There is a process called “Chroma Dots” that allows colourisation of material that was originally broadcast in colour but recorded in black and white. It was used to colour episode 3 of the old Doctor Who serial Planet of the Daleks when recently released on DVD.
ETA: Obviously this doesn’t help the photo in the OP <my bad>
That’s interesting about the Chroma dots. I was interested in the process in general as well as the particular photo as well so thanks for the link.
I, too, expect red to be dark in a black and white picture. Perhaps the author happens to know what color the dress truly is from seeing it, or from descriptions she’s heard or red. It’s her mother in the picture, after all.
Right. To summarize, they’re using the fact that in PAL (and NTSC) analog broadcast signals, the color values are encoded at high frequency and summed with the luminance (gray level) values. If you have a recording of the original composite signal, without too much noise in it, then you can extract the original color information — just like analog color TVs already do.
Black and white film doesn’t have that kind of fancy signal processing behind it, of course.
I suppose, theoretically, that a high-resolution photograph of a black and white TV which was, at the time it was photographed, showing a color broadcast could be analyzed in the same way, and from that you could derive the original color in the broadcast. You’d then be in the funny position that you could recover the color on the TV screen, but not the colors of the people and objects in the room.
This computed image would be the “opposite” of what was actually photographed: people and objects in color, with the TV showing black and white.
Somehow, computers figure out colors from B & W when films are colorized.
I thought they started with a known color. Perhaps a costume like the red one in Jezebel. Then interpret the others from that baseline.
If shades of gray doesn’t give shades of color. Then, how are movies colorized?
I’ve heard the colors in colorized movies are pretty close. Sometimes too close. The costumes were chosen to look good in B & W (like the Superman example mentioned earlier). Seeing their true colors is often disappointing.
It’s PEOPLE!
Sorry.
The “computers” don’t figure anything out, they just keep the colorizers from scribbling outside the lines.
I would like to know too. Grapheme-color synaesthesia is the most common, but I don’t see how those are related to each other.
In case panache45 doesn’t check back in, I’ll just mention that my family of synesthetes can “see sounds as colors”, or (rarely) greys as colors, but it’s just a function of brain wiring. In other words, our brains are making up colors. No good for determining what color something originally was.
Cute anecdote alert … When the BBC broadcast the Dad’s Army episode they’d used the process on, there was a lot of coverage of the achievement, but maddeningly little detail of how it’d been done. Having come up with a guess as to how it had been, I causally mentioned my hypothesis to a software engineer friend who works for a telecoms company. Turns out that not only was I basically correct, he could say so because he was one of the team who’d worked on the restoration of that episode.
Don’t know if he was involved in the Doctor Who project.
That would be my guess as well. It’s not a very helpful way to describe a black and white photo, since the rest of us can’t tell what color the dress was, but perhaps the caption-writer was (rather clumsily) trying to provide that information.
My thing was OK the dress of the lady sitting down in the picture is supposed to be red. But it looks about the same color as the bride’s dress.
So I guess this would make sense if the bridal dress was red. But how many brides wear a red dress?
If the bridal dress were white, it seems the dress of the lady sitting down, if it were indeed red, should be darker in a B&W photos.
That’s just the way it would seem to me, I don’t know anything about it so I could be wrong, but that’s what I’d guess
I think in order for that to work, you’d have to assume an unreasonable level of efficiency in the photographed B/W TV - that is, no smearing of the data due to phosphor persistence or a dozen other mitigating factors. I doubt the ‘chroma dots’ data actually survived through to the front of the CRT, even though it may have been there in the received signal.
The dress of the sitting lady is a little bit darker than the bride’s. But, even so, I don’t think the dress itself could be anything darker than a rose (not a deep red.) But it very well could have a red tint. It’s just impossible to determine what that tint is. Her dress is definitely lighter than the red of the roses and lips.
I have a book of baseball uniforms since 1900, and the author says that in old B & W photos, generally reds are darker than blues. Beyond that, guessing the color of things in B&W photos is definitely not an exact science.
According to Jay Neitz, normal color vision humans can distinguish between about 1,000,000 hues (100[sup]3[/sup]). Monochromats, who only see black and white, can distinguish about 100 hues. What this suggests is that lots of confusion between colors can occur if you can only see black and white.
If the pictures were taking with orthochromatic film yes, a red with equal luminance (brightness) to blue will look quite noticeably darker in the print, even they may have exactly the same “brightness” in real life. With panchromatic film (which has spectral sensitivity across all the color wavelengths), they will appear more-or-less equal. (It depends on the exact emulsion. Although panchromatic film is sensitive to reds, blues, greens, etc., it’s not necessarily equally sensitive.
That said, in either case, there are many, many colors that will map to the same shade of gray. Reds and greens are probably the most common with panchromatic film, as what we call a baseline red and a baseline green are of very similar luminance. This is a big part of the reason for using color filters in black-and-white photography: to create contrast between colors that otherwise have the same luminance and would appear as the same shade of gray in a straight black-and-white print.
Sorry to be so late with this.
I shouldn’t have said “some of US,” since my synesthesia doesn’t work that way. I “see” colors in letters, numbers and musical notes and keys. I don’t literally see the color, but my brain just “knows” what the colors are. So I know that the letter A is red, as is the number 8, and an e-flat-major chord is yellow-orange. Some synesthetes actually see the colors; I don’t.
Now consider a black & white photograph. You’re looking at a scene, and even though you’re not a synesthete, you just “know” that the grass is green. There are some synesthetes who claim to actually see the color. I don’t.
And when I look at a grayscale photo of an abstract painting, my mind assigns colors to the shapes, though of course the information from my eyes is grayscale . . . and the colors I get from my brain may bear no resemblance to the colors in the actual painting. Just as the colors of letters and numbers and musical notes and keys may bear no resemblance to their actual colors, if any. But if I see a letter A that’s green or a number 5 that’s blue, they seem as wrong to me as if I had seen purple grass.
From what I read in the early days of colorization - researchers who cared, particularly on classic films, would research to find old colour publicity photos of the sets, remaining props, etc. and interview anyone involved in the movie if they could to get some idea what colour various props (decorations, dresses, etc.) were.
The story goes that the original silent films used the film more sensitive to blue; so white makeup was preferable to skin tones, as it showed lighter. Blue shirts would show better than white (I presume because they reflected higher into the UV range?). Also, in B&W photos from those days, the red lipstick shows as very dark black, making high contrast between the lips and face.
This is why I suspect the example photo is panchromatic film - the subjects’ lips do not seem to be very dark, suggesting that deeper pink/red is showing as faily light on the film. The dress must ahve been a lighter red too, but in a hue too red to be described as pink? It also gets a lot of light reflection - maybe it’s shiny silk or satin; the lighting, based on shadows and satin highlights appears to be coming from upwards to our right (their left). It seems more like regular lighting or sunlight than a direct flash from the camera. The roses, assuming they are deep red, show as grey rather than dark black as would be expected for ortho film.
I agree with everything you said, but a studio photographer would, presumably, use some kind of diffuse off-camera flash in this situation, so it doesn’t look like a direct flash. Here’s flash portraiture examples from a book written in 1909.