Two-strip Technicolor… how exactly does that work? How were those films able to get the range of whites and blacks that they did with only two colors? The answer is a tricky one. Ultimately, experimentation was used to pick the particular color primary values that would get pretty good results given the limitations of using only two primary colors.
First of all, let’s evaluate the bare minimal extreme where only the lowest quality is achieved. Additive color with only red and green will only give you a range of colors from black, red, yellow, and green. You will not be able to achieve pure white because of the lack of a blue component. Thus, all images will generally be tinted yellow.
Again, please note that I am generating these color displays in a gamma-correct fashion, using gamma = 2.2 rather than full sRGB for simplicity.
Let’s review some very basic information on film processing. Typically, film is “negative film”: When it is exposed to intense light, it becomes dark (opaque), but when it is not exposed to light, it remains light (transparent). This is referred to as a “negative image,” of course. Negative images are typically not useful for viewing, so they must be negated again to generate a “positive image.” Less common is “reversal film” which, after development goes straight to a “positive image” and does not need further processing to be viewable. In the case of taking black and white film and toning it, the transparent area remains transparent, and the dark/opaque area is replaced with the asserted color tone. Color tones are described by the color of light that remains after filtering, not the color of light that has been subtracted, i.e. CMYK (cyan, magenta, yellow, black) printing tones.
In the case of Technicolor filming, we produce a red source strip by filming behind a red filer and a green source strip by filming behind a green filter. The film stock itself is panchromatic black-and-white negative film. Interestingly, the question of additive versus subtractive color affects only the presentation of a film; the capture process is the same in either case.
Now, let’s try the very basics of subtractive color. With subtractive color, you start with white light, then you can subtract light to control the display of red and green. But, here’s the key. If you only subtract away red and green (toned cyan and magenta), you’d be left with blue as “pitch black” rather than pure pitch black. So, the trick that you play to get pitch black is that when you subtract either red or green, you will also subtract an equal amount of blue. Essentially, this means that we process the red and green source strips so that the positive red strip is toned green (subtract red & blue) rather than cyan (subtract red) and the positive green strip is toned red (subtract green & blue) rather than magenta (subtract green). Unfortunately, when we layer two such strips together and are displaying shades of gray, we “subtract” equal amounts of the blue channel twice. Also, we sacrifice the ability to display pure yellows for the ability to display grays.
Please note that when we layer two filters together, we do not add the amount of light they subtract, but rather we multiply the percent of light that each filter transmits. These are, after all, filters that will allow, say, 50% of light to pass. So 50% times 50% is 25% light transmitted, for example. Because of this, pure shades of gray will end up being tinted yellow with this kind of subtractive color.
Compare that to this pure gray strip.
But, here’s the key that we can use to our advantage. Because we multiply filters when layering two of them together, we can adjust the blue light filtering so that rather than subtracting 100% of blue light at full intensity, we can only subtract up to 90% of blue light. When we want to display pitch black, we will get the product of 90% times 90% filtered, or only 1% of light will be passed through. This allows us to still be able to display pretty convincing blacks, while at the same time getting more pure grays with the filters. And, we minimize the amount we compromise on being able to display bright red and green hues.
Okay, so how does two-strip Technicolor really filter the light? For the red source strip, my technique and theirs agree head-on by toning cyan-green. But for instead of using magenta-red for the green source strip, they use orange-red.
20200222/https://en.wikipedia.org/wiki/Technicolor#Two-color_Technicolor
The best way to understand two-strip Technicolor is to watch one of the actual historical movies that used the technique. How about the oldest such movie, The Toll of the Sea? For a long while, it was thought to be lost, but in the past few years, it looks like the Wikipedia article on the movie has been updated to say that it has been found and restored in the 1980s. Now that it is over 90 years old and in the public domain, a copy of the movie has been uploaded to Wikipedia. All of the actors depicted are long gone, of course.
20200222/https://en.wikipedia.org/wiki/The_Toll_of_the_Sea
20200222/https://en.wikipedia.org/wiki/Anna_May_Wong
20200222/https://en.wikipedia.org/wiki/Kenneth_Harlan
20200222/https://en.wikipedia.org/wiki/Priscilla_Moran
These filter choices explain why I sometimes saw weird purple hues in The Toll of the Sea movie. Because the cyan-green tone choice for green strip did not subtract out 100% of the available blue light, this made it possible for purple hues to be displayed under some circumstances where there was bright reddish light.
What would explain the choice of orange-red toning for the red strip? I imagine this may have been chosen to optimize the display of skin tones, and maybe it also helps compensate for the fact that we can no longer display yellow as in red-green additive colors. It also completely eliminates the possibility of displaying cyan colors at all, which I did notice that all blues of any sort were conspicuously absent from the movie, being substituted with a variety of green colors. You can only see so much green before you start feeling tired and “green” about it. Also, in regards to over-subtracting for the blue channel… even though we have one channel that subtracts 100% blue, the fact that we still multiply the filters and have reduced the strength of blue subtraction on the other channel means that we still do get closer to displaying more ideal grays. Also, the scene where there Lotus Flower was looking at the black and white photograph of Allen Carver, the photograph was generally slightly tinted green, with the very darkest areas tending toward more of a yellow tint. The additional green tint in particular is probably due to the orange-red toning adding in a bit of extra green light.