Jump to content

Negative Pixel Values, ACES, and & DaVinci Resolve


AJ Young

Recommended Posts

I ran into a problem concerning Negative Pixel Values and ACES 1.1 (ACEScct) on a project in DaVinci Resolve 17. With the help of Tim Kang, we found the solution to my problem: incorrect pipeline set-up. Additionally, Elvis Ripley, Paul Curtis, and Nick Shaw gave incredibly valuable insight about this issue on Cinematography Mailing List.

For posterity's sake, below are the problem, solution, and insight:

Problem: Negative Pixel Values on an image in DaVinci Resolve 17 using ACES 1.1 (ACEScct). To over simplify what a negative pixel value is, essentially they're pixels that go below 0 but appear as bright and distinct noise. They can happen because pixel values (RGB) are floating point in color grading programs and somewhere along the line those negative values are displayed as positive, showing up as noise. Here's a screenshot of what negative pixels can look like:

Negative-Pixels.jpg.f29bebec4edb0081dacaee4c3f8ae72c.jpg

Solution: In this particular case, I had an incorrect pipeline set up.

The movie files (clips) are 10 bit Apple ProRes 422 from the Panasonic GH4 in vLog/vGamut. My first mistake was having DaVinci automatically choose the data levels of the file. Apple ProRes can either be legal (video) or full range data depending on the camera. Sony, Canon, and Panasonic cameras are full data ranges while ARRI logC is legal (video). However, DaVinci automatically chooses legal data ranges, and my Panasonic video file was being used incorrectly by the program.

My second mistake was using an incorrect ODT (Output Device Transform). Because I'm coloring this project on a traditional computer monitor (circa 2019), the ODT should be sRGB. I had my ODT set to Rec709.

Both of those mistakes combined created my particular negative pixel problem. Once I corrected them (set data range to full and ODT to sRGB), my negative pixel values disappeared. However, there may be an instance where this won't fix the problem because there are times a negative pixel value can occur with a correct pipeline.

One solution is to use a gamut compress created by the ACES community and available here: https://github.com/jedypod/gamut-compress

---

For anyone who stumbles on this post in the future, I hope it helps! To those who helped me fix this problem, thank you!

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...