vapoursynth-chromanrcuda

FFmpeg's chromanr for VapourSynth.
This version is made with cuda, if you don't have an nvidia GPU, check out the CPU version.

Usage

chromanrcuda.CNR(vnode clip[, float thres=4, float threy=20, float threu=20, float threv=20, int sizew=3, int sizeh=3, int stepw=1, int steph=1, int distance=0])

Parameters:

• clip
  A YUV clip to process (float only).
• thres
  Set threshold for averaging chrominance values.
  Sum of absolute difference of Y, U and V pixel components of current pixel and neighbour pixels lower than this threshold will be used in averaging.
  Luma component is left unchanged and is copied to output.
  Default value is 4. Allowed range is from 1 to 200.
• threy/threu/threu
  Set Y/U/V threshold for averaging chrominance values.
  Set finer control for max allowed difference between Y components of current pixel and neigbour pixels.
  Default value is 20. Allowed range is from 1 to 200.
• sizew
  Set horizontal radius of rectangle used for averaging.
  Allowed range is from 1 to 100. Default value is 3.
• sizeh
  Set vertical radius of rectangle used for averaging.
  Allowed range is from 1 to 100. Default value is 3.
• stepw
  Set horizontal step when averaging.
  Mostly useful to speed-up filtering, if > 1 it will skip some columns in averaging.
  Default value is 1. Allowed range is from 1 to 50.
• steph
  Set vertical step when averaging.
  Mostly useful to speed-up filtering, if > 1 it will skip some rows in averaging.
  Default value is 1. Allowed range is from 1 to 50.
• distance
  Set distance type used in calculations.
  0 (manhattan): Absolute difference.
  1 (euclidean): Difference squared.

Building

cmake -S . -B build -DCMAKE_BUILD_TYPE=Release -DCMAKE_CUDA_FLAGS="--threads 0 --use_fast_math -Wno-deprecated-gpu-targets" -DCMAKE_CUDA_ARCHITECTURES="50;61-real;75-real;86"

cmake --build build --config Release