2023-10-11, 05:18 PM
(This post was last modified: 2023-10-11, 05:18 PM by bitmap. Edited 1 time in total.)
For the 7700k, I would uncheck both LPE options as I don't believe they're supported at all. I would also uncheck HEVC encoding because that processor is not particularly capable of HEVC encoding and you'll rarely find clients that request it anyway.
For codec support, I'll have to see if I can find the matrix that TheDreadPirate pointed me towards (I think on Wikipedia) at one point, but there's probably some of those that your processor isn't compatible with. I don't think that's the issue here, but proper config now will save headaches later.
Seems like QSV just isn't configured properly. I like to specify the card directly on my compose file with both render and card calls, but you didn't necessarily have to. It gives me specificity of what device I'm passing into the container and specificity of what I can use later inside the container. Also used to be the only way I could get it to work, so...old habits die hard...
You should also ensure that the UID you've specified as owning this container has access to the groups that own the devices you're passing in. Generally they're render and video but I'm not sure on Unraid.
I'm just kind of re-hashing basic steps from the docker HWA guide that seem to trip people up a lot.
https://jellyfin.org/docs/general/admini...ualization
An ffmpeg log of unsuccessful, as well as successful transcoding would help. Probably on pastebin or sourcebin (dunno the difference, TDP recommended the latter so maybe they've got more insight).
EDIT: Also, why no cache directory mounted?
For codec support, I'll have to see if I can find the matrix that TheDreadPirate pointed me towards (I think on Wikipedia) at one point, but there's probably some of those that your processor isn't compatible with. I don't think that's the issue here, but proper config now will save headaches later.
Seems like QSV just isn't configured properly. I like to specify the card directly on my compose file with both render and card calls, but you didn't necessarily have to. It gives me specificity of what device I'm passing into the container and specificity of what I can use later inside the container. Also used to be the only way I could get it to work, so...old habits die hard...
Code:
devices:
- /dev/dri/renderD128:/dev/dri/renderD128
- /dev/dri/card0:/dev/dri/card0
You should also ensure that the UID you've specified as owning this container has access to the groups that own the devices you're passing in. Generally they're render and video but I'm not sure on Unraid.
I'm just kind of re-hashing basic steps from the docker HWA guide that seem to trip people up a lot.
https://jellyfin.org/docs/general/admini...ualization
An ffmpeg log of unsuccessful, as well as successful transcoding would help. Probably on pastebin or sourcebin (dunno the difference, TDP recommended the latter so maybe they've got more insight).
EDIT: Also, why no cache directory mounted?
Jellyfin 10.10.0 LSIO Docker | Ubuntu 24.04 LTS | i7-13700K | Arc A380 6 GB | 64 GB RAM | 79 TB Storage