# Colorization - is reproduction of the original colors even theoretically possible? **URL:** \ **Category:** Factual Questions\ **Created:** [June 11, 2011, 11:16pm UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165 "2011-06-11T23:16:24Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![KarlGauss](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/karlgauss/32/3713_2.png) [@KarlGauss](https://boards.straightdope.com/u/KarlGauss)\ **Post date:** [June 11, 2011, 11:16pm UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/1 "2011-06-11T23:16:24Z")
Colorization of old black and white movies (and even still photographs) got a lot of attention a few years ago and [could be quite controversial](http://en.wikipedia.org/wiki/Colorizing#Entertainment_make-overs). I am curious to know if it is possible, in theory, if not in practice, to deduce the original color scheme of a black and white image (say a [high quality photograph like this](http://1.bp.blogspot.com/_KKEVXDn98y8/TQJp0G1QisI/AAAAAAAAGrE/JYDLRNl4q10/s1600/Karcshwebsite30306flickr.jpg)). For example, could be there a unique correspondence in a given image between the gray scale of its pixels and specific (color) wavelengths? I assume _not_ (or it would be being done) and imagine the best that can be done is to assign a specific color to a “reference pixel” and then use that as an ‘anchor’ or reference to colorize the rest of the image. Can anyone elaborate on this? Thanks! ---
**Author:** ![aruvqan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/aruvqan/32/2891_2.png) [@aruvqan](https://boards.straightdope.com/u/aruvqan)\ **Post date:** [June 11, 2011, 11:48pm UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/2 "2011-06-11T23:48:57Z")
Problem being that for films and TV made specifically in black and white, costuming and makeup could be wildly different from the colors they were representing. If I remember correctly, makeup for B&W to look reasonably lifelike tended to be in greentones to get the shading and shadows correct. Costuming was similar. I wish Eve Golden was still here =( She probably has online pictures of the backscene makeup artists working that would show the colors of stuff. ---
**Author:** ![beowulff](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/beowulff/32/542_2.png) [@beowulff](https://boards.straightdope.com/u/beowulff)\ **Post date:** [June 11, 2011, 11:51pm UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/3 "2011-06-11T23:51:53Z")
I don’t think this is possible in practice. It’s theoretically possible to do if the color to monochrome transfer function is known, and the original colors are known, but even then there will probably be different colors that map to the same shade of grey. To try to deduce the original colors from a monochrome image is at best an educated guess (e.g. - skin is going to be pinkish rather than greenish, unless you are Vulcan). ---
**Author:** ![KarlGauss](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/karlgauss/32/3713_2.png) [@KarlGauss](https://boards.straightdope.com/u/KarlGauss)\ **Post date:** [June 12, 2011, 12:07am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/4 "2011-06-12T00:07:22Z")
> [@beowulff](#): > > . . . but even then there will probably be different colors that map to the same shade of grey. . . This, I think, is the essence of the problem. Is the function one to one, or many to one? ---
**Author:** ![j666](https://avatars.discourse-cdn.com/v4/letter/j/9de0a6/32.png) [@j666](https://boards.straightdope.com/u/j666)\ **Post date:** [June 12, 2011, 12:22am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/5 "2011-06-12T00:22:04Z")
> [@KarlGauss](#): > > I am curious to know if it is possible, in theory, if not in practice, to deduce the original color scheme of a black and white image (say a [high quality photograph like this](http://1.bp.blogspot.com/_KKEVXDn98y8/TQJp0G1QisI/AAAAAAAAGrE/JYDLRNl4q10/s1600/Karcshwebsite30306flickr.jpg)). One would never want to colorize a photograph of that quality. That was taken with the intent of producing a picture with the kind of contract and highlight only possible in a B&W. I am not some kind of luddite or purist - I prefer color photography. ---
**Author:** ![Lestrade](https://avatars.discourse-cdn.com/v4/letter/l/94ad74/32.png) [@Lestrade](https://boards.straightdope.com/u/Lestrade)\ **Post date:** [June 12, 2011, 12:26am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/6 "2011-06-12T00:26:52Z")
> [@KarlGauss](#): > > This, I think, is the essence of the problem. Is the function one to one, or many to one? It’s many to one. Churchill could have been lit by a red or blue gel when that photo was taken. There’s no way to tell from the image alone. You can see this if you play with a raster image editing program like [Krita](http://www.koffice.org/krita/). ---
**Author:** ![chacoguy](https://avatars.discourse-cdn.com/v4/letter/c/50afbb/32.png) [@chacoguy](https://boards.straightdope.com/u/chacoguy)\ **Post date:** [June 12, 2011, 12:38am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/7 "2011-06-12T00:38:25Z")
[This](http://media.thestar.topscms.com/images/c3/38/c927de5a40199e79a0c937316606.jpeg) image was colorized to [this](http://www.oneinchpunch.net/wordpress/wp-content/uploads/burning-monk-colour.jpg). \ It’s the famous shot of the monk in Saigon burning himself alive. It was only made possible by interviewing eye witnesses. ---
**Author:** ![minor7flat5](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/minor7flat5/32/258_2.png) [@minor7flat5](https://boards.straightdope.com/u/minor7flat5)\ **Post date:** [June 12, 2011, 1:19am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/8 "2011-06-12T01:19:23Z")
As others have said, B&W film only captures brightness/darkness, and many colors can map to the same level of brightness in the B&W photo. To muddy the waters more, many films are not sensitive to all colors in the same ways. In the early days, orthochromatic films were not sensitive to red, so red objects (e.g. flags) appeared black in photographs. Around 1915-1920, panchromatic films were introduced, with a more equal color sensitivity distribution, but the movie industry was slow to use them because of many factors such as the slower film speeds and the different response to lights (needing tungsten lights instead of arc lights). In addition, B&W photographers regularly use filters in order to enhance parts of the photograph. A red filter is great at making blue skies look beautiful dark, but then the trees sometimes look too dark, so they sometimes go for an orange filter as a compromise. Between different film color sensitivity curves, filters, and darkroom tinkering, there is really no way to match colors to shades of gray on a photograph. ---
**Author:** ![WhyNot](https://avatars.discourse-cdn.com/v4/letter/w/c77e96/32.png) [@WhyNot](https://boards.straightdope.com/u/WhyNot)\ **Post date:** [June 12, 2011, 1:34am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/9 "2011-06-12T01:34:10Z")
> [@aruvqan](#): > > Problem being that for films and TV made specifically in black and white, costuming and makeup could be wildly different from the colors they were representing. If I remember correctly, makeup for B&W to look reasonably lifelike tended to be in greentones to get the shading and shadows correct. Costuming was similar. I wish Eve Golden was still here =( She probably has online pictures of the backscene makeup artists working that would show the colors of stuff. Not just the colors, but the intensity of shading. High contrast was key to a natural look under the very bright lights and low screen resolution, which would wash street makeup out completely. Here are some (black and white) photographs of makeup done by Max Factor (yes, he was a real person) for black and white television. These were probably green, as the caption points out. [Cosmetics and Skin: Max Factor and Televison](http://cosmeticsandskin.com/aba/max-and-the-tube.php) ---
**Author:** ![KarlGauss](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/karlgauss/32/3713_2.png) [@KarlGauss](https://boards.straightdope.com/u/KarlGauss)\ **Post date:** [June 12, 2011, 1:47am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/10 "2011-06-12T01:47:49Z")
> [@j66](#): > > One would never want to colorize a photograph of that quality. That was taken with the intent of producing a picture with the kind of contract and highlight only possible in a B&W. Of course, I agree. I just wanted to give an example of a “high quality” photo. To the other people who have contributed: I appreciate your responses. What **minor7flat5** said was especially “illuminating”. And thanks for the excellent links! ---
**Author:** ![friedo](https://avatars.discourse-cdn.com/v4/letter/f/8edcca/32.png) [@friedo](https://boards.straightdope.com/u/friedo)\ **Post date:** [June 12, 2011, 1:51am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/11 "2011-06-12T01:51:42Z")
_Psycho_ would be a lot less scary if colorized “correctly.” The famous shot of blood going down the shower drain was actually something like chocolate sauce, because it showed up nice and dark on camera. That’s the problem with a lot of B&W movies: they deliberately used costumes and props of the “wrong” color to achieve a particular effect or contrast when rendered in B&W. So restoring the “original” colors would make even less sense than colorizing it to what it looks like it should be. ---
**Author:** ![Kent\_Clark](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/kent_clark/32/105_2.png) [@Kent\_Clark](https://boards.straightdope.com/u/Kent_Clark)\ **Post date:** [June 12, 2011, 2:23am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/12 "2011-06-12T02:23:25Z")
> [@friedo](#): > > _Psycho_ would be a lot less scary if colorized “correctly.” The famous shot of blood going down the shower drain was actually something like chocolate sauce, because it showed up nice and dark on camera. > > That’s the problem with a lot of B&W movies: they deliberately used costumes and props of the “wrong” color to achieve a particular effect or contrast when rendered in B&W. So restoring the “original” colors would make even less sense than colorizing it to what it looks like it should be. The police cars in the old B&W TV show _Car 54 Where Are You?_ were actually painted red and white, not black and white. The show did location shooting in New York City and no one wanted the TV cops to be confused with actual NYPD. Here’s a site that shows you [both B&W and color shots.](http://technorati.com/entertainment/tv/article/now-thats-funny-car-54-where/) ---
**Author:** ![Keeve](https://avatars.discourse-cdn.com/v4/letter/k/f07891/32.png) [@Keeve](https://boards.straightdope.com/u/Keeve)\ **Post date:** [June 12, 2011, 2:35am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/13 "2011-06-12T02:35:30Z")
> [@KarlGauss](#): > > This, I think, is the essence of the problem. Is the function one to one, or many to one? Many to one. Or at the very least, two to one. A very simple proof would be this: Given a particular shade of gray, how could one possibly know whether the original object was some color which ended up appearing as this shade of gray, or whether the original object really was that shade of gray to begin with? ---
**Author:** ![Keeve](https://avatars.discourse-cdn.com/v4/letter/k/f07891/32.png) [@Keeve](https://boards.straightdope.com/u/Keeve)\ **Post date:** [June 12, 2011, 2:45am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/14 "2011-06-12T02:45:39Z")
When the original _Adventures of Superman_ tv series was filmed in the early 1950s, the first two seasons were shot in black and white. But there wasn’t enough contrast between the blue and red, so the costume used brown for red, gray for blue and white for yellow. You can see a photo of it [here](http://www.originalprop.com/blog/2009/02/22/philip-weiss-auctions-cancels-special-superhero-weekend/). ---
**Author:** ![panamajack](https://avatars.discourse-cdn.com/v4/letter/p/47e85d/32.png) [@panamajack](https://boards.straightdope.com/u/panamajack)\ **Post date:** [June 12, 2011, 3:19am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/15 "2011-06-12T03:19:42Z")
Of some slight relevance, and possibly interest, is that some of the earliest color photographs were shot by using color filters on black and white film, which were then projected using colored light. [Sergey Prokudin-Gorsky](http://www.flickr.com/photos/prokudin-gorsky/) pioneered this technique (note the dates on those pictures). Computers make it quite easy to generate nearly perfect projections. You can see some mismatches in portions of the image with irregular motion, e.g. smoke. [This page](http://www.loc.gov/exhibits/empire/making.html) on the restoration shows what some of the filtered black and white images look like. ---
**Author:** ![minor7flat5](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/minor7flat5/32/258_2.png) [@minor7flat5](https://boards.straightdope.com/u/minor7flat5)\ **Post date:** [June 12, 2011, 3:34am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/16 "2011-06-12T03:34:01Z")
As an experiment, I did some B&W conversions of a very colorful clown in order to give folks an idea of the variations involved. I use an application called Silver Efex Pro to convert color photos to B&W. This is pretty much the gold standard for B&W conversion, with presets for many film stocks and an excellent rendition of film grain. I downloaded the [original photo from stock.xchng](http://www.sxc.hu/photo/1157681). [Original color photograph](http://img.photobucket.com/albums/v186/minor7flat5/bw.jpg) [Fuji Neopan ACROS 100](http://img.photobucket.com/albums/v186/minor7flat5/bw1.jpg) [Ilford PAN F Plus 50](http://img.photobucket.com/albums/v186/minor7flat5/bw2.jpg) [Agfa APX 400, with blue filter](http://img.photobucket.com/albums/v186/minor7flat5/bw3.jpg) Note how with the Fuji and the Ilford, the left-side lapel colors seem to switch. See how the blue filter on the Agfa darkened all the reds, making the nose much darker. ---
**Author:** ![zagloba](https://avatars.discourse-cdn.com/v4/letter/z/977dab/32.png) [@zagloba](https://boards.straightdope.com/u/zagloba)\ **Post date:** [June 12, 2011, 3:43am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/17 "2011-06-12T03:43:03Z")
A process called [colour recovery](http://en.wikipedia.org/wiki/Colour_restoration) has been used to recover the original color from black and white recordings of programs originally broadcast in color. ---
**Author:** ![Indistinguishable](https://avatars.discourse-cdn.com/v4/letter/i/90ced4/32.png) [@Indistinguishable](https://boards.straightdope.com/u/Indistinguishable)\ **Post date:** [June 12, 2011, 4:07am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/18 "2011-06-12T04:07:41Z")
> [@KarlGauss](#): > > This, I think, is the essence of the problem. Is the function one to one, or many to one? It isn’t mathematically possible to continuously map colors one-to-one into greyscale, given that greyscale is one-dimensional and colors can vary along more than one dimension. ---
**Author:** ![AaronX](https://avatars.discourse-cdn.com/v4/letter/a/7bcc69/32.png) [@AaronX](https://boards.straightdope.com/u/AaronX)\ **Post date:** [June 12, 2011, 6:51am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/19 "2011-06-12T06:51:14Z")
> [@Indistinguishable](#): > > It isn’t mathematically possible to continuously map colors one-to-one into greyscale, given that greyscale is one-dimensional and colors can vary along more than one dimension. But 3 \* infinity is still infinity, so there are as many colors as shades of grey. ---
**Author:** ![Indistinguishable](https://avatars.discourse-cdn.com/v4/letter/i/90ced4/32.png) [@Indistinguishable](https://boards.straightdope.com/u/Indistinguishable)\ **Post date:** [June 12, 2011, 7:57am UTC](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165/20 "2011-06-12T07:57:33Z")
In an arbitrary mapping sense, not a continuous mapping sense. [Next page](https://boards.straightdope.com/t/colorization-is-reproduction-of-the-original-colors-even-theoretically-possible/585165.md?page=2)