I am trying to set white balance as instructed in Lightroom prior to negative conversion. I am selecting the light border of the 35mm negative as directed, however, I am frequently getting the message unable to set white balance. Any suggestions would be welcome.
Alternative WB setting is AutoWB. I find that it works too, even the original WB can work. Nevertheless, results will differ and you’ll see, which interpretation you like best. Here’s a quick test that I did (using virtual copies of one shot) for illustration:
The images are, clockwise from top left
- Original raw, cropped (I use UniWB for ETTR, hence the green tint)
- Conversion, WB taken from between images
- Conversion using Lightroom’s AutoWB
- Conversion without WB
As you can see, I did not even crop off the borders in shots 3 and 4, instead, I increased border tolerance to 15%. From other tests, I found that Negative Lab Pro was fairly tolerant to exposure as long as I kept the histogram within bounds.
The settings I use normally are “Basic”, “Linear+Gamma”, and whatever “Film” and “Color” settings get me closest to what I want. If I must, I’ll use the sliders for contrast and brightness. If I want further processing, I have Lr make a TIFF and go from there.
Based on the above, I propose that you try AutoWB as a first step.
You can also white balance your image from a shadow area.
Maybe you overexposed, meaning there is too much clipping in the raw in the unexposed film, meaning you can’t set white balance there.
I used a dcraw command line to make a tif from my raw shots, without using the white balance that the camera stored. That way, you get raw, unaltered (but demosaiced) linear pixels.
I try to get that file as close to clipping (but not clipped in any channel) as possible (looking at the film image and unexposed filmstrip. The light through the sprocket holes is unimportant).
Keep trying taking shots till you have one at good exposure, then use the same exposure for all the shots of a roll, and throw the raws into Lightroom for NLP.
My solution to the same problem is to find a negative from that role that the eye dropper works on without issue and then just synchronize the settings.
My suggestion: Your exposure is likely too high for sampling the rebate. Suggest 1) shoot the rebate at a lower exposure, the 2) shoot all images on the roll at your usual exposure. 3) Use rebate from first image to set WB for the whole roll. I’m shooting color negatives at +1 or +1.5EV with no clipping.
Thanks for the feedback. Using AutoWB with the Negative Pro profile applied seems to be working fine.
My original post was regarding scans of 35mm negatives that were shot from the late 70s through the early 2000s. As far as the exposure goes, they are normally exposed.
just to make things clear(er), exposure can matter when we photograph the negatives. Overexposing the camera scan can help get cleaner bright areas like clouds in the sky - as long as the highlights don’t clip. Nevertheless, I find NLP to be fairly tolerant to exposure.
Exposure of the original negative is what it is and cannot be changed. NLP will make best use of it and Nate gave us useful hints in this post: Terrible Colour
Yes, I too find NLP tolerant to exposure changes. For camera-scan of color negatives, I find +1EV helps and is still quite save (vs. clipping). FWIW, after Import to Lightroom, I set WB, then adjust the exposure slider to center the histogram.
Does it make any difference if you change exposure or not?
What I see in the test above is that NLP equalizes exposure anyway, at least with exposures within a approx. 2 EV range, provided that both ends of the histogram are within bounds. The shots were made with exposure correction ranging from 0 to 1.67 EV. Some slight shifts in colour can be seen though.
I see some differences when the histogram shown in LR, after WB, it way up high. As long as not up near RH edge, I agree with you, that NLP is insensitive to exposure differences.
My current practice:
- Expose +0.5 or +1EV
- Import, WB
- Then move the exposure slider a bit down
Indeed. Although CMOS sensors are thought of as linear devices, exif data in my EOS 5D3 files indicate a “linearity upper margin” (see screenshot of exiftool output in Graphic Converter)
In order to see how far right I can expose below the linearity limit, I took a series of shots of the backlight and looked at them in RawDigger. The limit was reached when exposure was at +3EV. I can now set exposure manually at these settings and scan the films knowing that dark areas will not be compressed. I could even go a little bit higher because the film base and mask swallow some light. I just have to remember the location of “10’000” is in the histogram.