Is there any way to change the image taken under a sunlight (i.e. full and continuous colour spectrum) to look like taken under a light source with line spectrum (e.g. mercury or sodium street lamp)?

high pressure sodium lapm spectrum
mercury vapour lamp spectrum

Note: Setting the white balance is not enough. The line spectrum of such sources distorts colours in a specific way. (For example, mercury vapour lamp makes the white human skin look yellowish/greenish while the white painted wall in the background looks white.)

Thanks for any suggestion.


1 Answer 1


Yes, this is easy if you have a good ICC profile creation tool like i1Profiler and a color checker shot with your camera.

Step one: Create an ICC profile from the color checker data and use the current light source condition as the illuminant for your profiles white point..

Step 2: Create another profile with the color checker and the new illuminant. (e.g. mercury or sodium street lamp) Since it's not a standard Illuminant. You will have to plug in your trusty i1Pro2 or ColorMunki and go out and measure the light source you want your images to use.

Step 3: Now you have the record of the data for both illuminants built into profiles. Since what you want is for your images to have that casted look of a different white balance, so to do that you convert to the first profile with the correct white point. This should keep the colors the same on screen, then simply assign the new profile. That will show you what your image would look like under the new illuminant. If you like what you see then convert from that new profile back into a standard color space like Adobe98 or sRGB so you can share print or post it.

  • \$\begingroup\$ Does this method work w.r.t. the note in the question about human skin vs. a white wall? I have the feeling that the difference noted is caused by a difference in reflectiveness of a certain areas. I.e. the different surfaces "react" uniquely to a certain light source. This would mean that any method where the whole image is altered (without masking) would never be able to replicate the look of a specific lightsource. Not trying to bash your answer, I'm just curious. \$\endgroup\$ Commented Mar 24, 2014 at 11:47
  • \$\begingroup\$ As with any color management method, it's specific applications can not be predicted due to the fact that the spectral sensitivities of the specific camera are not known nor are those in any specific scene. Color management works best if all important colors are actually the ones that are known and managed. \$\endgroup\$ Commented Mar 24, 2014 at 13:59
  • \$\begingroup\$ Maybe my comment was not clear, but what I meant was that different materials (skin, wall paint) seem to "react" differently to e.g. mercury vapour lamps. I thought this might be a physical material property, which, if not recorded at capture time can not be emulated in post. Seeing that the answer is accepted I'm probably wrong. \$\endgroup\$ Commented Mar 24, 2014 at 18:46
  • \$\begingroup\$ @BartArondson It is correct that different materials have different spectral reflectances. Given that we are shooting with a RGB capture and not a device with known spectral sensitivities, metamerism is an issue that may require additional post correction \$\endgroup\$ Commented Mar 24, 2014 at 19:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.