Optimising colour rendition

Posted:

In theory, it should be possible to just shoot a digital photo of each film frame with a macro lens. But, is not as trivial as it may sound. As mentioned, I have thousands of photos, diapositives, colour negatives and black and white negatives.

As for diapositives, this is indeed straight forward, all that is needed is a good light source. Colours will then be reproduced correctly with a single exposure. I might need to adjust black and white point a bit, but that's all.

Black and white negatives may have a bluish cast, but since this is black and white, any colour cast can simply be removed. There may be a problem with blacks and whites, as both may end up greyish, but with Adobe Camera Raw you can set the black and white point with the tone curve tool. And the good thing is that this setting can then be applied to all frames on the strip. The problem is that Camera Raw does not allow to invert an image, but the tone curve may even do that job.

Colour negatives are the real problem due to the typically orange cast the negatives have. Wayne Fulton at scantips.com explained me that photo scanners compensate for this mask by exposing each channel differently. There is about one stop difference between the green and the red colour channel, and 0.5 stop between the green and the blue channel. So, just taking a photo of the negative, the red will be wildy overexposed or the blue and green underexposed.

I did a test shooting an unexposed film strip and trying to center the peak for each colour channel, the result was indeed 1 stop difference between red and green, and 2/3 stop difference between green and blue.

There are two methods for compensating for the orange mask:

Blue shifting the light source

Wayne Fulton suggested to try to compensate for the overexposure of the red channel using daylight colour temperature conversion gels which shifts the light towards the blue spectrum. These gels, CTB (Colour Temperature Blue) comes in stops or fractions of a stop and with the proper combination, each channel will be more equally exposed. Which combination works best requires some testing. The problem is that these gels affects all colours to some extent and doesn't just filter red.

Exposing for each colour channel

An alternative method is to use bracketing exposure that many digital cameras support. Shooting, say, 5 or 7 photos with a 1/3 stop difference and then merge each colour channel from different exposures into one resulting image. Based on my test, exposing for green then red must be exposed -1 stop and blue 2/3 stops, so a bracket of 7 with 1/3 stop interval will get a proper exposure for each colour channel. This should work, at least in theory, and possibly give better colour rendering than modifying the colour of the light source.

Post processing the negative

Either way, each colour channel will not be equally exposed. To correct for this, it is necesary to set the black and white point. This is best done sampling from the film containing unexposed parts as well as fully exposed parts, neither will be fully black or fully white. This can usually be found on the beginning of the film strip.

It becomes more complicated though, as depending on film type the result may vary greatly. I have shot LOMO 100 and 800 (LOMO branded films), Fuji Superia 100 and 200, Kodak ProImage 100, Kodak Supra 800, Color and Kodak Professional 100, and each may have a different colour cast that need to be corrected. I have photos from the last 20 years, and some colours may have faded. And on top of that, I have shot Fujichrome and Kodachrome for cross-development. Now, the cross-developed films does not have an orange cast, in fact, for Fujichrome it is green, and for Kodachrome it is transparrent as for the diapositives. So these need a different treatment.