AV1

[humanoid86]

The AV1 codec has become very popular.

I hope you will add "NEW" codec/filter ("drivers") with AV1 support?

[Selur]

av1 encoders Hybrid currently supports atm.:

• aomenc
  
• rav1ve
  
• svt-av1
  
• QSVEnc (requires Intel GPU; only in dev build available atm.) (Supported on Windows only atm.)
  

I plan to add av1 support through NVEncC once I get my hands on one of the NVIDIA 40xx cards, but both the 4090 and 4080 cards are way out of my budget range (unless I win the lottery).

This should already cover every free available av1 encoder I know of.
=> If you can name additional av1 encoders that are free, have a command line interface and supports piped input, I can look into supporting them.

Cu Selur

[humanoid86]

you didn't understand me again

fresh versions needed

not these old ones - they are experimental - not stable

[Selur]

Dev version should be up-to-date, but you can simply replace the binaries if the command line interface didn't change.

Cu Selur

[humanoid86]

Dev version should be up-to-date, but you can simply replace the binaries if the command line interface didn't change.

Cu Selur

Would a simple file replacement help?

without adjustments in the settings?

O_O.........

[Selur]

Yes

[karthauzi]

Anyone tried the --fp-mt=1 usage for better Threading yet? Will it speed up encoding a bit 6Core /12 Threads? Will the Video Quality drop like settings up tile columns and rows in AOM? I dont want to use chunked encoding yet, because i only have 16 GB of RAM, when i use chunked encoding with filters on, 2 gb ram left, vspipe will eat all of my RAM.

Greetings

[Selur]

I only did a few tests with SD and 4k content and couldn't find any speed gain when using --fp-mt=1, not sure whether there are any other requirements for it to do something. (Side note: when using chunked encoding, make sure to lower the threads count)

Cu Selur

[karthauzi]

Hi Selur, why lower the Threads count? even if i set full threads, AOM is not using all of my thread to 100% CPU Load is pending about 80%, with 2 chunks and still 12 Threads, i can get 100% cpu load all the time.

[Selur]

More threads mean more memory usage and overhead.
If you run multiple encodes in parallel, you might want to lower the per job threads count.