[Cin] Testing Nvidia Nvenc encoding

[Terje J. Hanssen]

Den 22.12.2024 00:04, skrev Andrew Randrianasulu:
>
>
> вс, 22 дек. 2024 г., 01:53 Phyllis Smith <phylsmith2017 at gmail.com>:
>
>     Andrew,
>
>         most likely our nv headers drifted from that
>         ffmpeg-7.0/proprietary driver assumes at runtime.
>
>     I have been wondering about nv-codec-headers as we are at:
>     https://github.com/FFmpeg/nv-codec-headers/releases/tag/n10.0.26.0
>     but I am unsure about updating to:
>     https://github.com/FFmpeg/nv-codec-headers/releases/tag/n12.2.72.0
>     because if you look at:
>     https://github.com/FFmpeg/nv-codec-headers/releases/
>     the release versions go from 12.xx to 8.x and it is really weird
>     AND there is no year on the release dates but just day and month.
>     Since it is such an important part of ffmpeg inside CinGG, I am
>     concerned but will at least try the 12.2.72.0 just to see what it
>     does.
>
>
> in theory it should give users of new nvidia hardware av1 encoding 
> ...... but not sure how it will work with older drivers and hardware.
>
>
>
>         you can try to install something like nv-codec-headers and
>         then add
>

Andrew,
As far as there might be a workaround also for the nvenc tff interlace 
issue, I didn't do more about the latter than searching the most similar 
package
ffnvcodec-devel (FFmpeg version of NVIDIA codec API headers)

Additional I think an AppImage built successful of my dynamic build 
ffmpeg-7.1 with
sh ./bld_appimage.sh bin_use_system_ffmpeg-71
It works on the build-machine, and I will test it on the older machines too.

Phyllis,
I was about to send a little comment to your first News version, 
regarding relative "new" Intel hardware.
The SkyLake/ KabyLake test machines are from 2015/ 2016 respectively :)
Of course they have lesser codecs support than the relative new 
bult-machine.



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cinelerra-gg.org/pipermail/cin/attachments/20241222/8f2633b2/attachment.htm>