How to use ffmpeg scale_cuda to implement 10bit transcode?

Gaming and Visualization Technologies Video Processing & Optical Flow
1

GPU : Quadro P2000
NVIDIA Driver Version: 387.22
CUDA Version : 8.0
FFMPEG version : 3.4.1
OS : CentOS7.3

Compile command line:
./configure --enable-static --disable-shared --enable-cuda --enable-cuvid --enable-nvenc --enable-libfdk-aac --enable-nonfree --enable-libnpp --enable-cuda-sdk --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64

How do I use it.

Here is my current command line:
/usr/local/ffmpeg -hwaccel cuvid -c:v h264_cuvid -i udp://238.1.1.1:62000?pkt_size=1316&overrun_nonfatal=1&fifo_size=658000&buffer_size=658000&timeout=3000000 -vf scale_npp=1920:1080 -c:v nvenc_h264 -preset slow -b:v 2000k -maxrate 2000k -minrate 2000k -bufsize 2000k -g 50 -c:a aac -b:a 128k -f mpegts udp://239.1.1.1:1234?pkt_size=1316&ttl=255&overrun_nonfatal=1&fifo_size=658000

2

thank you

3

There are FFMPEG forums where you can get help on FFMPEG usage. This forum is for nVideo SDK stuff.

4

As far as I know, scale_npp in ffmpeg does not support 10bit sclae.

5

Any solution for this ?

6

It is now supported. avfilter/scale_cuda: fix pitch calculation for >8 bit formats · FFmpeg/FFmpeg@a1553b0 · GitHub

1 Like