color determination from bw photo?

[jack_nordine]

<p>Even though I've been into photography for over 40 years, a friend recently asked a question that I couldn't answer. I'm hoping someone here can assist. My friend has an old black & white photo of a home he recently purchased. The photo is from the 1930s. He would like to restore the house with it's original colors, but cannot determine the colors due to the bw photo. Is there any way to identify true colors from a B&W photo? Thanks in advance.</p>

[john tonai]

<p>No</p>

<p>Colors are made up of 3 parts<br>

Brightness or Luminance (you have that in a black and white photo)<br>

Saturation (you have that in a black and white photo)<br>

Hue (this is missing in a black and white photo)</p>

[Mattman944]

<p>No, only clues based on the shade of gray in the photo. He could do some research to see what colors were popular in that area of the country at the time the house was built. If it is in a historical area there might be more specific records.</p>

[jack_nordine]

<p>Thanks for the responses.</p>

[jay_de_fehr]

<p>Color can be recorded in B&W negatives (plural) -- you need one each filtered through red, green and blue filters. These are known as separation negatives, and this was the very first color photo process, used in Russia as early as 1909:</p>

<p><a href="http://www.loc.gov/exhibits/empire/">http://www.loc.gov/exhibits/empire/</a></p>

[jack_nordine]

<p>After re-examining Matthew's response, would it be possible to guess with some amount of accuracy the original colors based on the shades of gray?</p>

[lex_jenkins]

<p>If you know the film used and the spectral response of that film you might be able to get pretty close to the original colors. Presumably it's a panchromatic film, but it would still be interesting to compare spectral sensitivities of 1930s era pan films.</p>

<p>Photographing standard color swatches in similar lighting might help as well in comparing gray tones.</p>

[Bill C]

<blockquote>

<p>After re-examining Matthew's response, would it be possible to guess with some amount of accuracy the original colors based on the shades of gray?</p>

</blockquote>

<p>No, very little accuracy. Some colors might be ruled out, for example: colors such as pink and yellow are never dark. Therefore, if the B&W shows certain areas as very dark, then you can probably presume these areas were not pink or yellow. But for a given tone on the B&W, there are many possible paint colors. So although you can rule out certain colors, you can never nail anything down.</p>

<p>The situation might change if you had a specific list of 5 or 10 possible colors. Then you might be able to narrow it down by "brightness signatures."</p>

<p>An item that could give some clues: if part of the house is in direct sunlight, and some in shade, there will be a clue as to whether the paint has blue content. A non-blue color, reddish, yellowish, greenish, etc, will not reflect the bluish shadow light as well, thus they'll get darker in the shadow than you would expect.</p>

[garethspics]

<p>Slightly off-topic, but of some interest is the fact that the BBC managed to recover colour information from monochrome film recordings. The film camera was pointed at a monochrome monitor which was replaying the original colour signal. The colour information was displayed as a fine patterning which is normally stripped of by a colour receiver and decoded to give the correct colours. When the film was scanned decades later this pattern was clearly visible and the correct colours were restored.</p>

[Bill C]

<p>It just occurred to me that you could get a sense of the difficulty by simply looking at a desaturated image of a Macbeth Colorchecker. Here's a link where someone has experimented with that. <a href="http://www.thelightfarm.com/cgi-bin/htmlgen.py?content=15Jan2012">http://www.thelightfarm.com/cgi-bin/htmlgen.py?content=15Jan2012</a></p>

[glen_h]

<blockquote>

<p>If you know the film used and the spectral response of that film you might be able to get pretty close to the original colors. Presumably it's a panchromatic film, but it would still be interesting to compare spectral sensitivities of 1930s era pan films.</p>

</blockquote>

<p>You could extract some color information if you had an orthochromatic and panchromatic picture of the same house.<br>

There probably isn't enough difference in color from sunrise/sunset lighting to noon lighting to extract, though, but in theory if you have pictures with even slightly different colors of light you can extract color information.</p>

[g_robinson]

Hi Jack,

 

Yes you can determine colours from many black and white photos!

 

We were asked at the Doerner Institut in Munich to do this for an organisation wishing to restore their Berlin property to its original colours

from b&w photos. We didn't have the resources to take them up on the challenge at the time.

 

About one in a million Britons suffer complete colour blindness and only see in black and white. To compensate they become very

sensitive to shades of grey, and with constant exposure to others telling them what colour objects are, they learn to recognise colours

independently. A bizarre consequence is that they can also correctly identify colours from original 405 line monochrome TV

transmissions, or from monochrome cine films!

 

As another member suggested, this arises from the interplay of known colours upon unknown colours. A deep blue dress reflects very

differently onto rosy cheeked woman, than a red dress. The corner of a building taken at 45° with direct sunlight at 5780K on one side,

and 8000K blue sky on the shadow side will produce a ratio of densities that depend on the colour of the wall paint. For a given set of

conditions that include the level of cloud cover, and geometry of the camera, sun & building, then by varying the building colour only, the

density ratio of the two wall faces will change!

 

But who has access to all that data, and can calibrate out the chemistry of the negative and paper too? We didn't, so we rejected the

request.

 

However, it might be simpler to find any one of the 60 fully colour blind individuals in the U.K., and ask them to cast their eye over the

photos. If the masonry paint is not too pale there's a chance that they can identify the colour from experience.

 

PS. Here's a related topic: Is it possible to identify colours in a scene when low pressure sodium lighting is the only illumination present?