BACK TO #NerdyFilmTechStuff

Thanks, man!

[In response to your question about the proposition that film’s look, as distinct from digital, might be attributable to film’s red layer being “softer” than its blue and green layer.]:

I don’t believe that the Red layer being less resolute than Green in film is a distinguishing look at all. I think that’s another speculation that’s not based on controlled observation. 

There is a very illustrative experiment that's super easy to execute which show pretty unequivocally that a blurry red channel is not a defining perceptual attribute that differentiates film from digital.

To execute the experiment, you need:

-A good film scan of a resolution chart (like, an uncompressed DPX. It can be a 2k down-res from a 4k scan).

-High quality digital images (say, uncompressed ArriRaw from an Alexa). For the digital, you’ll need both a focus chart AND one or more normal scenes from a film.

The method: 

1. In Nuke or something, view the film scans one channel at a time. Green is the sharp channel, so compare viewing red-only to viewing green-only and then compare them both to the full color image. First off, it’s not even certain that the red channel actually is “irresolute” compared to green. On my focus charts that I have here, the red channel looks slightly flatter so the focus lines don’t pop out quite as much as they do in the green channel, but they seem to actually resolve to almost the same physical point in space. [This seems tentatively to indicate that red may have reduced MTF, or perceptual sharpness, but not actual reduced resolution.] Already in step 1, the assertion that the red channel on film is blurry in any kind of meaningful way seems suspect, but let’s give it the benefit of the doubt and continue on with the experiment anyway.

2. Now let’s load up our digital images. Set up an adjustable blur on the red channel only (so you can increase and decrease the blur — and it’s only applied to red, not the other two channels). Apply this blur and look at the digital image of a normal scene from a movie (not the focus chart). Look at it in normal full color (not one-channel at time) and slide the blur up and down, you’ll see that you can actually blur the red channel quite a bit before there is ANY perceptual change at all. 

3. Now, slide the red blur up until you can just barely make out its effect, then slide it back down a bit so that it’s still turned up, but is completely undetectable perceptually. 

4. Now, with that amount of blur still turned on, switch from the normal scene to the focus chart and switch from viewing it in normal color to viewing the red-channel-only. You’ll see that it’s pretty damn blurry. Compare that to the film focus chart, also red-channel-only. You’ll see that you’ve actually blurred the digital image’s red channel WAY more than the film’s red channel can be said to be irresolute.

This proves that even way more red-channel blur than is present in film has no perceptual effect whatsoever. (In fact, the difficulty in seeing red and blue channel resolution is why Bayer pattern has twice as many green photosites as red or blue). 

Let’s take it even one step further: we know that it’s possible to aesthetically appreciate the “look” of film even in low resolutions like standard def — I mean that’s why commercials and TV movies and such were always shot on film in pre-HD days. If you did the same experiment described above but in standard def instead of 2k, you’d be able to blur the red channel in the digital image to a comical degree before you could see any perceptual change in the normal scene. Additionally, if we scale our film resolution chart to standard def and again compare the red channel to the green, it is now unquestionable that there is no resolution difference between the two . So the fact that you can appreciate the “film look” in standard def, when there is no resolution difference between red and green, again proves that the film “look” is not tied to the resolution [or perceptual sharpness] of the red channel.

These are all tests I’ve done before, but I actually did them again just now to double check while I was writing this.

Hope that helps.

-Steve