Andrew, when I check into Git the onevpl patch, I would like to checkin Terje's final tested 3 hevc_qsv render formats.  Even though there will be relatively few users who not only have the Intel hardware to use these as well as their own build, it seems like a valuable addition.  Any objections?

On Wed, Nov 13, 2024 at 5:17 AM Terje J. Hanssen via Cin <cin@lists.cinelerra-gg.org> wrote:

After finishing the three hevc_qsv presets for _8b420, _10b420 and _10b422 mp4
https://lists.cinelerra-gg.org/pipermail/cin/2024-November/009039.html

I thought to do the corresponding for hevc_vaapi

However I am struggling to get 10bit 422 output rendered from Cingg, which so far falls back to 10bit 420 chroma subsampling

fprobe -hide_banner hd01_hevc_vaapi_10b422.mp4
  Duration: 00:01:11.20, start: 0.000000, bitrate: 11082 kb/s
  Stream #0:0[0x1](und): Video: hevc (Main 10) (hev1 / 0x31766568), yuv420p10le(tv, bt709/unknown/unknown, top coded first (swapped)), 1920x1080 [SAR 1:1 DAR 16:9], 11080 kb/s, 25 fps, 25 tbr, 12800 tbn (default)


In a previous thread I proved that ffmpeg hevc_vaapi encoding has this 10bit-422 capability
3) 10-bit yuv422p10le hevc_vaapi encoding using -vf 'format=y210,hwupload'
https://lists.cinelerra-gg.org/pipermail/cin/2024-November/008982.html


And yes, there was also issues before hevc_qsv_10b422.mp4 combinations Pixel format/profile finally succeeded.
But here are some differences, yet. While hevc_qsv provided a Pixels drop down menu with pixel formats,
hevc_vaapi just has a single "vaapi" pixel format.

Any idea why, and what can fulfill the vaapi 10bit 422 encoding here?

--
Cin mailing list
Cin@lists.cinelerra-gg.org
https://lists.cinelerra-gg.org/mailman/listinfo/cin

--
Cin mailing list
Cin@lists.cinelerra-gg.org
https://lists.cinelerra-gg.org/mailman/listinfo/cin