By VIRGIL E. MILLER, A.S.C
from American Cinematographer March, 1942 p.107
WHEN Panchromatic film first came into use, I became a filter-Factor “fan,” because it opened up a new field in exterior photography. Orthochromatic negative, being sensitive to the upper or red end of the spectrum, gave us untrue renderings of color-values; the bright end of the spectrum photographed dark; reds were black on the screen ; orange and yellow and green took on nightly hues that beautifully camouflaged the true relative brightness of the objects being photographed.
Nearly, twenty years ago I made my first experiments with Wratten filters; at that time I made a split-screen test of all the filters then available, not only comparing the unfiltered to the filtered values, but further comparing one filter against the other. I will not go into the fundamentals as we all know then, i.e., a red filter permits the passage of the red rays and stops the blue, etc.—those facts are self-evident to all cameramen. But from the start I found certain limitations being imposed by the use of single filters; I could easily secure “over-corrections,” but they weren’t believable. For instance, I was sent out by the Universal Studios to photograph in a night-effect shot, a train of cars crossing the desert country just above Toluca, (later Lankershim and now called North Hollywood). I used my red filter; it turned my sky black, but all my red cars came out white—a ghost train was the result.
Knowing that panchromatic film of that period was almost insensitive to green, I decided that perhaps a red filter plus a light green filter might give me truer renderings of red objects and still give me a black sky and also permit the headlight on the locomotive (or automobile) to register.
I had purchased a book of filters from Rowland & Dewey (now Los Angeles’ Eastman Kodak Stores) and had one green filter to experiment with on this particular assignment. The test exceeded expectations and those who saw it declared it the nearest to real moonlight they had seen. I exhausted my supply of green gelatin, and could not secure any more in Los Angeles. I ordered three squares (3″) at $1.80 per square; with them I received a letter from Dr. C. E. K. Mees, A.S.C, of the Eastman Kodak Company, asking what I intended using them for.
I wrote back, saying that the No. 56 green in combination with the No. 23-A or the No. 2.5-A gave me very believable “night renditions” in the daytime.
I received a letter saying such a combination would give me nothing on the film, since such a combination was complementary. I sent Dr. Mees a roll of the film “shot” with the combination; his letter in reply said that they were astounded and that their film was possessed of qualities of which they were unaware, and that he would be on the coast shortly and wanted to talk to me. Needless
to say, we met a few weeks later and he commended me very highly for what I had done . . . thusly “combination filters” were “born” twenty years ago.
Down through the years panchromatic film has been improved to the point where it is sensitive to the entire spectrum; in addition, we have our “infrared” negative that lets us explore the invisible reds and completely eliminates all or most of the shorter wave-lengths. Night-effects are obtainable with this infra-red, but to most of us such effects are so greatly overcorrected that they are often unbelievable. It is for this reason that a truly panchromatic negative yields results more believable and therefore more satisfactory for picture purposes.
Since Twentieth-Century Fox Studios use Eastman negative films exclusively, my recent experiments have been confined solely to that film. The split-screen tests illustrated were made on Plus-X, and were made during the noon hour, inasmuch as the light changes vary less during that period. I used Wratten gelatin filters, behind the lens to make them less susceptible to extraneous light. The background remained the same for each half-portion of the frame; I “panned” over to include exactly the same area.
After ascertaining the footage desired for the entire series of tests of filter values, I matted off one-half the aperture and photographed the scene unfiltered, using a meter for determining a printing range suitable for the purpose. I then rewound the film and, using the identical background, immediately photographed about twenty feet with each of the filters used, viz: Aero 1, Aero 2, G or 15, 3N5, 5N5, 21, 25, 29, 72, 23A-56 and 25A-66.
The negative was developed normally and the positive was printed on ONE light, based on the best printing value for the unfiltered side of the negative.
My first test, in which I used the generally accepted filter factors issued by the various film companies, revealed one thing definitely; the factors used for day renderings were approximately accurate; those used for night-effects were much too high. This was evidenced in our running of the film, as it varied in its densities. Had the factors been correct, the densities would have varied only slightly.
Please understand that I am not talking now of color-corrections and their variations, but of over-all densities. Factors, as they are called, are primarily useful in securing proper densities; they have nothing to do fundamentally with color-renderings. In other words, a Director of Photography must know what filter to use in order to secure the color rendition he desires; only experience can accurately determine his choice; after that, he should know the factor to be used in order to bring his preconceived color renderings within the scope of the laboratory’s printing limits.
Using my first test as a key to factor changes necessary to balance my densities, I shot another test. This second test more nearly fulfilled the requirements for balanced densities. As mentioned, the “day” values were very close to those obtained by using the prescribed factor values ; the appended table shows only slight differences. But those factors used for night-effects in my second test show wide variations from those generally recommended, as may be seen in the table.
The following filters are those most used for night-effects.—Numbers 25, 29, 72, 23A-56 and 25A-66. It is this group that I found necessary to alter in factor values in order to give me suitable and
believable night-values well within the limits of our laboratory’s printing range.
| Filter | Commonly Accepted Factor | Recommended Factor |
|---|---|---|
| Aero 1 | 1,25 | 1,25 |
| Aero 2 | 1,50 | 1,50 |
| 3N5 | 4. | 4. |
| 5N5 | 5. | 5. |
| G-15 | 3. | 2. |
| 21 | 3,5 | 2,5 |
| 25 | 7. | 4,5 |
| 29 | 15. | 7. |
| 72 | Wide open | 12. |
| 23A-56 | 7 to 15 | 5. |
| 23A-66 | 7 to 15 | 5. |
I have used two combinations above shown; either are very acceptable, but personally I prefer the 25A-66 for this reason; the 25A better corrects for night values and the lighter 66 green retards
the heavier red in rendition of face values. Circumstances alter cases and perhaps the 23A and the heavier 56 green may give a little more detail where less sky correction is needed, and face values
are not so important.
A further discussion of factors would be only “painting the lily” since cinematographers are well aware of their general uses; a comparison of the illustrations will bear out the general theme of
this paper and emphasize the fact that I have tried to bring out, viz: that there has been a tendency to use factors too high to secure the intended results.
END.