- Posts: 38
- Joined: Thu Apr 21, 2016 4:52 pm
Thu Oct 27, 2016 8:35 am
I have a weird problem with some of my highlights turning black when i add a desaturation curve in Resolve 12.5.
This was shot 4.6k in 3.1 Raw. The stills are from a HD timeline. The only difference between the two stills you see is that I have added a lum vs sat curve to the shadows (see screen shot of curve). Obviously the desaturation to the shadows should not be affecting the highlights!
Link to DNG: DNG: https://www.dropbox.com/s/c1lxp3d9lg85f ... m.dng?dl=0
Someone on the facebook group said this:
"The problem seems to be the blue and purple lights are severely out of gammut, and the 4.6K (or maybe 3:1 compression specifically) doesn't know how to handle these values. If you lift the black levels in Resolve and look at the RGB Parade, you can see that there are values far below black in both the Red and Green channels for some reason.
If it's any consolation, if you run it through the BMDFilm4.6K to Linear LUT, then the Linear to BMDFilm4.6K LUT, you can do your saturation anywhere after in the signal chain. Running through these LUTs seems to truncate these problem values."
My experience is also that adding that LUT earlier in the node tree stops the problem. However I would like to able to add a desaturate curve when ever I feel like it. Putting it before a LUT is something I do semi-frequently.
This was shot 4.6k in 3.1 Raw. The stills are from a HD timeline. The only difference between the two stills you see is that I have added a lum vs sat curve to the shadows (see screen shot of curve). Obviously the desaturation to the shadows should not be affecting the highlights!
Link to DNG: DNG: https://www.dropbox.com/s/c1lxp3d9lg85f ... m.dng?dl=0
Someone on the facebook group said this:
"The problem seems to be the blue and purple lights are severely out of gammut, and the 4.6K (or maybe 3:1 compression specifically) doesn't know how to handle these values. If you lift the black levels in Resolve and look at the RGB Parade, you can see that there are values far below black in both the Red and Green channels for some reason.
If it's any consolation, if you run it through the BMDFilm4.6K to Linear LUT, then the Linear to BMDFilm4.6K LUT, you can do your saturation anywhere after in the signal chain. Running through these LUTs seems to truncate these problem values."
My experience is also that adding that LUT earlier in the node tree stops the problem. However I would like to able to add a desaturate curve when ever I feel like it. Putting it before a LUT is something I do semi-frequently.
- Odd problem_1.jpg (483.89 KiB) Viewed 413 times
- Odd problem_2.jpg (338.48 KiB) Viewed 413 times
- Odd problem desaturation.png (170.48 KiB) Viewed 364 times
