and to the vignetting effect of the diaphragms in projection ob- jectives, the rear row of filaments be mounted directly behind the front row, thus showing such an aspect to the edges of the condenser as to provide greater bright- ness in this directions, in order to compensate for the lessened efficiency of the optical system. The spaces between the filaments, as seen in the axial direction will, of course, be filled as well as possible with the reflected light from the con- centric spherical mirror.

MR. RAYTON: It is well understood that standardization is one of the later stages in the life's history of any development. In fact, it has been stated that the next step is obsolescence. Nevertheless, there is nothing to be gained by postponing standardization longer than is necessary, and I, therefore, wish to second the plea that it be delayed no longer than is necessary. None of us would wish to see standardization accomplished at the expense of further improvement, and it is now a question whether the lamp manufacturer has reached a limit beyond which he is not soon likely to go. Conditions have been chaotic, but out of the chaos improvements have come. The lamp designer and the lens designer have at times been distressed by an apparently needless multiplicity of demands, but nearly all developments viewed in retrospect disclose an immense amount of apparently wasted effort. In view of the increase in brightness of the projected image now possible, this effort in connection with 16 millimeter projectors has not been in vain.

The biplane filament lamps have been a very interesting, and will doubtless prove to be a profitable, development. From the standpoint of the design of optical systems, the filament area has a tendency to be too small. If we ask the lamp designer for a larger area, he can meet the requirement only by increasing current consumption. This means more heat and possibly a larger bulb. If the diameter of the bulb is increased, the filament-to-condenser distance increases, and possibly no gain in illumination is realized unless we increase the size of the condensers and, in turn, the size of the whole projector. Relief has been offered in the form of offset filaments, but this is a complication for the lamp maker that multiplies the number of kinds of lamps he has to produce and increases costs.

At the present time, we would say that the lamp competent to produce the brightest possible picture with the best available optical system, employs a bi- plane filament lamp with a filament area of 8.5 by 8.5 millimeters, with distance from filament to bulb wall not more than 16 millimeters. Such a lamp would consume about 600 watts, according to the best information we have, and would require to be set ahead of the center of the bulb. These dimensions would not allow for decentration of the lamp, but would assume perfect adjustment.

MR. FARNHAM: Mr. Mili has brought out very well the advantages of the biplane filament construction over the monoplane form, but he has unfortunately not presented the complete picture with regard to 16 millimeter projectors. There is no question that the biplane filament construction places more hot tungsten within a given source area, and were that the sole criterion, all motion picture projector lamps would have been changed to the biplane form just as quickly as our recent improvements in filament wire made it practicable. However, the biplane filament is fundamentally more costly than the monoplane, and there are yet other considerations.

July, 1932] BIPLANE FILAMENT CONSTRUCTION 841

For wattages ranging from 50 to 300, monoplane construction appears to be more desirable, considering all factors. Optical systems generally employed for 16 millimeter motion picture projectors are so designed as to utilize a source area represented by a circle from eight to ten millimeters in diameter. Tubular bulb projection lamps of the 115 volt class, in wattages ranging from 50 to 300 watts, can have monoplane sources that either do not fill this usable source area or just fill it. Conversion of any of these lamps to the biplane construction will result in practically no gain in screen illumination because, although the mass of the source is placed slightly closer to the optical axis, the loss of light due to shadowing of the rear coils by the front coils and the reduced mirror advantage are so great as to make any net gain negligible in proportion to the increased lamp cost.

Where the maximum amount of light is required — with wattages above 300 — the biplane construction possesses the advantage. Monoplane sources greater than 300 watts are larger than the 8 to 10 millimeters usable source area, and hence part of the source is not utilized (although there is some increase in screen illumina- tion because of the somewhat higher average temperature of the usable area).

Considering the problem of ventilation, most projectors using lamps of the 50 to 300 watt range depend on either a natural or a moderate volume of forced ventilation through the lamp house; hence substitution of the biplane construc- tion with its inherent doubling of the wattage would in many cases result in excessive temperatures and poor lamp performance. Where the selling price of the projector will permit, a high volume of forced ventilation, of the order of 40 to 50 cubic feet of air per minute, requiring a high-speed fan and introducing the problem of noise elimination, lamps of 500 to 600 watt ratings in small tubular bulbs, offsetting the filament if necessary, can be employed. The biplane fila- ments of such lamps make full use of the available source area, and maximum screen illumination is obtained.

Projectors of the low and medium priced class can hardly be expected to carry the added burden of a greater lamp cost, particularly when the screen results are only slightly better or not better.

MR. MILI: Mr. Gage's suggested modification of the biplane filament con- struction has already been considered, and lamps having such a filament con- struction have been compared, using a !3/< inch aspheric 16 millimeter lens sys- tem, with the present staggered construction. This comparison indicates that a somewhat smaller total screen lumen output, with no appreciable increase in brightness at the corners, will be obtained by using this modified construction.

Mr. Ray ton's data on the best light source dimensions for 16 millimeter pro- jectors are n good agreement with our own findings. Such an experimental lamp was used in the tests herein reported.

Replacing a monoplane filament by a biplane filament of the same electrical rating and life will result in a gain in screen illumination, if the monoplane light source is too large to be accommodated fully by the lens system. This gain is even greater when the biplane is substituted for the C-13A filament now standard for the 1000 watt line voltage projection lamp. For example, in the most com- monly used 35 millimeter optical systems, the C-13D filament produces about 40 per cent greater screen brightness than the C-13A.

ILLUMINATION IN PROJECTION PRINTING OF MOTION

PICTURES*

CLIFTON TUTTLE AND D. A. YOUNG**

Summary. — In projection printing of motion pictures, it is essential that the illumination in the printing plane be uniform and of high intensity. It is also desirable that the light in the negative plane be more or less diffused in order to mini- mize the effect of scratches and reduce contrast. Since the introduction of diffusion into an optical system is accompanied by a loss of intensity and possibly also of printing plane uniformity, experiments have been made to determine what conditions are most favorable from the points of view of uniformity and diffuseness.

In the design of projection printing optical systems there are three points of special interest to consider: First, it is essential that there be a certain degree of uniformity of illumination in the printing plane. Second, it is desirable that the intensity be as great as pos- sible in order that the time of exposure may be reduced to a minimum. Third, it is important that the illumination in the negative plane be not entirely specular. Diffuse light tends to eliminate the print- ing of accidental scratches and dirt on the negative and to reduce the contrast of the optical image.

The fulfillment of the first and third of these desiderata almost invariably is prejudicial to the best interests of the second, and the conditions favoring the first are not always desirable from the point of view of the third. Thus, a study of all of the factors involved may lead to the discovery of a set of optimum conditions.

We have, therefore, attempted a systematic determination of essential data applying to various optical systems in the hope that the possession of such data may aid the designer of a projection printer in the choice of his optical system and its component parts.

While these data are primarily intended for use in conjunction with motion picture printing, they can in most cases either directly, or by extrapolation, be applied to other similar projection printing problems.

* Presented at the Spring, 1932, Meeting at Washington, D. C. Communica- tion No. 496 from the Kodak Research Laboratories.

** Research Laboratories, Eastman Kodak Co., Rochester, N. Y. 842

ILLUMINATION IN PRINTING

843

THE LAMP

Since the principal purpose of this work is to find the conditions which will give a maximum amount of light with acceptable illumina- tion conditions of uniformity and diffuseness, we have selected a high intensity source — the 500 watt monoplane filament projec- tion type lamp — for special study.

Regardless of the type of optical system to be used, it is of interest to determine the intensity distribution in planes parallel to the fila- ment and at various distances from it. This can be computed,1 but for multiple filament lamps the computation becomes very laborious, and, moreover, this calculation would disregard the effect of the glass bulb itself, one side of which acts as a cylindrical mirror, while the other side acts as a weak negative cylindrical lens.

PL/vise p or

SURFACE OF CCUU

-[

LLAMP L SCREW-B

PLAN VIE1W

FIG. 1. Apparatus used to determine the distribution of intensity in optical printing systems.

Consequently, it was decided to do this part of the work by mea- surement rather than by calculation. We have used the set-up shown in Fig. 1, using a Weston Photronic cell and a Leeds & North- rup high sensitivity galvanometer. The sensitive surface of the cell was covered with an opaque diaphragm containing a slit 1 milli- meter wide by 3 millimeters long. This diaphragm was made of aluminum leaf and was in direct contact with the sensitive surface. By turning screw A, the light-sensitive cell would be moved in a plane parallel to the filament any desired distance and in measured increments. By adjusting screw B, the distance, D, from the lamp filament to the cell could be varied. Thus, it is possible to measure the intensity in any plane, P, along a horizontal line through the axis, since the galvanometer deflection is proportional to the illumination. The intensity variation along this line was adopted

844

C. TUTTLE AND D. A. YOUNG

[J. S. M. P. E.

as a criterion of evenness of illumination. The data are shown graphically in Fig. 2: the ordinate scale representing logio photo- graphic meter-candles* with the lamp operated at its rated volt- age, the scale of abscissas representing the distance from the opti- cal center line in millimeters. This plot shows that to illuminate a motion picture frame 23 millimeters wide with no measurable falling off of intensity from center to edge, the frame must be located at least 280 millimeters from the lamp filament.

-50

OflT

C/M. CENTER

UUHC. >N

O •> \Q

FIG. 2. Horizontal distribution of intensity from a 500 watt projection type lamp at various distances from the filament.

It is of interest to present these data in the form of a table. To facilitate the application of the information to the various systems to be considered later, the circle diameter, which, for a given bright- ness variation tolerance, can be used at various distances from the lamp, is shown in Table I. The column headings are self-explanatory

* The effect on positive film of a source operated at a color temperature of 5000 °K. is used throughout this paper as a standard of comparison. Thus, a value given as 10,000 meter-candles indicates that the same effect would be pro- duced on positive film by an illumination of 10,000 meter-candles (visually de- termined) from a source of approximately daylight quality.

July, 1932]

ILLUMINATION IN PRINTING

845

Distance

from Lamp

Filament

to Plane

32 42 52

62

72

82

92 102 112 122 132 142 152 162

1.02 6.0 10.0 13.3 13.3 15.2 22.0 20.0 22.0 23.2 30.0 28.0 30.0 33.3 40.0

TABLE I (All Dimensions in Millimeters)

Center-to- Edge Brightness Ratio

1.04

9

12

18

22 25 32

35 33 40

1.06 10.0 16.6 22.0 25.0 32.0 36.0 36.8 42.0

1.08

11.6

20.0 25.0 26.6 35.0 38.4

1.10 15.0 22.0 28.4 32.0 38.6

An examination of Table I shows that there is an approximately constant ratio between the filament distance and diameter of field which can be used at any given uniformity tolerance. The value of this ratio is significant since it is indicative of the amount of light which can be gathered either by a condenser to be imaged on the film or by the film aperture itself, if no condenser is used. This ratio of filament distance to field diameter may be called the "/ value" and it seems worth while to show in Table II how this value varies with the uniformity tolerance. The last three values are extra- polated.

TABLE n

Center-to- Edge Brightness Ratio

1.02 1.04

.06

.08

.10

.12

.14 1.16

/ Value 4.4

2.9 2.5 2.2 1.9 1.5 1.2 0.9

In addition to the type of non-uniformity shown by the preceding data, there is another phenomenon which occurs to a greater or lesser extent with any glass-enclosed source. In planes close to the lamp bulb there are present ghost images of the filament, due probably to the mirror action of the back side of the lamp bulb. While these

846 C. TUTTLE AND D. A. YOUNG [j. S. M. P. E.

images are faint, they may be of sufficient intensity to cause trouble in some optical systems if the aperture is placed within 100 milli- meters of the filament.

»

PRINTING OPTICAL SYSTEMS

Optical System A. — The simplest optical system that can be used for projection printing consists of a lamp, a negative aperture, a diffusing material in the plane of the negative aperture, a lens to image this negative in a plane known as the printing plane, and a printing plane aperture. Such a system is shown in Fig. 3. The lamp, A, is a 500 watt monoplane filament projection lamp of the type previously discussed; C is the negative aperture so arranged that a diffusing material, B, can be added at will; D is the printing lens, and E the printing plane. In this particular set-up, D is a 75 millimeter, //4.5 Kodak anastigmatic lens. The distance, a, was made 100 millimeters; b and c are each 150 millimeters, and the printing lens works, of course, at 1:1. Once having chosen the printing lens and magnification, the dimensions b and c are definitely fixed. A reasonable value for dimension a must, however, be chosen. This distance should be so assigned that the lamp is as close to the negative aperture as is compatible with the largest permissible bright- ness change from center to edge of the negative frame. Since the maximum allowable brightness change is the limiting factor in the placing of the lamp, a quantitative value for this ratio by an analysis of the optical system of photographic reproduction as a whole may be determined. First, let it be assumed that uniform illumination in the positive plane is desired.

It can be shown that the image of a uniformly bright object formed by any lens will decrease in brightness from center to edge as the fourth power of the cosine of the angle which the extreme ray makes with the optic axis. If this calculation be carried out for a 50 milli- meter camera lens covering a frame 23 millimeters wide, it is found that the edge-to-center brightness ratio is about O.9.* If a negative of a uniformly bright object be developed to a gamma of about

* Besides the falling off of intensity as the fourth power of the cosine, with nearly all commercial lenses there is some additional falling off caused by lens barrel vignetting. Measurements of this effect made on a number of commer- cially available motion picture camera objectives show that the effect of vignetting results in an edge image brightness in extreme cases only one-sixth that of the center brightness, while it is seldom greater than one-half.

July, 1932]

ILLUMINATION IN PRINTING

847

0.5, it will have a center transmission which is about 95 per cent of the edge transmission. If this negative is evenly illuminated and printed at 1 : 1 with a 75 millimeter lens, because of the unavoidable cos4 falling off from the center to the edge, the image in the printing plane will be about 2 per cent brighter at the edge than at the center. To compensate for this, the lamp to negative distance, a, should

FIG. 3. Projection printing system A.

be so chosen that there is about 2 per cent falling off in illumina- tion from center to edge. This distance (see Fig. 2) proves to be about 100 millimeters. System A, shown in Fig. 3, can be used only with some diffusion in the negative aperture, as otherwise the printing lens would project to the printing plane remnants of the, filament image. The printing plane distribution of illumination is shown in Fig. 10 for various kinds of diffusion.

FIG. 4. Projection printing system B.

Optical System B. — The second common type of printing optical system is shown in Fig. 4. In this system, a filament image is formed in lens D, by lens C. Lens C is a 13 diopter spectacle lens and forms a filament image which just fills the //4.5 aperture of lens D, thus making the system as efficient as such a system can be. The illumination in the negative aperture falls off less than 1 per

848

C. TUTTLE AND D. A. YOUNG

[J. S. M. P. E.

cent from center to edge, and if it were not for the falling off due to the cos4 of the half angle of the printing lens, the same degree of evenness should obtain in the printing plane. Thus, the addition of diffusing material to a system of this sort is intended solely for the purpose of lowering the print contrast and decreasing the visi- bility of scratches and other accidental marks on the negative. The

FIG. 5. Projection printing system C.

brightness fall-off in the printing plane is shown in Fig. 1 1 for several varieties of diffusion. It is noticeable that this ratio remains nearly constant, regardless of the type of diffusion.

Optical System C. — The third type of printing system consists of a rectangular bar of glass or quartz, B, one end of which is covered with flashed opal No. 3 (see Table III), the other end, C, being

FIG. 6. Projection printing system D.

covered with some diffusing material. This system is shown in Fig. 5. The integrating bar depends for its effect upon the total internal reflection of all light passed through diffusion B. The multiple reflections tend to even out whatever initial non-uniformity exists. The glass or quartz of the bar must be very clear, white, and free from bubbles and striations. The edges of this rather

July, 1932] ILLUMINATION IN PRINTING 849

expensive and fragile device must be very perfect. The brightness fall-off in the printing plane for this system is shown in Fig. 12.

Optical System D. — The fourth system of interest has as its basic component the Ulbricht integrating sphere. This device is com- monly used as a diffuse light source, but apparently has not been used to any great extent on printers. The system as used by us is shown in Fig. 6. The sphere is about 5*/2 inches in diameter and painted with lithophone-sodium silicate paint. Various degrees of diffusion were placed over the aperture, with results as shown in Fig. 13.

When a lamp is used within an enclosure such as the whitened sphere, the filament is heated by re-radiation from the walls. The wattage supplied to the lamp no longer is a criterion of its color tem- perature and its life characteristics.

Precautions must be taken to insure that the results obtained with the lamp in the sphere are comparable with those obtained in the other cases.

Lacking an optical pyrometer to determine directly the color temperature of the filament, it has been assumed that at the same value of resistance for the filament (measured by E/I) it must be at the same temperature.

It was found that for the same resistance, the lamp in one measured case was consuming 14 per cent more wattage outside of the sphere than it did inside.

The Condenser Relay System. — A conventional type of relay system consisting of a condenser near the lamp which is imaged by a field lens on diffusing material* in the negative plane may be suggested as a possible means of illuminating the negative aperture. Use of such a set-up enables one to increase the distance between lamp and aperture to almost any desired value, but has little else to recommend it. Since the condenser is imaged on the negative aperture, it is obvious that the illumination on the condenser must be at least as uniform as that desired in the negative aperture. Aperture values which may be used with various degrees of uniformity in the condenser plane are shown in Table II.

* It does not seem worth while to consider the case where a lens is used in the negative plane to image the relay lens in the projection lens. In this case, the aperture of the system is fixed by the projection lens. Better results can be secured with optical system B, since the loss due to the added glass-air surfaces of condenser and relay lenses would be avoided.

850 C. TUTTLE AND D. A. YOUNG [J. S. M. P. E.

If it were practically possible to place the negative aperture almost in contact with the 500 watt bulb, the frame would collect from the filament center a cone of light equivalent to about //1 .5. The falling of! of intensity from center to edge of the frame would, in this case, be about 12 per cent. From the standpoint of intensity alone, there is then no advantage to be gained by using a condenser relay system, unless the aperture is greater than //1. 5 and unless a falling off of intensity from center to edge greater than 12 per cent is permissible.

Practically, it would probably be undesirable to place the negative frame closer than about 50 millimeters from the filament, thus gathering a cone of light not greater than about //2. The center- to-edge falling off would be about 10 per cent.

The previous statement, regarding the advantage of a condenser relay system may be amended to read: A condenser system of effective aperture slightly greater than //2* would result in a gain of intensity over that practical with no lenses between lamp and negative frame. Such a system, however, would have a center-to- edge brightness difference of more than 10 per cent. If a more uni- form illumination of the negative is required, there is no advantage whatever in the use of a condenser system.

DIFFUSING MATERIALS

Diffusing materials naturally fall into one of two general classes: First, those which depend on surface roughness for their diffusing action (ground or sand-blasted glass are examples of this class of materials), and second, those which scatter light because of small imbedded particles of foreign material or small air bubbles. To this second group belong the opal glasses — pot and flashed.

Ground glasses are prepared by pressing flat glass surfaces on a rotating plate sprinkled with some abrasive material, usually emery. Grades of powdered emery commonly used vary in average particle size from about 0.025 to 0.0004 inch in diameter. Results obtained in grinding glass appear capable of being repeated, so the fineness of the abrasive used seems to be a sufficiently good specification for the resultant surface. Sand-blasted glass can not be so specified as the resultant surface varies in roughness with the air pressure of the blast as well as with the particle size.

Opal glasses2 are produced by the addition of cryolite or other

* Allowing for the four glass-air surfaces of a condenser relay system with the unavoidable light loss of 16 per cent, this value should be about //1. 85.

July, 1932]

ILLUMINATION IN PRINTING

851

alumina-bearing materials to the melt, and perhaps in some cases by the introduction of small air bubbles in the melt. It is believed that in the best opals the diffusing particles are themselves trans- parent but of a different refractive index from the glass matrix. Now, the scatter becomes greater and the absorption becomes smaller as the particle size is increased until some optimum value is reached. For small particles, the scatter is approximately pro-

too

10

*>O

GO

10

FIG. 7.

ZO ^O 40 DEPARTURE Or ANGLE Or OBSERVATION O

Brightness distribution curves for various diffusing materials.

portional to the inverse square of the wavelength. Thus, a diffusing material may be loaded with particles of just the right size to act as the best possible diffuser for a given wavelength. Such a ma- terial may be as completely diffusing at a thickness of about 0.2 millimeter as it is at any greater thickness. Other pot opal diffusers are much less dense and do not reach their maximum diffusing efficiency under a thickness of 2 millimeters.

852

C. TUTTLE AND D. A. YOUNG

[J. S. M. P. E.

Flashed opal glasses vary greatly in their diffusing characteristics, some transmitting specularly in the far red and infra-red while diffusing well the shorter wavelengths; others are much less selec- tive. The thickness of flashing is not a criterion of diffusion ef- ficiency.

In subsequent remarks use will be made of goniometer distribu- tion curves in evaluating the diffusion characteristics of the various

300

ZOO

100

F-L.A.SHEO OPAL. NO \

DEZP/XRTURE: F-RONI NORNAAL. \\

OP~ A.1SG>1_EL OF" Oe>^>EZRVATrtOM ft. V

IO

4O

<bO

TO

FIG. 8. Brightness distribution curves for various diffusing ma- terials used in conjunction with optical system C (integrating bar).

materials. These curves,* such as are shown in Fig. 7, were plotted from data obtained in the following way: A Weston Photronic cell diaphragmed to a 1 millimeter slit opening, as before described, was mounted on the arm of a spectrometer table. The arm was rotated about a vertical axis passing through the center of the diffusing medium. The diffusing medium was illuminated by a collimated

* Distribution curves for ground glasses Nos. 2, 3, and 5 are not shown in Fig. 7 because they lie very close to the curves for ground glasses Nos. 1 and 4.

July, 1932]

ILLUMINATION IN PRINTING

853

beam of light normally incident. Now, the quotient of the galva- nometer deflection divided by the cosine of the angle of departure of the goniometer arm from the line normal to the sample, is a measure of brightness. A perfect diffuser, by definition, would appear equally bright to the cell mounted on the goniometer arm at any angle of departure from the optical axis up to (90— 1/°°). The distribution curve from 0 to 90 degrees for a perfect diffuser would, then, be the

386

\

NORMAL. \\\

Or ANQ»L-E OF" OBSERVATT\OM O-

O 10 ZO 3O 40 5O feO HO *0 30

FIG. 9. Brightness distribution curves for various diffusing mate- rials used in conjunction with optical system D (integrating sphere).

straight line, AB (Fig. 7). If we now plot distribution curves from 0 to 90 degrees of the various diffusing media, we can say that the diffusion efficiency of the medium is the ratio of the area under the curve for the. medium to the area under AB, which represents the distribution curve of a perfect diffuser. The value of diffusion efficiency thus determined is given in column 4, Table III. Column 5 gives the transmission of the medium as measured by a densitome- ter with a diffuse light source.

854

C. TUTTLE AND D. A. YOUNG

[J. S. M. P. E.

In the case of optical system A or B where the incident light is more or less specular, the data of Fig. 7 can be used to specify the diffusion in the negative aperture. In the case of optical systems C and D, where the incident light is already diffuse to a great ex- tent, distribution data applying to these particular set-ups must be obtained. Fig. 8 shows the angular distribution in the negative plane when the integrating bar is used. In all cases, the end of the bar next to the lamp was covered with flashed opal No. 3 (see Table III). The other end was covered with flashed opal No. 1, flashed opal No. 3, and ground glass No. 1 in turn.

TABLE III

Characteristics of Diffusing Media

Material Ground Glass	No.	1	Description Abrasive "fine x"	Total Trans- Diffusion mission Comments Efficiency % Last stage before polish 14.0 85.0
Ground Glass	No.	2	Abrasive	"smooth x"	Very fine texture	15	.1	85	.0
Ground Glass	No.	3	Abrasive	220 x		16	.0	80	.0
Ground Glass	No.	4	Abrasive	150 x		17	.8	76	.0
Ground Glass	No.	5	Abrasive	70 x	Texture too coarse	18	0	69	.0
					to be used in focus
Flashed Opal	No.	1	Flashing	0 3 mm thiok		80	.7	42	.7
Flashed Opal	No.	3	Flashing 0 1 mm thirk	36	.0	55	0
Pot Opal	No.	1	3.0 mm	. thick	Slightly yellow	89	3	10	0
Pot Opal	No.	2	2.0 mm	. thick	Slightly yellow	89	3	13.8
Pot Opal	No.	3	1.0 mm	. thick	Slightly yellow	89	3	27,	0
Pot Opal	No.	4	0. 5 mm	. thick		89	3	40. _	g
Pot Opal	No.	5	0.25 mm	. thick		89	3	52.5
Schott Pot Opal No. 1	3.0 mm	. thick	Neutral in color	89,	0	21.	0
Schott Pot Opal No.	2	1 . 0 mm	. thick	Not uniform enough for			42.	0
					printing
Schott Pot Opal	No.	3	0.5 mm	. thick	Not uniform enough for	56.	0	60.	o-
					printing

Fig. 9 shows distribution as measured for the integrating sphere, the exit aperture being covered with flashed opal No. 1, flashed opal No. 3, and ground glass No. 3. Diffusion efficiencies for systems C and D are shown in Table IV.

TABLE IV

Diffusion Efficiencies for Optical Systems C and D Optical System C Ground Glass No. 1

Optical System D

Flashed Opal No. 1 Flashed Opal No. 3 Ground Glass No. 3 Flashed Opal No. 1 Flashed Opal No. 3

47.6 77.5 64.0 51.0 83.0 64.0

July, 1932]

ILLUMINATION IN PRINTING

855

MEASUREMENTS OF UNIFORMITY AND RELATIVE INTENSITY IN THE PRINTING

PLANE

Apparatus similar to that shown in Fig. 1 was used to measure the intensity distribution in the printing plane for the optical systems previously described.

45

O b 10

DISTANCE. FROt^l OPTICAL CENTER UlNE

FIG. 10. Horizontal distribution of illumination in printing plane of optical system A with various diff users in negative plane.

The results are shown graphically in Figs. 10, 11, 12, and 13. The intensities plotted logarithmically are in photographic meter-candles. The abscissa values show horizontal displacement off the optical axis in millimeters.

856

C. TUTTLE AND D. A. YOUNG

[J. S. M. P. E.

Optical system A, B, or C can be modified by the addition of a concave mirror placed behind the lamp in such a way that a filament image is interlaced with the filament itself, resulting in an intensity increase of about 60 per cent throughout the system and having a negligible effect on uniformity.

For the convenience of the reader, the essential data for the dif- ferent systems are summarized in Table V.

TABLE v

Essential Data Showing Characteristics of Optical Systems

Optical System A

Optical System B

Optical System C

Optical System D

Diffusing Material

Ground Glass No. 1 Ground Glass No. 2 Ground Glass No. 3 Ground Glass No. 4 Ground Glass No. 5 Flashed Opal No. 3 Schott Opal No. 3 Pot Opal No. 5 Flashed Opal No. 1 Pot Opal No. 3 Pot Opal No. 1 No Diffusion Ground Glass No. 1 Ground Glass No. 5 Flashed Opal No. 3 Schott Opal No. 3 Flashed Opal No. 1 Ground Glass No. 1 Flashed Opal No. 1 Flashed Opal No. 3 Ground Glass No. 3 Flashed Opal No. 1 Flashed Opal No. 3

Axial Intensity Without With Concave Concave Mirror Mirror	Center Intensity Edge Intensity	Diffusion Efficiency
12,600 20,200	4.0	14.0
8,500 13,600	2.7	15.1
6,300 10,080	2.2	16.0
5,250 8,400	2.2	17.8
4,360 6,980	2.0	18.0
922 1,470	1.3	36.0
536 860	1.12	56.0
302 480	1.07	89.3
288 460	1.12	80.7
170 272	1.05	89.3
64 100	1.05	89.3
85,000 130,000	1.05	0.0
6,450 10,300	1.05	14.0
2,040 3,250	1.07	18.0
675 1,080	1.17	36.0
316 510	1.12	56.0
170 270	1.10	80.7
2,615 4,180	1.10	47.6
950 1,520	1.07	77.5
1,690 2,710	1.10	64.0
13,500	2.3	51.0
2,040	1.11	83.0
4,060	1.11	64.0

Discussion of Table V. — System A, with any diffusion less than that offered by flashed opal No. 3, is probably quite out of the ques- tion for projection printing because of non-uniformity in the print- ing plane. With diffusion efficiencies greater than 50 per cent, the uniformity would probably be satisfactory.

It is, of course, possible to modify system A by placing the negative plane closer to the lamp filament. An approximate determination

July, 1932]

ILLUMINATION IN PRINTING

857

both as to the resultant intensity and uniformity for various dis- tances of lamp to negative can be made by reference to the data of Fig. 2.

Optical system B, compared with A, shows a loss of axial intensity if any diffusion is placed in the negative plane, but a great improve- ment in uniformity over that of system A for diffusion efficiencies less than 18 per cent. It is evident that nothing is to be gained by

• NO OIF-ru5ION

-C.HOUMD C,LO,Sr> NO.

•GROUND <M-A.S% NO.%

•> (0

DIVIANCC PROM

FIG. 1 1 . Horizontal distribution of illumina- tion in printing plane of optical system B with various diff users in negative plane.

the use of system B, when diffusion efficiencies as great as 36 per cent are used.

Comparing the results of optical systems C and D, it seems that the uniformity of illumination is slightly better with the integrating bar, but that the available intensity is slightly greater with the sphere. Both are excellent as diffuse sources. The sphere covered with ground glass unfortunately can not be used because it does not illumi- nate the printing plane with satisfactory uniformity.

858

C. TUTTLE AND D. A. YOUNG [J. S. M. P. E.

O 5 10

DISTANCE: FROM OPTICAL. CENTER LINE: FIG. 12. Horizontal distribution of illumination in printing plane of optical system C with various diffusers in negative plane.

3.0

J FRAME: EIDOE:

|\G>ROUNO

OPAL_ NO.

FLASHEIO OPAL. tHO. »

O «b 10

OJSTANCEI FROM OPT\CA\_

l_\NE:

FIG. 13. Horizontal distribution of illumination in printing plane of optical system D with various diffusers in negative plane.

July, 1932] ILLUMINATION IN PRINTING 859

THE FUNCTION OF DIFFUSION IN AN OPTICAL SYSTEM

The most important function of diffusion in a projection printing system is to decrease the imaging of scratches which may be present on the negative. Since an increase in diffusion efficiency usually results in loss of printing intensity, it seems important to find out just how much diffusion is needed to accomplish the desired end.

FIG. 14. Microdensitometer traces of scratch images.

It is difficult to get quantitative data on the subject, but the results of three different experiments are discussed briefly below.

The first experiment consisted in printing with optical system B the image of a scratch made with a diamond pyramidal point on a flashed photographic density. By placing various degrees of diffusion immediately behind the negative it was possible to vary the visi- bility of this scratch on the prints. From a visual inspection of

860

C. TUTTLE AND D. A. YOUNG

[J. S. M. P. E.

these prints, it appears that there is a very little improvement in scratch elimination beyond a diffusion efficiency of 60 per cent.

In the second experiment, a long scratch which varied in depth from end to end was produced on a piece of flashed photographic density. This wedge-shaped scratch was made by causing a pyrami- dal cut hardened steel point to travel along the strip a distance of 150 millimeters at an increasing load. The density was then mounted under a traveling microscope in such a way that the scratch could be observed throughout its length at a magnification of

80

60

40

20

I 0

SCRATCH IMAGEl-CONTRAVT RATIO

1.0 I.I I.I I.* I.* l-t 'ft

FIG. 15. Curve showing relation between contrast of a scratch image and the diffusion efficiency existing in the negative plane.

Various degrees of diffusion were then placed under the scratch and readings were taken of the disappearance point. Three observers seemed to be in remarkably good agreement as to the whereabouts of these points. Results of the experiment are shown in Table VI. The third method of attack on the scratch elimination problem was as follows: scratches about 0.01 millimeter wide on negative material were projection printed at a magnification of 250# on posi- tive film. Density variation of these scratch images was examined by means of a recording microdensitometer. Typical traces ob-

July, 1932J

ILLUMINATION IN PRINTING

861

TABLE VI

Minimum Load to Cause a Scratch Visible with Various Degrees of Diffusion

Diffusion Efficiency Load (Cms.)

1.2 3.0 3.0

for Disappearance of Scratch

0.0 36.0 40.0

6.7 6.7 6.7

42.0 80.7 89.3

tained on this instrument are shown in Fig. 14, the density values of the scratch image and its background being marked on the traces. By applying the usual methods of photographic photometry, it was possible to determine the contrast existing between the scratch image and its background for various degrees of diffusion in the negative plane. Fig. 15 shows a curve which correlates the contrast of the optical image of a scratch and the diffusion efficiency in the

30

20

0*7. DIFFUSE

FIG. 16. Family of characteristic curves showing the effect of the use of various degrees of diffusion in printing the same negative.

negative aperture. A glance at this curve shows that the contrast between a scratch image and its background will continue to de- crease as long as the diffusion efficiency increases. The curve shows very plainly that a diffusion" efficiency in the neighborhood of 55 per cent is about as high as is necessary since the scratch-background contrast ratio is changed very little by any increase in diffusion efficiency beyond 55 per cent. For values below 55 per cent, the

862

C. TUTTLE AND D. A. YOUNG [J. S. M. P. E.

scratch-background contrast increases very rapidly as the diffusion efficiency is decreased. Ground glass seems to be of very little use as a diffuser. These data seem to bear out the conclusion which might be drawn from the other two experiments.

The second important point in regard to the amount of diffusion used in a projection printing system is its effect on the contrast of the projected image. It is well known that the contrast of an image in a specular system is much greater than that in a diffuse system. To study this effect in the various optical systems, a series of flash densities on motion picture panchromatic negative were measured under various diffusion conditions. The results are shown in Table VII.

TABLE VII

Density as Determined in Various Optical Systems

DH-	Dm (« = 89.3%)	Dm	/DM-	Dm 80*7%)	Dm	/DH-	Dm (« = 56%)	Dm/D'f	Dm (f = 36%)
0	.17	0.16	0.	95	0.17	1.	00	0.19	1.12	0.	21
o	.48	0.46	0.	96	0.47	0.	98	0.51	1.06	0.	53
0	.63	0.63	1.	00	0.66	1.	05	0.70	1.11	0.	75
0	.72	0.75	1.	04	0.78	1.	08	0.83	1.15	0.	90
1	.0	0.98	0.	98	1.06	1.	06	1.11	1.11	1.	21
1	.4	1.38	0.	99	1.44	1.	03	1.52	1.09	1.	68
1	.7	1.68	0.	99	1.80	1.	06	1.89	1.12	2.	10

Average Density Factor

0.99

1.04

1.11

D-H- Dm/DH-	Dm (« = 17.8%) Dm/D-H-	Dm (« = 15%)	Dm/Djf	Dm 0%)	Dm/DI
0.17	.23	0.24	.41	0.24	1.41	0.29	1.70
0.48	.10	0.64	.33	0.70	1.46	0.85	1.77
0.63	.19	0.89	.41	0.96	1.52	1.15	1.83
0.72	.25	1.08	.50	1.15	1.60	1.34	1.86
1.0	.21	1.47	.47	1.47	1.47	1.82	1.82
1.4	.20	2.00	.43	1.89	1.35	2.59	1.85
1.7	.23	2.52	.47	2.68	1.57	3.00	1.76

Average Density Factor 1 . 20

D-ff = diffuse density Dm = measured density e = diffusion efficiency

1.43

1.48

1.80

It is of interest to show the effect of the projection system on the effective gamma of the negative. The family of curves shown in

July, 1932] ILLUMINATION IN PRINTING 863

Fig. 16 is illustrative of the effect on negative gamma when various degrees of diffusion are used.

CONCLUSIONS

The following is a list of the systems which give a diffusion ef- ficiency of over 55 per cent and which, at the same time, are capable of a high illumination and uniformity in the negative plane :

1. Optical System A with mirror with Schott Opal No. 3, 860 meter-candles.

2. Optical System A with mirror with Pot Opal No. 5, 480 meter-candles.

3. Optical System C with mirror with Flashed Opal No. 1, 1520 meter-candles.

4. Optical System C with mirror with Flashed Opal No. 3, 2710 meter-candles.

5. Optical System D with Flashed Opal No. 1, 2040 meter-candles.

6. Optical System D with Flashed Opal No. 3, 4060 meter-candles.

Optical system B is not mentioned in this table, since there is nothing to recommend its use with diffusion efficiencies greater than 36 per cent.

Optical system A can be used to produce higher printing intensity if the distance between the lamp filament and negative plane is de- creased. When this is done, the uniformity of illumination is, of course, decreased.

For the usual run of studio negatives of which the densest have a total diffuse transmission of about 2 per cent,3 a maximum printing exposure of about 80 meter-candle-seconds would be required. As- suming that the projection printer pull-down is designed to expose for one-half of the cycle, one can compute the requisite illumination approximately

160 R = I

where R is the rate in pictures per second and / is the intensity in meter-candles which must be available in the printing plane.

The procedure which was followed to determine the value of diffiusion efficiency is not easy to follow without special equipment. It is also true that even with a reliable goniophotometer, the adjust- ments must be carefully made in order to get true distribution curves. Considerable reliance may be placed upon the distribution data which have been presented, not because of any certainty as to the absolute accuracy, but because all measurements have been made with the same set-up and therefore the relative values of areas be- neath the curves should be significant.

As a simple method of determining the relative diffuseness of

864

C. TUTTLE AND D. A. YOUNG

various systems, it seems possible that the measurement of density in a system may be a reliable criterion of diffusion efficiency.

With this end in view, the relation between diffusion efficiency and the ratio of effective density to diffuse density for panchromatic negative film is shown in Fig. 17. By using the same photographic material, one should be able to determine the diffuseness of a system on the basis of the observed density ratio. If the ratio of density

90

40

20

10 1.6 ^ 20

EFFECTIVE: DENSITY/DIFFUSE DEN

FIG. 17. Curve showing the relation be- tween diffuseness of a printing system and the effective "q factor" or density ratio (for Eastman Type II panchromatic negative).

measured in the system to diffuse density is less than 1.13, it should be safe to assume that the diffuseness of the system exceeds the arbitrary value, 55 per cent, which appears to be sufficient diffusion for the avoidance of scratches.

REFERENCES

IGLASEBROOK: Diet, of App. Phys., 10 (1923), p. 422.

2 RYDE, W., AND YATES, D. E.: "Opal Glasses," J. Soc. of Glass Technology, 10 (1926), p. 274.

3 TUTTLE, C.: "On the Assignment of Printing Exposure by Measurement of Negative Characteristics," J. Soc. Mot. Pict. Eng., 18 (February, 1932), No. 2, p. 172.

A NEW LIGHT CONTROL FOR PRINTING MACHINES* KURT SCHNEIDER**

Summary. — This paper describes a new device for automatically controlling the intensity of the printing light so that the successive scenes of each print receive the respective exposures that have been previously assigned to them. By means of a keyboard having numbered keys corresponding to twenty-four light intensity steps, the control can be rapidly set for a negative having as many as one hundred and sixty scenes. A recording indicator enables a quick check-up with the timing card, and facilitates an accurate control of printing.

In the printing of motion picture film, the necessity of employing a means of varying the exposure intensity to obtain a print of uniform density, regardless of negative density, has long been recognized. One of the earliest methods used to accomplish this consisted in employing a device which allowed the lamp used ki the printer to be placed at various distances from the printing aperture, thus varying the intensity of the light at the point of contact between the negative and positive during the printing exposure. The device mentioned was manually operated, depended entirely on the skill of the operator for results, and could be used only with very slow printers. As the length of pictures and the number of scenes increased and large numbers of prints became necessary, automatic devices for accom- plishing this purpose were developed.

These devices may be divided into three classes, namely:

(1) Devices in which the exposure is varied by changing the area of the aperture through which the exposure takes place. This type of device, which operates with constant speed printers in which the time of exposure is fixed, has been used in various forms either as a manually operated, semi-automatic, or fully automatic light control.

(2) Devices used with the usual type of printer operating at constant speed and with fixed lamp position, in which the intensity of the printing light is changed by varying the amount of resistance in the lamp circuit. This method is fully automatic when the control

* Presented at the Spring, 1932, Meeting at Washington, D. C. ** Oehler Machine Co., Inc., Long Island City, N. Y.

865

866

KURT SCHNEIDER

[J. S. M. P. E.

is properly connected with a circuit interrupter, a device which is actuated by notches on the edge of the negative film.

(3) Devices in which the intensity of exposure is varied by inter- posing a specially printed film or a filter having the proper opacity between the light source and the printing aperture. This type of light control is fully automatic, although it has not been widely used.

The device about to be described comes under the second classifica- tion. The same method of control, however, can be used to control

FIG. 1. Front view of the keyboard auto- matic light control, showing the keyboard, the scene numbering drum, and the recording indicator.

the opening of an optical diaphragm instead of the amount of resis- tance in series with the printing lamp.

The keyboard automatic light control is a new device to be used with motion picture film printing machines for automatically con- trolling the intensity of the printing light, so that the successive scenes of each print receive the respective intensities of exposure pre- viously assigned to them. Before a negative is printed, the correct printing intensity for each scene must be charted. During the print- ing operation, the intensity of the printing light is regulated to give

July, 1932]

A NEW LIGHT CONTROL

867

the charted light intensities by automatically varying the amount of resistance in the lamp circuit. In this way, differences in the negative density are compensated for, and a print of uniform density is ob- tained.

The assignment of printing exposure to each scene of a negative is commonly known as "timing," and in practice is accomplished visually by an expert judge of negatives or with the aid of a sensi- tometer. From these assigned printing intensities, the keyboard automatic light control can be rapidly set by means of numbered keys corresponding to a series of light intensity steps. Although twenty- two is the largest number of steps of light intensity ordinarily used, the keyboard automatic light control has twenty-four keys for registering light intensity values, the two additional keys serving special pur-

FIG. 2. Side view of the keyboard automatic light control.

poses. The device can be set for a negative having as many as one hundred and sixty scenes.

As the light control is set, the number of each successive scene and the number representing the degree of exposure intensity are shown by a recording indicator. This feature of the device consists of a bank of pilot lamps, one of which illuminates the number representing the exposure intensity being set or with which the film is being printed at the time; and a drum dial graduated from 1 to 160, representing the numbers of the scenes being set or printed. The graduated drum is rotated an equal step each time a key is pressed. It is viewed through a rectangular lens through which the scene numbers and the fixed scale indicating the "timing," "printing," and "clearing" positions may be clearly seen. The recording indicator allows ac-

868 KURT SCHNEIDER [J. S. M. P. E.

curate setting, enables a quick check-up with the timing card, and facilitates an accurate control of printing.

The operation of the keyboard automatic light control is very simple. In order to set the control for the light intensities given by a timing card, the switch at one side of the control must first be placed in the timing position, indicated by T. This releases the keys, which are all automatically locked when the switch is in the printing posi- tion. Pressing the key marked "Dial" then permits the graduated drum dial to be rotated by means of the dial knob at the left of the control. When scene number 1 is opposite "Timing," as viewed through the rectangular lens, the control can be set by pressing the numbered keys in the same order as the succession of light intensity values given by the timing card. A good plan is to have the timing card give the scene numbers in their consecutive order as well as their required light intensities. This will permit very accurate as well as rapid setting, because the number of each successive scene and its corresponding light intensity are simultaneously given by the recording indicator.

It hardly seems necessary, but a check-up of the timed scenes can be quickly made by returning scene number 1 of the graduated drum dial to its original position opposite "Timing" and then operating the "Shift" key, which causes the successive scene numbers and their corresponding light intensities to be repeated by the recording in- dicator.

If it is desirable to change the light intensity of any scene, for instance, scene number 158, the graduated drum dial is rotated until scene number 158 is opposite "Clearing," after which the operation of the key marked "One Clear" releases the light intensity setting of that particular scene. To assign a new light intensity to scene 158, number 158 of the graduated dial is brought opposite "Timing" and the key representing the new degree of light intensity is pressed.

The keyboard automatic light control is connected electrically to the printer lamp and circuit interrupter of the printing machine. After the control has been set, according to the light intensities given by the timing card, the graduated dial is brought opposite "Printing," and the switch placed in the printing position. The device is then ready to control the printing operation so that the successive scenes of each print receive the exposures that have been assigned to them. During printing, the intensity of the printing light is controlled by automatically varying the amount of resistance in the lamp circuit.

July, 1932] A NEW LIGHT CONTROL 869

The change-over mechanism is quick and positive in action. The contacts of the light control are made of an alloy of platinum and iridium and are entirely enclosed, thus insuring against contact faults. When the required number of prints has been made, the control can be set for a new negative in the following manner. The switch is placed in the timing position and the drum is given one complete revolution while the key marked "All Clear" is pressed down. The light control can then be set for the light intensities given by the new timing card in the same way as described before.

The essential elements of the keyboard automatic light control consist of the following:

(1) A drum adapted to carry steel balls in selected circumferential positions on its outer face.

(2) A system of key-operated levers by means of which the balls are placed in their selected circumferential positions or removed from the face of the drum.

(3) A contact mechanism consisting of a number of plunger contacts which are closed one at a time by the cam-like action of the balls as the drum is intermittently rotated.

THE DRUM

The drum is concentrically mounted within a stationary cylinder on a common shaft with a ratchet gear, a ball elevating wheel, and a graduated dial. It is intermittently rotated by a specially designed solenoid movement or by levers operated by numbered keys. On the outer face of the drum are a number of lateral carrier grooves which are transversely cut by longitudinal clearing grooves. The inner face of the cylinder also contains longitudinal grooves which are called guide grooves. As the drum is intermittently rotated, the balls that have been placed on its face are maintained in their relative circumfer- ential positions by means of the drum carrier grooves and the cylinder guide grooves.

THE KEYBOARD

The keyboard has twenty-four keys numbered from 1 to 24, and four keys marked respectively, "Shift," "Dial," "One Clear," and "All Clear." A safety device prevents the operation of more than one key at a time. Operating a numbered key causes a ball from one of the channels in the cylinder to be pushed into one of the carrier grooves of the drum. The channels, of which there are twenty-four, each corresponding to a numbered key, receive balls from a magazine

870 KURT SCHNEIDER [J. S. M. P. E.

and are constantly kept full. When a numbered key is pressed, a ball is placed in a selected position on a lateral line of the drum surface. Each time a numbered key is released, the drum is rotated an equal step.

By means of the "Shift" key, the drum is rotated in equal steps without placing balls on the drum surface. The "Dial" key releases the ratchet mechanism, allowing the drum to be rotated freely by means of the knob at the left of the machine. Operating the "One Clear" key gives a lateral movement to a sliding member in the cylinder. This sliding member has longitudinal grooves which are in line with the grooves of the cylinder, and are somewhat greater in length than the diameter of one ball. Its lateral movement across the face of the drum causes one of the balls to be transferred from its position in a carrier groove to a clearing groove, from which it passes to an elevating wheel and is returned to a magazine located above the cylinder.

Operating the "All Clear" key gives a lateral movement to another sliding member in the cylinder. This sliding member also has longitudinal grooves in line with the cylinder grooves. In addition, it has angular grooves, one of which is placed next to each longitudinal groove. When the "All Clear" key is pressed down, the "All Clear" slider is moved laterally across the face of the drum and the ratchet mechanism is released. As the "All Clear" key is held down and the drum is manually rotated, the balls are shunted by means of the angular grooves and transferred from the guide grooves to the clearing grooves. In this way the balls are all removed from their selected circumferential positions with one complete revolution of the drum.

THE CONTACT MECHANISM

The contacts are made of a special alloy and are entirely enclosed. They are of the plunger type and are closed under pressure by the circuit-closing balls which act very much like cams as the drum is rotated. Two independent circuits are closed by each contact plunger, one being the lamp resistance, while the other is the pilot bulb indicating the degree of printing light intensity. Each time a numbered key is operated, a pilot lamp illuminates the number corresponding to the key number and the required degree of light intensity. The bank of pilot lamps may be said to act as a recording indicator because, after the machine has been set, the series of light intensities that have been assigned to the scenes of a negative may

July, 1932] A NEW LIGHT CONTROL 871

be repeated by operating the "Shift" key which causes the numbers representing the required degrees of exposure to be illuminated successively in the same order as originally given by the timing card. During printing, the bank of pilot lamps, in combination with the graduated dial, simultaneously give the number of each scene and the light intensity with which it is being printed.

Before printing, number 1 on the graduated dial is brought opposite "Printing," and the selector switch at the left of the machine is placed in the printing position. In this switch position, all the keys are locked. The first circuit-closing ball is already in position as the printing machine is started, and remains there until the first notch of the negative film passes the circuit interrupter of the printing machine. When the interrupter circuit is closed at the beginning of the notch, a solenoid magnet in the light control compresses a me- chanical force which is released when the interrupter circuit is broken. In this way, an extremely fast change-over to the succeeding light intensity is obtained and arcing is reduced to a minimum.

For each additional positive print all that is necessary to be done is to return number 1 on the graduated dial opposite "Printing" and to place the selector switch in the printing position before running the film through the printer.

CONCLUSION

The following advantages may be found in the keyboard automatic light control:

(1) The large number of scenes that can be controlled by a small and compact device.

(2) The ease and rapidity with which the settings may be made.

(3) The recording indicator which enables an accurate control of timing and printing, and allows a quick check-up to be made.

(4) The fast and positive change-over mechanism which eliminates light lag caused by magnet sluggishness and the slow-acting escape- ment movement of the usual type of light-board control.

(5) The prevention of contact faults by the use of an alloy of platinum and iridium for contacts.

SOUND MOTION PICTURE EQUIPMENT FOR THE U. S.

NAVY*

S. W. COCHRAN**

Summary. — A brief account of Die origin of sound motion pictures in the U. S. Navy; treats of the problems involved in designing equipments satisfactorily for naval service, and describes in detail an RCA Photophone Type PG34B1 equipment, the largest of three types of equipment being supplied to the U. S. Navy. The equip- ment described herein is designed for service on the largest type of fighting ship. It is devoid of batteries, performs in accordance with naval specifications, is easily serviced, and is provided with a remote volume control.

HISTORICAL

Within the past year the U. S. Navy has contracted for sufficient sound motion picture reproducing equipment to equip practically every ship and shore station within that vast protective organization of the nation. The unsuspecting motion picture engineer is proba- ably unaware of the fact that motion pictures form the basis of approximately 50 per cent of the entertainment of the U. S. Navy.

It is the purpose of this paper to present a brief history of sound motion pictures in the Navy, to relate briefly the origin of the Naval specifications covering this type of equipment, and to discuss the problems involved in the design of equipment satisfactory for Naval Service. The solutions to the problems confronted will be treated in conjunction with the description of an RCA Photophone PG34B1 equipment, the largest of three types being supplied to the U. S. Navy, and designated by the Navy as the Type 1, Class A, equipment.

The magnitude of the present motion picture organization within the Navy and the extent of its facilities are amazing. There are 287 exhibiting units, with a potential audience of approximately 80,000 men, distributed over the major portion of the globe. In the past, the Navy has leased two prints of each of approximately 300 feature pictures each year.

Naval film is serviced and distributed by three principal exchanges located in New York, N. Y., San Diego, Cal., and Cavite, P. I.

* Presented at the Fall, 1931, Meeting at Swampscott, Mass. ** RCA Victor Co., Camden, N. J. 872

PICTURE EQUIPMENT FOR U. S. NAVY 873

These exchanges are supplemented by fleet exchanges, one of which is assigned to each major operating unit of the fleet, and by smaller exchanges at Naval bases.

There are two very important reasons for the advent of sound mo- tion pictures in the Navy — first, the ever-prominent policy of the Navy to be modern and, second, the shortage of high quality silent feature pictures. The Navy has already purchased approximately 255 sound features and will continue to purchase them at the rate of approximately 300 sound features per year. A film library consisting of 1200 features will be maintained and replenished at the rate indicated. This briefly introduces the subject of this paper.

On what basis does the average exhibitor select the sound equip- ment for his theater? It is probably safe to assume that listening tests, price, and manufacturer's performance data on competitive equipment are the determining factors. It is Naval routine to purchase materials and equipment from an approved list of suppliers by virtue of open bidding on a specification. The first logical step in this system is the completion of the specification. The Navy called in the products of the suppliers of sound reproducing equipment. Competitive equipment was tested scientifically on the same test floor, with the same test equipment, and under the same test condi- tions.

These laboratory tests furnished a positive indication of what was available in the industry, and provided a means of determining what characteristics of sound reproducing equipment best fitted the theater. They did not indicate the specific requirements of the Naval service. It was for the purpose of ascertaining the major specifications of a sea going motion picture sound reproducing equipment, that apparatus of an approved supplier was assigned to sea duty aboard a battleship. Many interesting and complicated problems, which will be treated in the following paragraphs, were studied, and an abundance of valuable data and information was collected. This is the origin of the present Naval specification for sound motion picture equipment.

A modern high quality sound reproducing equipment, of the type used in the better theaters, is permanently installed in a location free from serious atmospheric effects and devoid of severe radio frequency fields. The major units of such an installation are grouped relatively close together. The audience which the equipment serves may vary in magnitude, but at all times occupies an area well distributed about the source of sound.

874 S. W. COCHRAN [j. S. M. p. E.

Each large ship of the Navy carries a battery of radio transmitters operating at wavelengths covering nearly the entire radio frequency band. The high frequency radiation is extremely serious. It penetrates practically every part of the ship. The problem of ground- ing becomes complicated. Potentials as high as 1000 volts have been measured between the muzzle of a gun and the armored deck below.

Salt air and salt water are ever-present on shipboard. Sea water is the general purpose cleansing agent on "men of war," and it can not always be confined to the object of the cleaning process. Equipment and instruments aboard ship are constantly exposed to sea air and are often exposed to sea water. The effects of immersion or exposure are twofold — the metal corrodes, and the insulating materials acquire a highly conducting coating that greatly lessens the value of the material as an insulator. It is therefore very apparent that metal contacts, terminals, and machine parts must be inherently self-protecting against the elements if they are to do Naval service. Insulating materials must be non-hygroscopic.

The topside or main deck of a battleship serves a multitude of purposes. Among other things, it is the ship's theater. A deck area which may be a gymnasium and a general work area in the morning, and the scene of gun drills in the afternoon, serves as the theater of the ship in the evening. This arrangement implies that certain parts of the sound motion picture equipment must be on the topside to operate during the performance. From experiments at sea, it was found advis- able to locate as much of the total equipment as possible below the weather deck, so as to receive as much protection from the elements as a location of that nature can afford. A layout of this sort necessitates that the projectors and associated sound mechanisms and apparatus be linked with the heavier amplifiers, power supply units, and controls below decks by means of cables. The audio cable and transmission system must be so designed and constructed as to prevent serious effects from the radio frequency fields, previously mentioned, and yet maintain the frequency response standards required by Naval specifications. It is essential that complete control of the equipment be centered at the projector station, the logical place for control operations. Under these circumstances, overload and automatic protective devices on the equipment below decks are imperative. A complete system of remote control is also involved.

A visit to a Naval vessel is very impressive, and is particularly so to the mechanical engineer. Each part of the ship and each part of the

July, 1932] PICTURE EQUIPMENT FOR U. S. NAVY 875

equipment is the symbol of ruggedness and durability. That is Navy standard. If one wonders why it should be so necessary, the havoc wrought by concussion from gunfire and vibration would be ample evidence to be convincing to any one. Our floating fortresses are designed primarily for one purpose which they serve nobly. There is little room allotted for entertainment facilities. The service demands, therefore, that the sound equipment for shipboard use be sturdy, durable, and rugged, yet condensed in form and light in weight.

The spread of a sea going audience is hardly comparable even with the opera house, with its innumerable balconies. On the battleship the picture performance is staged usually on one of the quarter decks, an area approximately 25 feet wide and possibly 180 feet in length. The size and shape of the audience varies with ships, but it is safe to suggest that the average audience on the deck is approximately the shape of the quarter deck described. The turrets, masts, rigging, and superstructure form the balconies and boxes. The sky seems to be the vertical limit. Picture the audience then as a long narrow body with great vertical depth made up of attentive movie fans, all of whom are anxious to hear the sound accompanying the picture but who possess no desire to be blasted by volume approaching the threshold of feeling.

This sea going audience is a regular one and, barring very inclement weather or ship's activities which prevent, the motion picture per- formance is a nightly feature. In the era of square riggers, the problem of securing a practical screen might have been an easy one, but sails on "men of war" have long since passed. In the past, the Navy has witnessed silent features projected on a heavy screen of canvas usually laced over a rugged pipe frame, which, when properly braced by guys, is the vertical support. The advent of sound compli- cates the screen situation. The old canvas screen is not sufficiently porous to permit its use as it is, and can not be satisfactorily punched because the punching process does not leave the holes clean. The problem was that of obtaining a screen that had a transmission efficiency of 90 per cent at 1000 cycles, that would stand a grab test of 335 pounds in the warp and 250 pounds in the filling (equivalent to that of very heavy canvas), and that would submit to washing in salt water. It must also stand up under the siege of fuel oil soot of high sulfur content, salt spray, and rain.

In addition to the problems which originated due to the sea going nature of the application, there was that part of the Naval specifica-

876

S. W. COCHRAN

[J. S. M. P. E.

tion that required that no batteries be utilized in any part of the equipment. It was generally known that copper oxide rectifiers were not desired by the customer. Naval specifications specified 115 volts d-c. as the available power supply for all equipment. As a result of these three limitations it must be apparent that the final design must incorporate some form of power conversion unit, and also utilize the available power source to the maximum extent, if complica- tions were to be limited to a minimum.

In the foregoing paragraphs an attempt has been made to describe

PROJtCTOR-B

PROJECTOR -A

XRHH »

"*l*a->'>» -a* 4"T :m*l]I-

I t '" * JT ro (is V/JIT cTc

MOTOR-GENERATOR

o t r

FIG. 1. Schematic diagram illustrating the major parts of a Type 1, Class A, equipment.

the conditions and limitations under which Naval sound motion picture reproducing equipment must operate. The remainder of the paper is devoted to the description of equipment designed and con- structed to comply with Naval specifications. Every effort has been made to associate the problems at hand with the solution as exemplified in the final product.

It seems logical for the sake of clarity, first to present an ap- proximate diagram of the equipment to be described. Fig. 1 shows the major parts of a Type 1, Class A, equipment. The location of the units varies with the ship. In general, two projectors with sound

July, 1932] PICTURE EQUIPMENT FOR U. S. NAVY 877

heads and associated control equipment, a remote volume control station, a screen, and four reproducers comprise that portion of the total installation located on or above the weather deck. The amplifier, motor-generator, and magnetic starter are assumed to be mounted below decks.

The projectors may be permanently mounted in a compartment in the main-mast, on the superstructure, or in a portable booth on the quarter deck. The sound screen is erected on the quarter deck possibly 100 feet from the projector station. The four reproducers are suspended behind it or above it by means of a ship's boat crane or suitable frame work. The remote volume control station is situated in the first few rows of the audience which occupies the quarter deck- area in front of the screen. The amplifier, motor generator, and associated equipment are located adjacent to each other in a protected area below decks. At best, this is an approximation of the possibili- ties involved in making shipboard installations on the various types of battleships. With this picture of the general arrangement of units in mind, each part can be described and its relation to adjacent units more easily understood.

PROJECTOR

The projector utilized in this application is a standard Simplex Type "S" machine modified to meet the specific requirements of Naval service. All metal contained in the machine is inherently resistive to corrosion. Sheet metal is required to be brass or an approved aluminum alloy. Bolts and rivets in sheet metal are of the same material as the metal, and bolt heads in exposed locations are copper plated. The projector is finished with a weather-resisting enamel. It is safe to assume that metals which are inherently resistive to corrosion and which are further protected by a suitable finish must withstand the siege of a humid salt atmosphere and salt water, in so far as corrosion is concerned. There can be no extensive electrolytic action due to contact of dissimilar metals when due precaution is exercised in joining similar metals with bolts and rivets of the same metal. These are the conditions under which the projector is constructed.

The Navy leases two prints of each feature picture it sees fit to exhibit, which must serve the entire Navy. It is only natural that the organization should be vitally interested in the manner in which this film is handled. The projector contains sprockets which are

878 S. W. COCHRAN [j. s. M. p. E.

hardened, ground, and polished to prevent excessive wear and de- formation of sprocket teeth. The shape of the sprocket teeth, the size of the sprockets, and the general alignment of the projector parts are such that the projector will pass a continuous film loop a minimum of 400 cycles without any indication of wear on the film. This film loop should indicate no need of repair at the completion of 800 cycles through the machine.

SOUND HEAD

The sound head, although a major unit, may well be classed as an accessory to the projector in that no adapters or machining operations are required to fit it to the projector.

It is composed of a main casting which contains the film handling mechanism and sound take-off devices, and a sheet metal phototube housing. The sound head is shown in Fig. 2. The main casting and film handling parts embody the same general features of construction as does the projector head, in that all parts resist corrosion and are designed to produce maximum film life.

The sound head derives its power from the picture head through a train of gears. The constant speed sprocket and the take-up sprocket are the driven members. Film enters the sound gate through a set of guide rollers by means of which the sound track is aligned with the image of the optical slit. The film is held snugly, though gently, against the gate shoe by means of adjustable tension springs on the gate. This particular feature is very important in handling film in the torrid temperatures of the tropics, where the wax protective coating on the film is very sensitive to abrasion and pressure and is very apt to pile up on the gate shoe. After leaving the sound gate the film wraps about the impedance roller and then passes to the constant speed sprocket and take-up sprocket in that order. The impedance roller is attached to a small flywheel, the inertia of which tends to keep it running at a constant speed, thus imparting a constant speed to the film. The take-up sprocket maintains a loop between itself and the constant speed sprocket to prevent changes in film speed introduced by the take-up mechanism from carrying through to the constant speed sprocket.

The exciter lamp is mounted in a socket which is removable from the projector and which permits prefocusing. It is possible to change exciter lamps with less than 30 seconds' interruption in a performance. The optical system is contained in a single barrel which is bolted to the

July, 1932]

PICTURE EQUIPMENT FOR U. S. NAVY

S79

center plate. It is focused by rotating a thumb nut on the objective. A spring lock prevents variation of the adjustment.

The shell of the phototube housing is of brass, heavily copper plated to increase its conductivity. The layer of copper on its surface materially increases its effectiveness as a shield against high frequency radiation. The phototube circuit is wholly contained within a metal compartment which is removable from the phototube housing as a unit by removing four bolts on the face of the housing. The UX-868 phototube (caesium cell), which is utilized in this equipment, is coupled to the amplifier system, all of which is below decks, by means of a low impedance transmission line and two line coupling trans-

Foale-r Switck

FIG. 2. The sound head.

formers. One of the transformers with its loading resistor, a protec- tive resistor, and suitable filter capacitors is cushioned in sponge rubber (to prevent microphonics) and mounted in a metal container. This container is placed in a second container and is sealed in wax to prevent the entrance of salt air and water. This sealed unit is mounted within a third metal compartment, previously referred to as being removable from the phototube housing, and is wired to the phototube socket, the fader switch, and the terminal board, all of which are a part of the removable compartment.

The fader switch is installed in the front face of this compartment, on the operating side of the projector, as indicated in Fig. 2. The

880 S. W. COCHRAN [j. S. M. P. E.

phototube socket is bolted to the upper face of the unit in a manner such that the UX-868 operates in the vertical position.

The circular opening in the upper face of the housing, through which the phototube is inserted and removed, is normally covered by a metal plate to which has been added a suitable pad of sponge rubber. A design of this nature reduces extraneous noise due to vibration of the phototube.

At the beginning of this paper, the necessity of successfully combat- ing the effects of radio frequency fields, vibration, salt water, and salt air was highly stressed. The descriptions of the projector and sound head have been presented with the sole object in mind of relating the problems to the solutions.

CONTROL PANEL

In direct contrast to the usual theater installation, the reproducing equipment on a battleship is of necessity widely scattered. It must be installed where there is room for it, and where it can be utilized. It is only logical to believe that the controls in an installation of that type should be centered at the vantage point, the projector.

For this purpose a control panel was designed to be mounted on the projector. Fig. 3 shows the control panel so mounted. It includes an ammeter for measuring exciter lamp current, a rheostat, and three fixed resistors for reducing the 115 volt, d-c. supply to 10 volts for the exciter lamp, a power switch for the exciter lamp circuit, and a push button station and indicator lamp for remote control of the motor generator.

It might be well to explain at this point that the 10 volt, 5 ampere exciter lamp utilized in the sound head is operated from the ship's power supply, thereby eliminating the need of batteries or low voltage generators. It is desirable to use the 5 ampere lamp for this applica- tion in preference to a higher voltage lamp because of the strength of its filament, its ability to withstand vibration, and its large light output.

The control panel is of heavy brass properly finished.

PROJECTOR DRIVE MOTOR AND CONTROL

Anticipating fluctuations in ship's supply voltages, the Navy has specified that the projector speed shall not vary more than plus or minus 1 per cent with fluctuations in line voltage of plus or minus 15 per cent, without any manual adjustment to take care of the line fluctuation.

July, 1932]

PICTURE EQUIPMENT FOR U. S. NAVY

881

The motor on the projector is a d-c. machine containing a centrifu- gal governor and control in a special end bell. The governor is very sensitive to slight changes in speed. It opens and closes a shunt on one end of several tapped resistors in the field circuit of the motor. The control contacts which the governor operates are spaced by rotating an adjustment on the end bell. By this means it is possible to select the constant motor speed desired. The flywheel on the motor shaft

It***.

MfcAt> V/JTH SIMPLEX PROJECTOR.

FIG. 3. The sound head and control panel mounted on Simplex projector.

stores enough kinetic energy to carry the system over the time interval required for the motor to respond to the action of the governor.

A small drum controller operates the motor. The switch mecha- nism of the controller is so arranged that the motor is braked dyna- mically by placing full excitation on the field and simultaneously short-circuiting the armature. The operating lever returns to the neutral position from the braking position as quickly as the hand pressure is relieved.

The assembly of the complete projector is shown in Fig. 3.

882

S. W. COCHRAN AMPLIFIER

[J. S. M. P. E.

The amplifier units, consisting of one voltage amplifier and two power amplifiers, are concentrated in one vertical rack together with a motor generator control panel. The complete unit is shown in Fig. 4.

The construction is particularly rugged and durable. The general structural features are to be noted in Fig. 5. The rack is made of channel iron supported by heavy angle iron feet. The lighter trans- former and capacitor units are supported in the horizontal position

Control Ktiob Voltage Atn|>li{i<st-

A.C, Volt m&ce-r -

FIG. 4. Front view of amplifier.

from vertical panels which are bolted to the rack. This construction is particularly apparent in the voltage amplifier. The transformer and capacitor packs on the power amplifiers are heavy and are there- fore mounted feet down on a heavy horizontal base. Each amplifier unit is a complete operating unit and is removable as such from the front of the rack. The dimensions of this rack are 72 by 24 inches, the maximum permitted by the Naval specifications. The lowest panel is 1 8 inches from the deck. The rack, panels, transformer and capacitor

July, 1932]

PICTURE EQUIPMENT FOR U. S. NAVY

883

cases, and all other parts that are liable to corrode are heavily cadmium- plated.

A front view of the interior of the rack is shown in Fig. 6.

The voltage amplifier is a three-stage unit consisting of one UY- 224 Radiotron resistance coupled to the second stage, a UY-227 Radio tron which is, in turn, transformer coupled to the third stage, and two UX-245 Radiotrons connected for push-pull operation. This amplifier is entirely a-c. operated, and contains its own power supply

FIG. 5. Rear view of amplifier.

including a power transformer, one UX-280 Radiotron rectifier, and a complete filter. The phototube also derives its polarizing voltage from this power supply. The volume control is a potentiometer between the first and second stages of the amplifier. It consists of a tapped switch, between the contacts of which are connected individual metallized resistors of a type approved by the Navy.

The volume control is operated by a handle protruding through the face panel or by a push button station located at a remote point.

884

S. W. COCHRAN

[J. S. M. P. E.

The push button control operates a capacitor motor installed in the voltage amplifier and geared to the volume control shaft. The capacitor motor is a two-phase, low-voltage induction motor which operates in conjunction with a capacitor which may be placed in series with either phase, depending upon the desired direction of rotation.

As a means of checking the operation of the amplifier under static conditions, a plate current metering jack is provided for each stage. The jacks are so shunted that all plate currents may be read accurately

Fcteee, Pla CirectA

lE£leciric»'l Inlet •'Switch. ,

FIG. 6.

Front view of amplifier showing the interior of the rack.

on one milliammeter provided on the amplifier control panel, to be described later.

Each important transformer in the voltage amplifier contains three varieties of shielding. The shielding prevents the transfer of undesir- able frequencies from one winding to another, and materially reduces the effect of powerful magnetic fields from within and of electrostatic fields from without. Supplementing the shielding of transformer units, the first two stages of the voltage amplifier, along with their

July, 1932]

PICTURE EQUIPMENT FOR U. S. NAVY

SS5

immediate accessories, are contained within a heavy metallic shield. Although the volume control is within this shield the capacitor drive motor is without. Shielding alone is not adequate when economically used in combating the effects of magnetic fields set up by power trans- formers and reactors within the amplifier. It is applied after the offending units have been so positioned and placed with respect to the audio units that the inductive effect is reduced to a minimum. Care- ful consideration of these points permits compactness of design, as exemplified in this amplifier, with an exceedingly low ratio of hum to audio power output.

FIG. 7. Limit curves showing specifications for amplifier frequency characteristics .

The metal tube cradles of the voltage amplifier rest on rubber cushions, and make no contact with metal at any point. The volume control is the only manual adjustment on the amplifier system. All other adjustments are automatic.

There are two power amplifiers on the rack, identical in design and removable as units. Each amplifier consists of one stage of two UV- 845 Radiotrons connected for push-pull operation. The power for these tubes is derived from the power supply system consisting of a filament transformer, a plate transformer, a filter, and two UX-866 Radiotrons (half -wave rectifiers). A tapped input transformer couples the power amplifiers to the low-impedance line which inter-

886 S. W. COCHRAN [j. S. M. P. E.

connects the voltage and power amplifiers. Normally the two power amplifiers operate with the full primary windings of their input trans- formers in parallel. However, if it becomes necessary to remove one power amplifier from the rack for any reason, proper operation of the remaining unit may continue by changing over to the tap on the input transformer. This maintains the correct line loading. It also permits the use of the same type of power amplifier in another Naval equipment of one-half the power rating. Each power amplifier normally feeds two reproducers. The output transformers are tapped to accommodate more or less than that number, however, making the amplifier versatile in its application.

FIG. 8. Motor-generator set.

The general features of construction described in the voltage amplifier apply to the power amplifier.

The UX-866 Radiotrons (mercury vapor rectifiers) require the application of cathode voltage thirty seconds in advance of the anode voltage. For this purpose an oil dash-pot time-delay relay has been incorporated in the design.

Inasmuch as each power amplifier supplies two of a total of four reproducers under normal operating conditions, it is important that some provision be made for the continued operation of the four re- producers in the event of the failure of either unit. This function is performed by one relay in each amplifier. These relays are energized with power from the rectifier circuits. When both relays are closed

July, 1932]

PICTURE EQUIPMENT FOR U. S. NAVY

887

(indicating normal operation) each power amplifier carries two re- producers. When only one relay is energized (indicating one power amplifier not operating) the four reproducers operate from the re- maining power amplifier. When the rectified current in either power amplifier is sufficiently low (indicating the failure of a unit in the rectifier circuit) one relay opens. When the load on the rectifier circuit is too great (indicating a flashover or tube failure) fuses in the a-c. supply circuits to the filament or plate transformer burn, causing one relay to open.

L-OtJO ePEAkJBR.

"BAFFLE

FIG. 9. Loud speaker with baffle.

Jacks are provided for monitoring the audio input and output to each power amplifier.

The tubes are set forward of the rack in order that the heat radiated may be dissipated in the surrounding air and be prevented from ascending between the rack channels.

The units are protected by heavy perforated metal cages. The removal of either of the cages releases an interlock which opens the a-c. supply circuit of the high potential power transformer, thereby eliminating the shock hazard. The total output of the rack is 80 watts of undistorted audio power.

The remaining panel on the amplifier rack is the control panel. Its

S. W. COCHRAN

[J. S. M. P. E.

contents include an a-c. voltmeter for measuring the a-c. voltage applied to the rack, a d-c. milliammeter and plug cord for measuring plate current in each amplifier stage, a push button station for motor generator starting, and motor and generator field rheostats. The field rheostats are mounted in tandem. They are controlled by concentric knobs appearing on the face of the panel. This construction is very economical of space. >

The push button station supplements similar controls on .the projector control panels. This is the only switch on the amplifier rack. It controls the power supply to the entire audio system. Depressing the start button causes the solenoid in the magnetic

FIG. 10. Frequency response curve of the unit.

starter to be energized, which, in turn, closes the d-c. circuit of the time-delay relays in the power amplifiers and places the motor- generator in operation. By this means power is applied to all the units concerned in the proper sequence. This eliminates the possi- bility of operating the power amplifiers without having the time-delay relays functioning as intended. Fuses in the rear face of the panel protect the remote control lines. (. *fc"j. V* I

In concluding this description of the amplifier, it is fitting that its performance be adequately treated. The limit curves shown in Fig. 7, established by the Naval specifications, best attest to the quality of amplifier performance.

July, 1932]

PICTURE EQUIPMENT FOR U. S. NAVY

889

Returning to the original problems for a moment, bear in mind the thought that the amplifier rack and its components were designed to bear up and perform in accordance with Naval specifications under the severest onslaughts of salt water, humidity, vibration, concussion, and radio frequency radiation. The need of an efficient system of remote control and of equipment adequately protected against possible injury from overload was duly emphasized at the beginning of this paper. These points are to be associated with the actual amplifier design, involving special shielding features, shock mountings, the application of protective finishes, automatic switching, remote control, and protective features as described above.

TO «0 10 KX> 110 130 130 MO ISO \6O PD DISTANCE IN FEET FBOM SCREEN

FIG. 11. Sound pressure distribution limits at 1000 cycles. MOTOR-GENERATOR

It is the purpose of the motor-generator to supply all the a-c. power required by the equipment. The motor is a 3 hp., 115 volt, d-c. compound wound unit, and the generator is a 110 volt, 60 cycle, iy2 kva. unit with a special series field which carries the motor line current. Motor speed and generator voltage are controlled by the field rheostats previously described as part of the amplifier control panel. This motor-generator is shown in Fig. 8.

A magnetic starter controlled by push button stations on the amplifier and projector control panels performs the starting function.

REPRODUCERS

Four directive baffles of the type indicated in Fig. 9 are supplied with this equipment. The entire baffle is of metal, heavily cadmium-

890 S. W. COCHRAN

plated. The driver is a cone dynamic unit sealed, where possible, with wax. The cone is protected by two wire mesh screens and two cloth screens placed across the throat of the baffle. The driver is bolted to a heavy metal plate which is the cover to the rear housing of the baffle. This rear housing (with driver) is detachable as a unit from the baffle by removing four wing nuts.

The maximum dimensions of the unit do not exceed the 26 by 48 inch limits set by the Naval specifications. It is easily carried by two men as required. The frequency response curve of this unit falls within the limits indicated on Fig. 10, as specified by the Navy.

To obtain a sound pressure distribution curve that will fall within the limits indicated on Fig 11, it was necessary to mount the four baffles in two tiers of two baffles each, with the highest point of the upper tier at a maximum height of twenty feet above the level of the test area.

Upon these units falls the burden of reproducing and distributing high quality sound under many atmospheric conditions and over a

wide area.

SCREEN

The problem involved in the design of a screen was that of securing a material that was 90 per cent efficient in the transmission of sound at a frequency of 1000 cycles, yet capable of standing a grab test of 250 pounds in the filling, stand washing in salt water, and resist the effects of fuel oil soot of high sulfur content. Heavy duck would meet these requirements with the exception that it will not transmit sound at the required efficiency. Neither does it perforate with a clean cut.

The final product is a combination of a durable cloth that is easily perforated and a checker network backing composed of ribs of several ply duck.

The ribs of duck have the required strength to withstand the grab test. There is sufficient perforated material exposed in the open spaces of the network to permit the required transmission of sound.

THE DUPLICATION OF MOTION PICTURE NEGATIVES* J. I. CRABTREE AND C. H. SCHWINGEL**

Summary.— In 1926, Capstaff and Seymour1 published a paper giving directions for making duplicate negatives using a new film manufactured specifically for that purpose. Good quality and tone reproduction were possible by this method although the graininess of exhibition prints was not entirely satisfactory.

Since then, improved films have been available, and the present paper contains a description of the tests performed during a search for the most satisfactory sensitive materials and processing technic. The experiments showed that in order to minimize graininess, the master positive must be developed to a relatively high gamma (1.85) in a positive developer, and the duplicate negative to a correspondingly low gamma (0.55 or less}. There are also given data which explain why the high gamma master positive in conjunction with a low gamma duplicate negative gives the most satis- factory graininess.

I. INTRODUCTION

The unsatisfactory quality of many of the duplicate motion picture negatives produced to date has been due partly to the use of un- suitable film emulsions for their preparation, and partly to a lack of understanding of the conditions under which good duplicates can be made.

A good duplicate negative is one capable of giving a print which is a facsimile of a print made from the original negative. It should reproduce accurately the tones of the original negative unless the contrast range has been modified to correct some fault. Also, the definition and the graininess of the duplicate should be of the same order as those of the original.

Heretofore, no extensive survey has been available for the pur- pose of choosing the best photographic material and processing tech- nic for the production of good duplicates. In an article entitled "The Duplication of Motion Picture Negatives"1 by J. G. Capstaff and M. W. Seymour, the use of Eastman duplicating film (emulsion series No. 1503) was advised for both master positive and duplicate

* Presented at the Spring, 1932, Meeting at Washington, D. C. Communica- tion No. 488 from the Kodak Research Laboratories.

** Research Laboratories, Eastman Kodak Co., Rochester, N. Y.

891

892 J. I. CRABTREE AND C. H. SCHWINGEL [j. S. M. P. E.

negative, both of these being developed to a gamma of unity. Fol- lowing this, C. E. Ives and E. Huse2 published additional notes and gave precautions necessary for the production of good quality duplicate motion picture negatives utilizing the above procedure.

The results obtained were very satisfactory with regard to quality and tone reproduction, but the graininess of the exhibition prints left much to be desired.

It was stated by some laboratories that better results with re- spect to graininess were obtained by making the master positive on motion picture positive film with lavender support developed to a high gamma, and the duplicate negative on par-speed negative film developed to a correspondingly low gamma.

Early in 1929 two new films were put on the market for use in duplicating. They were Eastman duplicating positive film with lavender support (emulsion series 1355) and Eastman duplicating negative film (emulsion series 1505). These films were recommended as being more suitable for duplicating purposes than the 1503-1503 combination. Emulsion 1505 was faster than 1503, and with emul- sion 1355 it was possible to print under the same printer conditions as with motion picture positive film.

In an article entitled "The Graininess of Motion Picture Film,"3 one of the authors showed that the graininess of a positive print increased with increasing density of the negative from which it was printed. He also pointed out that the graininess of a positive printed from a negative developed to a high gamma was greater than that printed from a negative of the same density contrast, but which was developed to a lower gamma.

With these facts under consideration and in order to determine the most suitable sensitive material and processing technic for du- plicating purposes, the present investigation was undertaken.

II. METHOD EMPLOYED FOR DETERMINING A SATISFACTORY DUPLICATING PROCESS

More than 200 separate tests were required for the purpose of finding the most suitable procedure to follow in the making of good duplicate motion picture negatives. Each test consisted of the printing of a master positive from the original negative and then printing a duplicate negative from this. In all cases, the duplicate negatives were made with a printing contrast equal to that of the original negative. The contrast of the duplicate negative was

July, 1932] DUPLICATION OF NEGATIVES 893

judged by the degree of development required to produce a print of given contrast. Each test embodied some change in the proc- ess, either in the sensitive materials, the developers, or the process- ing technic employed.

Matched exhibition prints on Eastman motion picture positive film were prepared from all original and duplicate negatives and these were compared on projection, the measure of graininess being taken as the ratio of the distance of the observer from the screen to the vertical dimension of the screen. By distance from screen is meant the minimum distance at which graininess was not apparent to the eye, with the line of sight normal to the screen at its lower edge.

It is true that the distance at which graininess was zero varied with the observer, but comparisons have shown that the variation was not more than 10 per cent among individuals having normal eye- sight. A greater accuracy can not be expected and this method of comparison seemed satisfactory from the standpoint that quality and graininess were compared under practical projection conditions. The screen size was 43 by 57 inches, with a screen brightness between 10 and 11 apparent foot-candles. The projector was operated at a speed well above that at which flicker was noticeable.

The standards used for comparison were prints made from the original negatives, and new standards were made for every series and at all other times when the processing needed to be checked.

Gammas reported in this work were determined from sensitometric strips made on the film under discussion and exposed in the printer. These sensitometric strips were processed along with the master positive or duplicate negatives. All densities were measured on a Capstaff densitometer.4

(A) Sensitive Material. — An emulsion suitable for duplicating must fulfill the following conditions: (a) The emulsion must have sufficient latitude to permit the correct reproduction of the greatest scale of tones likely to be met with in the original negative; (b) it must have the ability to reproduce fine detail, otherwise a serious loss in picture definition will occur; (c) the developed image must have a minimum graininess, otherwise the cumulative effect in mak- ing the master positive and duplicate negative will produce excessive graininess in the exhibition print; (d) it must have sufficient speed to permit printing the master positive or duplicate negative with- out extensive modifications of the printer optical system.

894 J. I. CRABTREE AND C. H. SCHWINGEL [j. s. M. p. E.

The emulsions listed below were those among the eighteen dif- ferent emulsions tested which showed the most promise of being suitable for duplicating purposes. Only the data relative to these will be given and discussed in this article.

Emulsion

Series Number Film

1302 Eastman Motion Picture Positive Film, Lavender Support

1355 Eastman Duplicating Positive Film, Lavender Support

1503 Eastman Duplicating Negative Film, Yellow-dyed

1505 Eastman Duplicating Negative Film, Yellow-dyed

1201 Eastman Negative Film

(B) Original Negatives. — The original negatives chosen for this work were of average density contrast. Since graininess is most apparent in the lighter tones, such as in the face and in other uni- form areas of low density, the scenes selected were principally close- ups.

(C) Type of Printer Employed. — The master positives, dupli- cate negatives, and exhibition prints were printed on a Bell & Howell continuous printer which had been tested and approved as giving good definition. It was found to be very important that the film be in uniform contact over the entire area of the printer aperture, otherwise a patchy, uneven image was produced. To test the printer for uniformity of contact, a print was made from a strip of evenly fogged and developed negative film. Perfect adjustment of the printer was indicated by an even density in the print, while imperfect adjustment gave the patchy unevenness referred to above. The printer was also tested for steady operation by exposing a length of positive film in the printer, without a negative, and then examining it for unevenness after development.

(D) Processing Methods. — Several methods of processing the films were used, including (a) rack and tank, and (b) continuous machine. There appeared to be no difference in the graininess of the images obtained by the different methods, although machine processing gave the most uniform results.

Too much emphasis can not be placed upon the necessity for good development technic. It must be remembered that the produc- tion of a final print in a duplicating process requires four distinctly separate development operations and, since all defects are cumulative, these defects will be greatly magnified in the exhibition print. It

July, 1932]

DUPLICATION OF NEGATIVES

895

is the accumulation of the small defects and errors which gives the "duped" appearance to prints made from unsatisfactory duplicate motion picture negatives.

(E) Developers. — In addition to such standard developers as formulas D-16 and .D-76,6 a large number of developers of special composition also were tried.

Only the experimental results obtained from the use of two of these special developers will be given. The formulas will be known as Special Developers No. 1 and No. 2. Developer No. 1 contained potassium iodide, and Developer No. 2 was one capable of giving high contrast.

III. DISCUSSION OF RESULTS

Table I contains the data relating to the emulsions tabulated above. All the duplicate negatives listed were made so as to have

TABLE I

(Print from Original Negative — Graininess Ratio = 5.4}

Emul- sion Num-	Mini- mum Den-	Emul- sion Num-	Devel-		Mini- mum Den-	Graini
No.	her	Developer	Gamma	sity	ber	oper	Gamma	sity	Quality	ness
1	1503	D-76	Approx.	0	.45	1503	D-76	Approx.	0.40	Good	7.5
			Unity					Unity
2	1503	D-76	Approx.	0	40	1503	D-76	Approx.	0.62	Good	7.8
			Unity					Unity
3	1505	D-76	Approx.	o	.20	1505	D-76	Approx.	0.42	Fair	8.1
			Unity					Unity
4	1505	D-76	Approx.	0	.50	1505	D-76	Approx.	0.45	Good	8.4
			Unity					Unity
5	1355	D-16	1.52	o	30	1505	D-76	0.69	0.50	Good	6.5
6	1355	D-16	1.52	o	30	1503	D-76	0.69	0.50	Good	6.7
7	1355	D-16	1.85	0	30	1505	D-76	0.55	0.25	Very Good	6.3
8	1355	D-16	1.98	0	40	1505	D-76	0.52	0.25	Fair	6.3
9	1355	Special	1.61	0	,28	1505	D-76	0.62	0.40	Very Good	6.7
		Dev. No. 1
10	1355	Special	1.62	0	51	1505	D-76	0.61	0.37	Good	7.0
		Dev. No. 1
11	1355	D-16	1.60	0	50	1505	D-76	0.60	0.45	Good	6.7
12	1355	D-76	1.05	0	51	1505	D-76	1.00	0.50	Good	8.3
13	1355	Special	2.02	0	21	1505	D-76	0.50	0.42	Good	7.5
		Dev. No. 2
14	1355	Special	2.04	0	50	1505	D-76	0.50	0.40	Very Good	7.2
		Dev. No. 2
15	1302	D-76	2.00	0	50	1201	D-76	0.50	0.45	Fair	7.9
16	1302	D-76	2.00	0	50	1505	D-76	0.50	0.42	Fair	8.1
17	1302	D-76	2.62	0	30	1201	D-76	0.38	0.30	Fair	6.5

the same printing contrast as that of the original. The contrast

of the duplicate negative was judged by the degree of development

required to produce a given contrast in the positive printed from it.

(A) Results Obtained with Various Emulsions in Which Both

896 J. I. CRABTREE AND C. H. SCHWINGEL [J. S. M. P. E.

Master Positive and Duplicate Negative Were Developed to a Gamma of Unity*. — Tests Nos. 1 to 4, inclusive, and No. 12 (Table I) were made for the purpose of determining the effect of varying the type of duplicating emulsion on graininess and quality. All master positives and duplicate negatives in these tests were developed in D-76 developer to equal degrees of contrast (gamma 1.0).

From the data it will be seen that emulsion 1503 produced images with less graininess than the faster duplicating emulsions 1505 and 1355. The quality was good in all cases except that of test No. 3, where the loss occurred in the master positive. The reason for the poor quality in the master positive was that the low densities were printed too low on the density scale and, therefore, the high- light densities were printed in the region of underexposure. It was found necessary with emulsions 1503 and 1505 to print to a minimum density greater than 0.4 for a gamma of 1.0 in order to insure good tone reproduction. With emulsion 1355, lower mini- mum densities were permissible. Comparison of the results from tests Nos. 1 and 2 showed that an increase in density in the duplicate negative caused a very slight increase in graininess. Likewise, from the results of tests Nos. 3 and 4, it was seen that an increase in density of the master positive caused the same effect. Results from tests Nos. 9 and 10 also confirmed this observation. This fact is not new and is in agreement with the previous findings of one of the authors.3

(B) Results Obtained Using the 1355-1505 Process— In tests Nos. 5 to 14, inclusive, Eastman duplicating positive film, emulsion series No. 1355, was used for the master positives and the yellow- dyed duplicating negative film for the duplicate negatives. With the exception of test No. 6, Eastman duplicating negative fast, emulsion series 1505, was used for the duplicate negatives. Tests Nos. 5 and 6 were for comparing the merits of emulsions 1503 and 1505, when used for the duplicate negative and developed to a low gamma. There appeared to be practically no difference in the graininess or quality when comparisons were made between the exhibition prints.

Test No. 8 shows the results obtained when the master positive was developed to a very high gamma in developer D-16. The du- plicate negative, which had been printed to the lowest minimum

* For explanation of emulsion characteristic curves with regard to exposure, latitude, tone rendering, gamma, etc., see references 6 and 7.

July, 1932] DUPLICATION OF NEGATIVES 897

density in the region of correct exposure, was developed to a corre- spondingly low gamma. The result was a very low graininess ratio (6.3), when compared with the results of the other tests. How- ever, the quality was not as good as with the lower gamma master positive due to the necessity of printing the high master positive densities in the region of overexposure. By lowering the gamma slightly (1.85), the quality was improved. This was illustrated by the results obtained from test No. 7 in which there appeared to be no increase in graininess over that in test No. 8.

Tests Nos. 9 and 10 were for the purpose of determining whether or not a special contrast iodide developer would permit the use of lower minimum densities for the master positive. It apparently had no advantage in that respect and was objectionable because the graininess was slightly worse than in the case where D-IQ was used (tests Nos. 10 and 11).

Tests Nos. 13 and 14 were made using a special high contrast developer. Results showed that the quality was good, but the graininess was worse than in those tests where D-16 was used for de- veloping the master positive.

(C) Comparison between 1355-1505 and 1302-1201 Processes — Tests Nos. 15 and 16 were made for the purpose of comparing the results obtained from the 1355-1505 and 1302-1201 processes; also for determining the effect of using emulsion 1505 for the duplicate negative in the 1302-1201 process. The results showed that for equal gammas the graininess was appreciably better when the 1355- 1505 process was employed. No advantage was gained by using emulsion 1505 instead of emulsion 1201 in the 1302-1201 process.

Test No. 17 showed that graininess results comparable with those obtained in test No. 7 were obtained when 1302 was developed to a very high gamma (2.62). The quality was poor and develop- ment defects were noticeable, caused by the very low degree of de- velopment necessary for the negative.

IV. FACTORS AFFECTING GRAININESS DURING EXPOSURE AND DEVELOPMENT

In the foregoing tests, the graininess of the original negative ex- erted a pronounced effect on the graininess of the duplicate nega- tives and prints. In order to eliminate this effect and to determine the influence of exposure and development on graininess, the follow- ing tests were made.

898 J. I. CRABTREE AND C. H. SCHWINGEL [j. s. M. p. E.

Lengths of the various films tested in Table I were given uniform flash exposures and developed by the rack and tank method. The degrees of development were determined from step tablet readings from exposed strips developed on the racks with the flashed films. The negative types of material were developed in the borax de- veloper D-76, and the positive material in the positive developer, formula ZM6.

It has been found throughout this investigation that, whenever a negative emulsion was developed in a positive type of developer, graininess was greater than when the material was developed to an equal degree in the borax developer. The borax developer is not suitable for positive development, however, because of its inability to produce the necessary high gamma.

The graininess of the developed flashed strips was judged by the method described for all preceding tests. In these observations the assumption was made that the graininess of the photographic material was proportional to the distance from the eye of the ob- server to the screen. Since the visual acuity of the observer was subject to variations due to such factors as adaptation level, fatigue, and general physiological conditions, allowances were made for these whenever measurements were made.

Before making measurements, the person chosen for the viewing was allowed to remain for a length of time in a room which had an illumination level approximating that encountered when viewing the screen. This preliminary precaution was necessary in order to fix the adaptation level of the observer and minimize errors arising from variations of this. Numerous check determinations were made, and in no case were values found deviating more than 10 per cent.

Each screen test consisted of the projection of not more than 225 feet of film to be viewed, after which the observer was allowed to rest for a period of ten to twenty minutes before continuing. In this way, errors arising from eye fatigue were minimized.

The values for graininess reported were the result of a large num- ber of observations made by three observers of normal vision, which were averaged when drawing the curves.

(A) Variation of Graininess of a Constant Density with Degree of Development. — Lengths of the flashed film were developed for varying degrees to give gammas covering the useful range for each of the emulsions used. The exposures were varied to give a den- sity of 0.8 in every case. The graininess ratios determined by pro-

July, 1932] DUPLICATION OF NEGATIVES S99

jection were plotted against the gammas to give the curves shown in Fig. 1. The results show that for negative emulsions the grain i- ness increased very rapidly with increase in the degree of develop- ment, while for the positive types of film, the graininess rapidly reached a maximum and then remained practically constant, or even decreased slightly with increasing degrees of development, over the useful range of the material.

These results seem to show why the method using a high gamma master positive and a low gamma duplicate negative, which were printed on positive and negative emulsions, respectively, gave less

*?l"rf)

6

THC VAtRIATTlON OF"

WITH ALL. OdlH^lTlE:?> O.fc.

o. -

A. i«

Q. " NtOTlON P\CTURt POSITIVE.

•

2

.4

o

a

z I

o

O.S I.Q l.% X.O

FIG. 1. The variation of graininess with gamma — all densities 0.8.

graininess than the earlier recommended method in which both master positive and duplicate negative were printed on a negative emulsion and developed to a gamma of unity.

These experiments also confirm those of Crabtree and Carlton6 who predicted that "the graininess-gamma curve for a negative material over the useful range of gamma (0.5 to 1.0) is probably straight and rather steep, while the graininess-gamma curve for the positive (gammas 1.2 to 2.2) has a long shoulder which must be almost parallel to the gamma axis."

For this discussion it can be assumed that the graininess of a print is the additive result of the inherent graininess of the master posi-

900

J. I. CRABTREE AND C. H. SCHWINGEL [j. s. M. p. E.

tive and duplicate negative materials, although actually it appears to be somewhat less than this total. The graininess ratio of a flashed length of duplicating negative film (emulsion series 1505), developed to a gamma of unity, was approximately 5.5 units, and a print from this on the same material and developed to a gamma of unity on this assumption, would therefore have a graininess approximately double this, or 1 1 units. Considering a second example, where the master positive was printed on duplicating positive film (emulsion series 1355), and developed to a gamma of 1.85 when the graininess ratio

THE: VARIATION OF-

OF PRINTS WITH DENS\TV OF" NEIOATT VVE1S.

PRINTS OELVElUOP>e.O TO

o-ifc oeve.uopE.R ANO TO *K

OF O.B.

X. PRINT F-ROf\ O. " "

OF" 2..Z.

FIG. 2. The variation of graininess of prints with density of negatives. Prints developed to a gamma of 2.2 in D-1Q developer and to a uniform density of 0.8.

was approximately 3.8 units; and then printing the duplicate nega- tive on duplicating negative film (emulsion series 1505), developed to a gamma of 0.55, which furnished additional units of 3.8 (see curve in Fig. 1), it is seen that the duplicate negative would have a graininess ratio approximating 7.6 units. It is apparent, there- fore, that the graininess should be much less in the case of a dupli- cate negative prepared by the latter method than one prepared by the former method.

In Fig. 1 the graininess curve for duplicating positive film is con- trasted with that of motion picture positive film for the purpose

July, 1932] DUPLICATION OF NEGATIVES 901

of showing that the graininess ratio is lower for the duplicating ma- terial at high gammas.

(B) Variation of Graininess with Density. — Lengths of film were given varying flash exposures, and developed to the gammas recom- mended. It was considered that the graininess could not be judged correctly from these because of the varying screen brightnesses, so prints were made on motion picture positive film from the various densities, and exposed so as to give a density of 0.8 with equal de- grees of development.

The graininess ratios of these films were determined and plotted against the negative densities to give the curves of Fig. 2, which show that for the master positive the densities should be as low as possible on the density scale, while in the case of the duplicate nega- tive printed on negative materials, the graininess increases only slightly with increase of density of the negative. It is also seen that the graininess of Eastman supersensitive panchromatic film is greater than that of Eastman duplicating negative film, but the curves run parallel to one another.

V. FACTORS AFFECTING GRAININESS DURING PRINTING

(.4) The Effect of Loss in Definition on Graininess. — Tests were made to determine the effect on graininess of imperfect negative- positive contact in printing. These were accomplished in two ways: (a) by adjusting the printer gate so as to permit the nega- tive to be out of contact with the positive stock during exposure, and (b) by printing through a thickness of Kodaloid.

The results were similar in both cases, and showed that whenever a loss in picture definition occurred there was also a diminution in graininess. The slight loss in picture definition was not objectionable in certain types of prints, particularly with close-ups where fine de- tail was not essential.

(B) Effect on Graininess of Printing with Diffuse Light. — The gate of a motion picture step printer was fitted with a piece of pot opal glass in such a manner as to insure perfect contact between the glass and the negative during the printing operation. This arrangement permitted the printing to be carried out with diffuse light. Duplicate negatives and prints from these and from original negatives showed no appreciable difference in graininess, although printing with diffuse light slightly impaired the picture definition.

902

J. I. CRABTREE AND C. H. SCHWINGEL [J. S. M. P. E.

FIG. 3. Characteristic curves for Eastman duplicating negative mo- tion picture film, series 1505. Developed in D-76 at 65 °F. by rack and tank method.

FIG. 4. Characteristic curves for Eastman duplicating positive mo- tion picture film, series 1355. Developed in ZM6 at 65 °F. by rack and tank method.

July, 1932]

DUPLICATION OF NEGATIVES VI. TONE REPRODUCTION

903

The photographic characteristics of emulsion series 1505 are shown in Fig. 3, from which it is seen that at the low gamma required for duplicate negatives (0.5-0.6), it was possible to print to a minimum density of 0.3 and still retain all the negative densities on the straight- line portion of the characteristic curve, which is the requirement for correct tone reproduction.7'8 For higher gammas it was necessary to increase the minimum density values. A development time- gamma curve is also given in Fig. 3 (a).

Emulsion series 1355 (Eastman duplicating positive film) was

3.0

z.e

Z.fe 24

RCLPROOUCTtOtH CURVE A

A. THeO*eTtC*M_t_V

B.O— ef&4«-l&O6 REPRODUCTION. C.»r — KVCRV POOR. "

or PRINT FROM

FIG. 5. Curves showing the degree of perfection attained in the dupli- cation of negatives.

found to be the most suitable material for use in the making of the master positive. Its latitude permitted the use of the high gamma (1.85) without impairment of the tone reproduction. Owing to the great density range to be covered in the master positive for best gamma conditions, no other material was found which was entirely suitable in this respect. Fig. 4 gives the characteristics of emulsion series 1355, and it is seen from the curves that at a gamma of 1.85, it is necessary to print at a minimum density of not less than 0.40, otherwise a loss in highlight quality will occur. Fig. 4 (a) gives the development time-gamma relationship for this emulsion.

904 J. I. CRABTREE AND C. H. SCHWINGEL [j. s. M. P. E.

(A) Reproduction Curve for the 1355-1505 Process. — Fig. 5 is a tone reproduction curve for the 1355-1505 process, the densities of a print from the duplicate negative being plotted against the cor- responding densities of the print from the original negative. The negatives and prints were perfectly matched and the prints from which the densities were taken received identical development.

It is obvious that with this method of representation, perfect tone reproduction is represented by a straight line at 45 degrees to the axis and commencing at the origin (Curve A). Curve B represents the tone reproduction with the 1355-1505 process, and it will be seen that when this curve is compared with Curve A, the process gives almost perfect tone reproduction, but only if care is taken not to print too low on the density scale when exposing the master positive and duplicate negative. Curve C (Fig. 5) shows the distortion that resulted when these precautions were not taken.

VII. THE DUPLICATION OF SOUND NEGATIVES

Frequency records with three modulation levels and frequencies varying from 100 to 6000 cycles were duplicated. Listening tests indicated that the upper frequency limit for duplicate negatives was approximately 6000 cycles, which frequency was only discernible at high and medium modulation levels. Tests also showed that a slight increase in ground noise occurred which became objectionable only in the frequency range from 5000 to 6000 cycles. Prints from duplicate negatives of piano records and vocal selections, for prac- tical purposes, were indistinguishable from the original prints.

VIII. PRACTICAL INSTRUCTIONS FOR MAKING DUPLICATE NEGATIVES

Duplicate motion picture negatives may be made in the usual way without departing from ordinary methods of exposure and de- velopment. The quality of these duplicates will be better than could be obtained on other existing emulsions using similar technic.

(A) The Master Positive. — The master positive should be made on Eastman duplicating positive film, emulsion series 1355. The speed of this material is approximately the same as that of ordinary motion picture positive film. A lavender support serves the purpose of reducing halation effects and also for identification. The emulsion is capable of giving very fine grained images with good contrast on full development. Its latitude is such as to insure correct repro- duction for the greatest range of tones likely to be met in an

July, 1932] DUPLICATION OF NEGATIVES 905

original negative. Also, it has the ability to reproduce the detail registered on the original.

(B) Printing the Master Positive. — When printing the master positive, the first requirement is that sufficient exposure must be given so that all the tones and fine detail of the original are recorded faithfully. A good master positive appears denser than the average projection positive, and to the inexperienced eye seems to be overprinted. The least dense portions ("catch-lights") of any master positive should have a measurable density — a graying-over of the highlight areas — otherwise the reproduction of tones in the final print will be unsatisfactory, giving the print a "duped" appearance. Timing of the master positive should be for the highlights, allowing the shadows to take care of themselves. Allowance must be made for the contribu- tion of the lavender support to the density when judging exposure.

(C) Development of the Master Positive. — The master positive should be developed in any good positive film developer such as formula D-IQ. Developers of the borax type (D-76) are not capable of giving sufficient contrast.

The degree of development must be such that the contrast of the master positive is equal to or greater than that of a normal exhibition print (gamma approximately 1.8, for average good quality originals). This high contrast of the master positive permits the desired low de- gree of development of the duplicate negative which insures a mini- mum graininess of the image in the exhibition print. Using rack and tank methods, with fresh motion picture D-IQ developer at a temperature of 65 °F., development time for the master positive will be from 9 to 11 minutes, depending upon the manipulation technic. With continuous machines the time of development differs widely with their design. The fixing, washing, and drying processes are identical for master positives and motion picture positive films.

(D) The Duplicate Negative. — The duplicate negative should be made on duplicating negative film, emulsion series 1505. This film has sufficient printer speed so that enough exposure can be ob- tained through the dense master positive without changing printer lamps. The yellow dye in the emulsion, which absorbs the wave- lengths of light to which the emulsion is most sensitive, reduces irradiation or scattering of light and, therefore, insures good defini- tion ; greatly extends the latitude (ability to reproduce a wide range of tones correctly); and lowers the contrast of the emulsion. Du- plicating negative film must be handled in the darkroom with the

906 J. I. CRABTREE AND C. H. SCHWINGEL [J. S. M. P. E.

same precautions as ordinary par-speed negative film, using the Wratten Series 2 safelight.

(E) Printing the Duplicate Negative. — As with the master posi- tive, sufficient exposure must be given to the duplicate negative so that every tone and detail of the master positive will be faith- fully reproduced. Good duplicate negatives have no clear shadows, even when they are present in the original negative. The shadows of the duplicate are always somewhat gray, and while those in some scenes may be more dense than others, depending upon the range of brightness in the subject, none of them should be glassy clear. Lack of exposure in printing the duplicate negative produces a lack of quality in the shadows of the exhibition print.

(F) Development of the Duplicate Negative. — The duplicate nega- tive should be developed in the borax developer, formula D-7Q. The degree of development should be such as to reproduce the con- trast of the original negative. If the master positive be fully de- veloped, the average time of development for the negative when employing rack and tank methods will be approximately 7 to 8 minutes in fresh D-76 developer at 65°F. This degree of develop- ment corresponds to a gamma of approximately 0.5-0.6.

Original negatives may be divided into three general classes, ac- cording as they are normal, flat, or contrasty. In the process of leveling up the different scenes for the duplicate negative, so that the final printing operation can be carried out at a single light setting and that all the prints will require equal times of development, the first rigid requirement is to keep the degree of development of the duplicate negative at or below the specified degree (gamma 0.55 or less), for this insures minimum graininess in the exhibition prints. When it is desired to change the contrast of the duplicate negative from that of the original, it is best accomplished by chang- ing the contrast of the scenes in the master positive by varying the times of development. In general, studio negatives are quite uni- form in quality so that usually enough variation can be obtained through slight changes in the master positive.

The duplicate negative should be fixed for 20 minutes in a properly compounded acid fixing bath. If the film is fixed in an incorrectly compounded bath or is allowed to fix for too long a period, the rate of washing out of the dye will be retarded. After fixing, the film must be washed for 45 minutes. If the washing is incomplete or uneven, the dye which remains in the film will cause the print from

fuly, 1932] DUPLICATION OF NEGATIVES 907

the negative to be mottled and uneven. Rinsing for several minutes water, before fixing, will greatly accelerate the removal of the dye. (G) Printing Precaution. — It is very important when making iplicates to clean the original negative and master positive, be- cause defects are cumulative. Dirt on the original negative or master positive will show up objectionably in the final print. After the original negative and master positive have been timed, they should be carefully cleaned to remove all traces of dirt.9

IX. SUMMARY

(1) As a result of tests with a large number of combinations of emulsions, the best duplicate negatives, with respect to both tone reproduction and graininess, were obtained with Eastman du- plicating positive film, emulsion series 1355, and Eastman dupli- cating negative film, emulsion series 1505.

(2) In order to maintain the graininess at a minimum, it was necessary to develop the duplicate negative to a low gamma (0.55) and the master positive to a correspondingly high gamma (1.85). Since the graininess of a positive image of given density increases with increase of density of the negative from which it is printed, it is necessary to maintain the minimum density of both the master positive and duplicate negative as low as is consistent with good tone reproduction.

(3) It has been found that the graininess-gamma curve of the negative emulsion 1505 is straight and rather steep, but that of the duplicating positive emulsion 1355 has a long shoulder which runs almost parallel to the gamma axis. This explains why it is de- sirable to develop the master positive to a relatively high gamma so as to insure a low gamma and, therefore, minimum graininess for the duplicate negative. Although increasing the gamma of the master positive to a value greater than 1.85 would result in slightly improved graininess, the latitude of this material is diminished so that with fairly contrasty negatives it is not possible to obtain perfect tone reproduction.

(4) A fine grain borax type of developer was found to be most suitable for the development of the duplicate negatives, while any good positive developer was satisfactory for the development of the master positives. The low degree of development required for the duplicate negative tended to introduce certain development defects

908 J. I. CRABTREE AND C. H. SCHWINGEL

which were at a minimum in the case of duplicating negative film, emulsion series 1505.

(5) Poor contact between the negative and positive during printing improved the graininess, but at the expense of loss in picture definition. The use of diffuse light caused a slight loss in picture definition without an apparent effect on graininess.

(6) Duplicating positive film, emulsion series 1355, has great latitude, so that when the master positive was printed on this ma- terial and developed to a gamma of 1.85 and the duplicate negative printed on duplicating negative film (emulsion series 1505) and de- veloped to a gamma of 0.55, a practically perfect reproduction of the tones of the original negative was secured. It was necessary, however, to print the minimum densities of the master positive at a density of at least 0.40* and of the duplicate negative at a density of at least 0.3 to avoid distortion due to underexposure.

(7) Prints from duplicate negatives of sound records were prac- tically indistinguishable, by listening tests, from the original prints, although tests with frequency records indicated an increase in ground noise and some losses at the frequencies of 5000 and above.

The authors are indebted to Messrs. H. A. Doell, A. J. Miller, and H. Parker, of this laboratory, for their assistance.

REFERENCES

1 CAPSTAFF, J. G., AND SEYMOUR, M. W.: "The Duplication of Motion Picture Negatives," Trans. Soc. Mot. Pict. Eng. (1926), No. 28, p. 223.

2 IVES, C. E., AND HUSE, E.i "Notes on Making Duplicate Negatives," Trans. Soc. Mot. Pict. Eng. (1928), No. 34, p. 382.

3 CRABTREE, J. I.: "The Graininess of Motion Picture Film," Trans. Soc. Mot. Pict. Eng. (1927), No. 24, p. 77.

4 CAPSTAFF, J. G., AND PURDY, R. A.: "A Compact Motion Picture Densitom- eter," Trans. Soc. Mot. Pict. Eng. (1927), No. 31, p. 607.

6 CRABTREE, J. I., AND IVES, C. E.: "A Replenishing Solution for a Motion Picture Positive Film Developer," /. Soc. Mot. Pict. Eng., 15 (Nov., 1930), p. 627.

6 CRABTREE, J. I., AND CARLTON, H. C.: "Some Properties of Fine Grain Devel- opers for Motion Picture Film," Trans. Soc. Mot. Pict. Eng. (1929), No. 38, p. 406.

7 MEES, C. E. K.: "Fundamentals of Photography," Eastman Kodak Co., Rochester, N. Y. (1931), Chapter VIII.

8 NEBLETTE, C. B.: "Photography — Principles and Practice," D. Van Nos- trand Co., Inc., New York, N. Y. (1930), Chapter IX.

9 CARLTON, H. C., AND CRABTREE, J. I.: "Cleaning Liquids for Motion Picture Film," Trans. Soc. Mot. Pict. Eng. (1927), No. 30, p. 277.

* This value is that of the actual silver deposit. To obtain the total film density, the support density of 0.18 must be added to it.

THE SCREEN— A PROJECTIONIST'S PROBLEM* FRANCIS M. FALGE**

Summary. — The importance of the projection screen as affecting the visibility of the projected picture and the box-office receipts of the theater is briefly discussed. In addition, the paper presents various phases with which those who are concerned with the selection and installation of screens should be acquainted, points out the difficulties attendant on the deterioration of the screen surface, and the additional cost of operation produced thereby.

Unlimited time and money are spent in improving, usually to a small extent, the many aspects of projection until, at length, the screen is taken into consideration. Here we stop; and yet, it is a fact that an improvement in projection of 100 per cent could be realized by making the appropriate corrections of the screen condi- tions. Not only would a large saving in electric current be achieved, but an improvement in box-office receipts would also be realized be- cause of the better appearance of the picture, and the improvement as regards visibility of the picture and the attendant strain on the eyes.

In the average American theater, the manager, whether employee or owner, assumes full responsibility for all details of operation of the theater. Various departmental subdivisions are made, one of the most important, if not the most important, being the department of projection. The management sees that the proper film is delivered to the theater at the proper time, purchases such equipment as is needed for projection of the picture, and relies on the department of projection to coordinate all details in such a manner as to assure the best picture.

The projectionist, then, is answerable to the manager for problems connected with the picture. But how far does this go ; where does his authority cease? An analysis of the facts shows that the projectionist is trained and equipped to assume direct responsibility for all prob- lems pertaining to the picture, whether inside or outside the booth. His familiarity with the principles of light control and of the mechani- cal details of operation is needed to decide what carbons should be used or what screen should be installed.

* Presented at the Fall, 1931, Meeting at Swampscott, Mass. ** Beaded Screen Corp., New York, N. Y.

909

910 FRANCIS M. FALGE [j. s. M. p. E.

The proper presentation of pictures, however, does not end with the placing of the responsibility on the projectionist. He must be provided with technically good film, of the proper density, and in the proper condition. Furthermore, his equipment must be in a satisfactory condition, and especially his screen must efficiently reflect the picture to the eyes of the patrons. Properly backed up by the manager, there should be no reason why a projectionist, with full responsibility for projection, should not have a picture at all times that is a real box-office asset.

It is a rule, rather than an exception, that the theater or mainte- nance manager of one or more theaters has, in discussing screens, said that he knew they were in a very bad condition, but that the condition of the business would not permit him to spend money now. Therein lies a fallacy, and a good reason why business is not so good. Further- more, profits are lowered because a dirty or improper screen actually causes a waste of money.

A careful study of this situation has shown that the only method of correcting this unfortunate state of affairs is to face the issue squarely, and to place the responsibility on the proper person, the projectionist. This would result in bettering projection as a whole, and in simplify- ing the problems of the management.

Some of the facts with which the projectionist should be familiar are as follows:

(1) The selection of the correct type of screen is dependent upon individual theater conditions, and especially upon the type of lamp, the angle of projection, and the width of the house.

(A) White diffusing screens reflect light about equally in all directions. They are best for wide houses; houses having large projection angles, and using high intensity lamps.

(B) Beaded directive screens redirect light into a beam so that the light reaches a majority of seats in a house suited to this screen. A smaller per- centage of light passes in all directions, so that a satisfactory picture can be seen from any seat not directly in the beam. These screens are best for me- dium width or narrow houses having projection angles under twenty degrees.

(C) Metallic reflective screens concentrate light into a narrow beam, with no diffusing element, and are suitable only for very narrow houses.

(2) When the type has been chosen, the most efficient of that type should be purchased, and it should be the one that will provide the best results throughout its useful life. The reputation of the manufacturer should be considered, and unusual claims should be carefully investigated.

(3) The proper size should be chosen :

04) Minimum desirable width is one-sixth the distance of the screen from the farthest seats.

July, 1932] A PROJECTIONIST'S PROBLEM . 911

(5) Maximum desirable width is eight-tenths the distance of the screen from the front seats.

(Q Intensity of lamps is an important factor in limiting the size.

(4) The screen should be properly installed by following carefully the direc- tions of the manufacturer. It should be masked in dead black. Screens should be placed not less than 18 inches from the stage floor, and as far from the front seats as possible.

(5) The projectionist should take note of the house lighting, and should suggest the elimination of glaring lights near the line of vision, and of spilled light on the screen.

(6) Great care should be taken to keep dust from settling on the screen.

(A ) All overheads and maskings should be kept clean.

(B) Doors and other openings that cause drafts through the screen should be kept closed.

(C) A front curtain should close in the screen when it is not in use.

(D) In many cases, it is essential that the screen be backed up to the horns with a non-porous material.

(7) Screens should be cleaned regularly once a week by brushing, by using a vacuum cleaner on the back surface, or by blowing.

(8) If recommended by the manufacturer, screens should be cleaned according to instructions every three to six months, depending upon the local conditions.

(9) For diffusing screens, it has been shown that resurfacing by spraying is possible, though the process is still in the experimental stage.

(10) Screens should be replaced in nine to eighteen months, depending upon the local atmospheric conditions and the care given the screen. Screens con- stantly diminish in efficiency, and, as a result, the picture constantly grows dimmer. When the efficiency of a screen has decreased about 30 per cent, the cost of the additional electric current is usually greater than the cost of a screen.

To assure that a satisfactory and efficient projection surface is present at all times, it is not only necessary to observe the above recommendations, but to provide for proper inspection of the surface. This should be the duty of the projectionist. At weekly intervals, a white booth light should be projected on the screen and the surface inspected for streaks, clouds, and discoloration. Then a small sample of a fresh piece of material should be placed against the screen and the loss of light estimated. A decided difference should be a warning that the screen either needs brushing, or that, due to age, it has deteriorated beyond the useful economical limit.

Recent tests made by the Projection Screens Committee of the Society show a loss of about 50 per cent in reflectivity of screens after a year's use. This means that with a low intensity arc, 30 amperes produce the results of 15; with a hi-lo, 75 amperes produce the results of 37, and with high intensity arcs of 120 amperes, only 60 are really being used effectively.

912 FRANCIS M. FALGE

Sound screens are porous and act as filters. The air passes through the screen, and the dust and dirt stay on the surface. Moisture and temperature conditions cause the dust to adhere to the surface in varying degrees causing streaks, cloudiness, and discolora- tion.

Taking a conservative 10 per cent loss of reflectivity for each three months' period, we find that at the end of the year, screens used ten hours a day, with electric power costing five cents per kwh., are caus- ing a loss of light and money as follows :

Loss Effective Daily Minimum 2nd

Lamp Intensity Amperes Amperes Loss Year's Loss

Low

25 Amps. 7 18 $0.37 $135.05

Hi-Lo

75 Amps. 20 55 1.10 401.50

High

120 Amps. 33 87 1.82 664.30

The second year's loss will be greater because of the cumulative effect on the loss of light.

Good business based on true economy will dictate that a screen that loses more than the cost of a new one should be replaced. The probable loss due to the patron's dissatisfaction with the dim, life- less picture and its harmful effect on his eyes is even more serious, though less easily evaluated.

The steady progressive decrease in light is constantly compelling the projectionist to devise ways of increasing the brightness of the picture. This is especially true when a dense print is used. Constant forcing of the equipment causes inefficient burning of carbons as well as troubles with the light source; an increase in the cost of carbons; trouble with the automatic feed; trouble due to the increase of heat in the lamp house, especially in the reflectors, condensers, and meters. Furthermore, the feed lines may not be capable of carrying the larger load, and trouble and loss may result from this cause. All this can be eliminated by keeping the screen surface in good condition at all times.

PHOTOGRAPHIC EMULSIONS' LEWIS W. PHYSIOC**

Summary. — A short story of the evolution of the photographic process, describing the various stages of improvement in speed and quality. The historical development of the photographic process is treated briefly, from the camera obscura through the early experiments of the daguerreotype, the discovery and use of hypo, the use of col- lodion and gelatin emulsions, to the panchromatic emulsion and the reversal process.

Those who are engaged in pursuits directly connected with the art of photography probably know that there is an interesting history represented in the development of this beautiful combination of art and science. When we compare our modern processes with the early experiments, our idea of mere history is enlivened by the elements of a thrilling romance. In fact, when we begin to study the various stages of development we see them unfold like the acts of a drama, reaching its climax in the most recent achievements.

For the benefit of those who have not the time to search through the many volumes devoted to the subject, it has been suggested that some of the salient features of this development be set forth in a brief, chronological order.

THE CAMERA OBSCURA

Although apparently known for a long time before, it was not until 1569 that the camera obscura was put to a practical use, when Baptista Porta devised a toy, later improved by Guyot, that was used for tracing natural landscapes and views to be reproduced. The various designs embodied the simple principle of passing rays of light into the darkened chamber through a pinhole aperture. W. H. Wollaston, in 1812, found that by using a single, meniscus lens in lieu of the pinhole, the image was rendered more brilliant and well defined, and after numerous developments of this principle the two Chevaliers of Paris made, in 1840, the first real photographic objective. It is interesting to contemplate that our most modern apparatus is merely an elaboration of this anciently observed phenomenon.

* Presented at the Fall, 1931, Meeting at Swampscott, Mass. ** Hollywood, Calif.

913

914 LEWIS W. PHYSIOC [J. S. M. P. E.

EARLY CHEMICAL EXPERIMENTS

With the chemical department, as with the camera obscura, we may revert to the ancient students, for as soon as men were capable of rational speculations they observed the effect of light upon various substances. A shield removed from a panoply after a long term of peace had left its outline upon the wall. The bleaching of hides was observed in very remote times. The early painters were well ac- quainted with the action of light upon various pigments.

The first important advance in this respect occurred when W. K. Sheele discovered that silver chloride became darkened by the action of sunlight. His experiments were interesting for three very signifi- cant reasons. In exposing the chloride under water he discovered that, as decomposition took place, a soluble compound was formed, and that by adding silver nitrate new silver chloride was precipitated. This was the first hint of the principles involved in the manufacture of an emulsion. Then, by adding ammonia to the darkened chloride, the insoluble metallic silver was left behind, a palpable suggestion of the developing process. Also, he noticed that the violet element of light acted more powerfully on the silver chloride, a condition that has placed a limit on the results of photography through many phases of its development.

From the time of Sheele 's discoveries other experiments followed with an enthusiasm that reminds us of those who are working over a puzzle and are close upon the heels of the key, producing silhouette prints of various objects, reproducing the symmetrical outlines and delicate tracery of the venal structure of leaves, etc.

But, unfortunately, the same light that made them destroyed them, and they could be viewed only for a short time. We can appreciate the disappointment expressed by one of those experimenters when he said "nothing but a method of preventing the unshaded parts of the delineations from being colored, today, is wanting to render this process as useful as it is elegant."

FIXING THE CAMERA IMAGE

After considering these experiments, it may readily be seen that the paramount idea was to fix the camera image. The first record of any sort of success is accorded a young Frenchman, Nicephore de Niepce. He strayed a little from the efforts of his brother workers, and combined some mechanics with his chemistry. He found that asphaltum was also subject to the action of light, but in a somewhat

July, 1932]

PHOTOGRAPHIC EMULSIONS

915

different manner. He covered a metal plate with a solution of asphaltum and oil of lavender, and exposed it for several hours in the camera obscura. The parts exposed to the light became insoluble, and the unlighted portions were dissolved away with a solution of oil of lavender and petroleum. The metal could then be etched, and the picture so obtained was probably the first example of photoengraving. There is nothing that might furnish a more interesting comparison between those experiments and our modern developments than Niepce's statement that it required from seven to eight hours to expose a landscape in open light.

THE DAGUERREOTYPE

Coincidental with Niepce's operations, there was a scene painter who likewise was obsessed with the idea of fixing the photographic image. He was a good scenic artist; nevertheless he was not satisfied with his own efforts. Scene painters have ever been men who liked to look around. His interest in photographic processes was so great that he neglected his painting, and his good wife became con- cerned about his mental condition. In 1829 he entered into a partnership with Niepce. Let us pause here to consider the earnest- ness of a man so set upon a project as to acknowledge his limitations in prosecuting his designs and to call upon his competitor for help. Such a man seldom fails, for he is more interested in the actual achieve- ment than in the glory usually attendant upon great accomplish- ments.

They plodded along together, not greatly improving their "heliog- raphy." They dipped in different chemicals and tried various metals for their plates — tin, copper, glass, and, finally, they knew not why, polished silver. There was a peculiar persistence about this idea of silver. Some say it was Niepce's idea, others that Daguerre insisted upon it. But the real truth had not dawned upon them. They were discouraged.

Finally Daguerre recalled some of the experiments of the earlier workers. They had shown that silver combined with iodine was sensitive to light, and he was convinced that Niepce's slow asphaltum should give place to the iodized silver plate. They tried it, but the images were so faint as to be useless. He and Niepce disputed, but soon Niepce died and his son carried on with the fanatical artist until accident solved the problem. The solution of the problem seems to have come about through the spilling of a bottle of mercury.

916 LEWIS W. PHYSIOC [j. s. M. p. E.

It is common knowledge that it is in the nature of some chemicals to give off fumes and of others to be greedy to combine with those fumes. So, when Daguerre returned to his investigations, he found that one of his old discarded plates had been acted upon by the fumes of mercury, and that each little particle of the faint silver image had become coated with condensed mercury. Daguerre discovered, to his joy and surprise, a well-defined delineation instead of the discarded failure. Here was another detail introduced into the foundation of the future structure of photography — a hint of the ultimate process of developing the "latent image."

The next development in the "fixing of the image" was introduced by another artist, Fox Talbot, a contemporary of Daguerre. He was out sketching, tracing a view from Wollaston's simple little camera obscura; and he, too, was inspired with the idea of capturing the fleeting image.

Talbot' s experiments were important in having pointed off another period in photographic history. He substituted the paper support for the metal and the glass plates and raised the peg another notch in the scale of sensitivity, for, as he said, "Subjects such as white sails in full sunlight may be obtained in half a second" a great advance over Niepce's eight-hour exposure.

In the adoption of the paper support we also recognize the fore- runner of our modern printing process.

HYPO: (SODIUM THIOSULFATE OR HYPOSULFITE)

In the discovery of the use of "hypo" our romance is given a dramatic mood. When we review the experiments already described we recognize one serious embarrassment. After having succeeded in fixing the camera image, the next difficulty was to make the reproduc- tions permanent, "to make them as useful as they were elegant."

The requirement was a solvent, strong enough to dissolve the un- lighted (or unchanged) silver salts, but not powerful enough to remove the metallic silver. The search was long and tedious, be- cause of the distinction between a physical solvent and a chemical reagent, and most solvents are powerful reagents. Various solvents were employed but were found unsatisfactory. Ammonia was tried, but this required a very strong solution which affected parts that needed protection; cyanide of potassium nearly met the require- ments, but was extremely poisonous and expensive.

It has been recorded that Francois Chaussier as early as 1799 dis-

July, 1932] PHOTOGRAPHIC EMULSIONS 917

covered the salt. Sir John Herschel, in 1819, was probably the first to discover that hypo could dissolve the unreduced silver salts without attacking the metallic silver. However, it was not until 1837 that J. B. Reade used it in photography.

The remarkable feature of this hypo is the fact that from the time it was first used, and through its many stages of development, its posi- tion as one of the most important elements in the art has been un- assailed by the most modern developments. We may be thankful that industrial chemistry and nature's stores have supplied us so bountifully with hypo. Its present price also furnishes an interesting comparison with that in the time of Reade, when it cost "half a crown an ounce."

THE WET PLATE

Despite the popularity of the daguerreotype, photographic enthusiasts strove for something better. They were probably spurred on by the agony of the victim who sat for a portrait, with body and head rigidly set in a brace, and whose face was covered with white powder make-up to reduce the lengthy exposure. Another item to be considered was the cost. The silver plates were expensive. It seems ridiculous when we consider how much more silver than was necessary was contained in one of those plates. Another undesirable feature was the limitation to a single picture. The natural induction, then, was to secure a negative from which could be produced many positives. In this conception lay the real germ of modern photography.

Paper negatives were tried, and were subsequently improved by waxing and soaking them in oil, to render them more translucent. Glass was the ideal support, but it involved new difficulties and demanded further experiments. It was necessary to find something that would adhere to the glass and would, at the same time, hold in suspension the sensitive silver salts. Albumen, starch, and serum of milk were tried, which for a while furnished a little hope.

Finally, Le Gray suggested collodion, a mixture of pyroxaline, alcohol, and ether. F. Scott Archer accepted the suggestion, and in 1848 announced the first real photographic emulsion. This marks an important period and involves several features. The virtue of the collodion is its nature to support the silver compound and yet not suffer from the effects of the silver nitrate necessary to form those compounds.

The collodion emulsion had one disadvantage, which led to the name "wet plate." Although the collodion held the silver salts, it

918 LEWIS W. PHYSIOC [J. S. M. P. E.

entered into no compound with them and, when allowed to dry, the salts crystallized and destroyed the fabric and transparency. It was necessary to expose it wet. After development, the unexposed salts were then removed and the plate allowed to dry without the crystal- lization.

The use of iodide of silver was another feature. With the exception of Daguerre, other workers had used the chloride which blackened directly in the light. The iodide, except under specific treatment, gave no visual effect of light. Daguerre's mercuric accident probably suggested the development process. The most important feature, however, was the marked increase of sensitivity over that obtained previously, and there were introduced a quality and a practicability that startled even the most enthusiastic students. The wet plate combined such features as a fineness of grain that has never been excelled, crystalline transparency and brilliancy, and a broad range of values. Even at the present time, process workers in the publish- ing business use this method when excellence of reproduction is re- quired.

While the sensitivity was low, compared to the modern dry plate, it was a great improvement over the Daguerre and Talbot processes, as the average exposure required was only ten to fifteen seconds.

THE DRY PLATE

Even the beauty of the wet plate could not long compensate for the tedious method of preparation and the inability to reproduce anything but still objects. In attempts to improve the process, it was observed that the soluble bromide salt entered into a finer solution with collodion and instead of floating the silver nitrate over the plate it could be introduced into the collodion solution drop by drop, forming an emulsified combination that could be spread on the plate; and that after washing out the solubles it could be allowed to dry. This is the