Replies: 3 comments 4 replies
-
This is CPU usage during "Slow" preview generation on 5950X |
Beta Was this translation helpful? Give feedback.
-
The SLOW mode on this video file processing time: Started: 13.45 Finished: 17.37 Total: 3h 53m Preview is more or less same size and resolution and still not that smooth in operation.
|
Beta Was this translation helpful? Give feedback.
-
I did a test with changing codec to vpx-vp9, due to license uncertainties around GPL. I think this is why it's slow. Because vpx is a pretty slow codec. Will change back to x264, which should use all cores and be a LOT faster. (this can be confirmed by testing an older version of losslesscut.) Planning on changing lossless-cut to GPL also, so that licenses are in order. |
Beta Was this translation helpful? Give feedback.
-
Converting preview is not very efficient. Only utilizing maybe 2 cores only (looks like 4 threads).
I want to use to chop up relevant parts of some 4K security footage which is in HEVC format as below
Doing the "high quality + audio" preview option, I can see 5950X utilisation of only 4 threads out of 32 possible and 16-18% usage only.
So converting 4K 2 hour footage is slow even on 5950X.
Is there not a multi-core FFMPEG available? This sort of job would be much better suited offloaded to an NVidia GPU to really speed it up. Or even better, native HEVC playback and preview using GPU decoding in the program?
Beta Was this translation helpful? Give feedback.
All reactions