I’m actually curious if they made these black on white and then inverted it in post - the red is a little too deep, and there’s some interesting shadowing on the text, whereas if you invert the video, it’s incredibly crisp. Hard to really tell after 40 years and a YouTube compression algorithm, but it seems like it would’ve been easier to do it that way in the pre-digital era.
While what you’re seeing is almost certainly a compression artifact or digitization difficulty (keep in mind, theater screens were the target media,) it was almost certainly a product of phototypesetting, so it would have to be inverted at least once. In phototypesetting, you’d load a large dark negative with a typeface developed into it into a phototypesetting machine. A typesetter would enter the text and whatnot into the machine, and the machine would locate the appropriate letter on the film using some kind of timing mark (the machines were all different) and as the letter passed over the target area, it would shoot light from xenon bulb through it to develop that letter, in the correct place, onto the target film. This is how a lot (most?) typesetting was done in the decades before digitized fonts were a thing. That’s how they got such nice clean lettering on posters and things like that.
I don’t know for sure, but they probably shot a negative of the completed type with a red gel behind it to get the final frames.
IIRC early copies of BladeRunner in the digital era, like the original DVDs of this, really suffered from the lack of detail on red, specifically, from the 4:2:0 color space encoding ( https://spearsandmunsil.com/portfolio-item/choosing-a-color-... ). So some level of oddity around red could come from that.
Oooh that’s interesting - I’d been aware of the color space encoding, but hadn’t fully grokked how it manifested.
It’s interesting because it straddles analog & digital so many times - film is a chemical process, broadcast is an analog process, CRT screens are analog processes, and at the end of the whole thing, our eyes are an analog process, so you’ve got this wonderful bundle of path dependency (B&W -> Color, Analog -> digital, film -> broadcast -> DVD) where you’re trying to capture the characteristics of a dozen different physical phenomenon so you can throw a photon into someone’s eye Just Right repeatably.