Don't Clip the Red Channel

Hello All,

Recently I made a breakthrough in my own process, and that is to use the RGB histogram when scanning and make sure to not clip the red channel. When I avoid that clipping, the conversions in NLP are so much better! Many times they are just right out of the box, sometimes they need a little work. But at least I have something to work with! (Maybe everyone already knows this but I wasn’t able to find a direct reference, so I feel obligated to share.)

Backup information:

Over time I had several real problems (poor light source, bad method of holding the negative, etc.) but about a year ago these were all resolved with this setup: Canon 6D, 180mm macro, copy stand, Kaiser Slimlite Plano light source and Negative Supply holders and scanning in the dark (except for the light of the computer screen as I tether with the EOS Utility).

Even with this setup, my scans were all over the place. Sometimes LR would not be able to set the color balance when I clicked the edge rebate with the eyedropper, most of the time the NLP edited scan would have really weird colors in it, and no amount of work I could do in NLP was able to fix it (that last part could be me though).

I was exposing to the right, seriously. To try something I hadn’t before, I started dropping the overall brightness to around the ‘middle’ of the histogram, at least not letting any part go too far from the middle. This yielded better results, although the scans were very bright.

Then, somewhere I read that clipping the red channel can be a problem, but it was a stray comment without a lot of background information. But that got me thinking, so on the next try I switched the histogram to RGB and made sure that the red was not clipped. That made a huge difference! Now I’m almost always getting something that can be sorted out.

The only downside to this practice is that the scan tended to be very bight, due to the blue channel being underexposed. Yesterday I tried sliding a 1/2 CTB gel for a speedlite under the Negative Carrier MK1. While I don’t see a difference in the color of the scan, the brightness is a lot better, e.g. not super bright. The negative aspect of this setup is lengthening of exposure times.

I hope this all makes sense! I’m so happy to be able to get a handle on color negative scans, so now shooting, developing and scanning are all within my grasp. Maybe one day RA4 printing too…

It is certainly a good idea to not clip any channels. But I’m not sure if the red channel is the one that is most important - at least not on a Canon EOS 5D Mark III. Your mileage may vary.

I mostly shoot negatives with the camera set to a UniWB (more about it in the forums or the internet) and find that the channel, that will clip first, is the green channel. I ETTR, but never up to the maximum, but to about 2/3 stops below clipping. This is where Canon says that the sensor’s linear top end is.

Here’s a shot of a negative taken on Kodak around X-mas 1987 in Australia as seen in RawDigger and the respective raw histogram.

I find this page to be helpful for a first contact with UniWB.
An other interesting procedure to get UniWB is described here.