[FFmpeg-devel] ./configure and nvidia/cuda sdk

[Anton Khirnov]

(I am not a lawyer, the following is my personal opinion and not legal
advice)
Quoting Top Secret (2024-09-11 14:21:21)
> Hello,
> 
> We are trying to redistribute a ffmpeg binary commercially. We use free version for generally decoding video only(some demuxing). We want to use nvidia's suggestion of using the GPU with ffmpeg. They give a sample configure string like so:
> ./configure --enable-cuda --enable-cuvid --enable-nvdec --enable-nvenc --enable-nonfree --enable-libnpp --extra-cflags=-I/usr/local/cuda/include  --extra-ldflags=-L/usr/local/cuda/lib64
> 
> In here we notice --enable-nonfree which basically means we can't give ffmpeg commerically right?

No, AFAIU it means you cannot distribute such a build at all, since
there is no well-defined license for it.

> Well looking into this legal wise... If we get NVidia's written consent that a lawyer can use in a court of law, alter configure script to say libnpp is not considered "non-free" and ditch that string would it be okay?
> 
> NVidia seems to advocate for FFmpeg use. If its not permissible, then would it make sense to just use opencv's implementation of ffmpeg which then advertises a completely different license(Apache).

If you want a redistributable build with NPP, Nvidia would have to
relicense NPP to something LGPL-compatible.

-- 
Anton Khirnov