One of the most important lessons I learned from Peter Krogh's excellent multimedia book, Digitizing Your Photos With Your Camera and Lightroom, was to use filters to compensate for the orange mask of color negative film. I had read various pros and cons in forum discussions and other blogs, and even saw one direct comparison that claimed no benefit, but Peter makes a convincing argument for light source compensation, particularly in a Lightroom-only workflow. Based on that recommendation, I ordered a set of colored filters, and based on my own testing, I am now convinced of the benefits.
The most obvious benefit that I am seeing is for photos with lots of blue sky, but I think the benefits apply to all images. I shot a roll of Ektar last fall in Colorado. In addition to the gold and orange aspen leaves, every one of those images has a lot of clear blue sky. I originally digitized the roll using my LED light pad with plain white light, and I had to clean up the sky in every image in Photoshop because I was seeing very obvious blotches that ruined the photos. I just assumed that it was a problem with my exposure or with the processing by the lab that damaged the negatives. By the time I had finished cleaning up these images, I never wanted to touch a roll of film again. Look at the image below for an uncorrected example.
Digitized using white light source. Note lighter vertical streaks on right side of image. I recently re-digitized a few of those negatives using a set of colored filters on the same light source, and there is no blotchiness at all in the skies. I am also getting a nicer blue whereas the original scans had somewhat of a greenish tint in the blue.
Digitized with color compensated light source. Based on the explanation that Peter provides in his video, I think the problem I was experiencing might be a limitation of Lightroom/Camera Raw not having the latitude to correct the difference in the color channels resulting from the orange mask of the film base. When the negative image is converted from a raw file to a TIFF, that color information is baked into the file. When we apply the extreme color conversions needed to invert the negative and remove the mask in Photoshop, the color information is stretched too far causing the infidelities in the converted positive image. By using the filters to compensate for the orange mask in the raw capture, the required color corrections are not as extreme, so the end result is a much cleaner image. The image below is a 1:1 comparison that shows the difference in scan quality (please pardon the lack of focus-I obviously need more practice with the rangefinder).
Color filtered light on the left; white light on the right. Incidentally, in preparing the sample images for this comparison, I noticed a distinct difference in image quality between the image converted in Photoshop as a TIFF as compared to the image converted in Lightroom using a profile applied to the raw file. The converted raw file is much cleaner-see below. I did not apply any additional noise reduction to the image on the left. If you have not read my guide on converting negatives in Lightroom using profiles, please check it out.
Comparison of converted image. Lightroom processed raw on the left; Photoshop TIFF on the right. Peter Krogh recommends using a dichroic light source that can be tuned to neutralize the mask of each specific type of film. It is a great solution, but unfortunately I don't have one of those and they are large and expensive. I did, however, find a recommendation from John Fechner (fechnerimaging.smugmug.com) in a discussion forum for a combination of color gels that he found does a good job of neutralizing the orange mask. The combination of filters that he recommends are Rosco Cinegel #3202: Full Blue (CTB), Cinegel #3204: Half Blue (1/2 CTB), and Cinegel #4415: 15 Green. My experience so far is that this combination works really well.
Here's a comparison of a negative captured without and with the color correction gels. Note that straight out of camera, the negative on the right is already nearly white balanced and the orange tint is almost completely removed. The histograms below the image show the separation between the blue and red channels in the unfiltered negative. Even though these are raw images, the inversion and color correction includes some extreme adjustments to the color channels. By using the color filters, the raw image starts out more balanced and the end results are cleaner because less extreme adjustments to the red and blue channels are required. For a more detailed explanation of why this works, refer to Peter's excellent book.