Hello. I trying to encode an x264 mkv to h265 mkv with NVenC. It works well but I wonder if it's possible to either skip the audio extraction job and convert the audio directly to new file or simply be able to extract the audio from the original file while also encoding the x264 to new video and simply merge them after they're both done.
As you can see in the image. Sub Job 01_audio started extraction and the next subjob is converting. That, obviously needs to wait. But is it possible to process 05_video and convert video while also extracting the audio?
Short: Not in Hybrid.
Longer:
In theory: Yes
In Hybrid: No, since Hybrid isn't designed to do this. Adding this would require a rewrite of the job creation and the job processing, since it would need to add functionality to:
a. mark that subjob X depends on subjob Y in the job creation (and the necessary logic that in any possible scenario could properly set these marks)
b. a job processing which would honor such marks.
-> I doubt this will come any time soon, if this is a real problem for you, you should look at ffmpeg or handbrake based tools, since they tend to do everything in one pass.
As a side note: Hybrid did at some point in the past try to do the extraction and encoding of audio in one go, but that often caused sync problems and harder to trace errors which is why I dropped that. Nowadays audio first gets extracted and then reencoded.
Cu Selur
Thanks for the fast reply.
The problem for me is that Audio extraction takes about 2-3 more time than audio convertion + video convertion + remuxing the files into one mkv container. Would changing source and final result location to 2 different HDD's boost the speed? As I notice it's mostly limited by "Disk Usage" being up to 99% most of the time of the extraction.
Or do you have any other workaround or recommandation for extracting faster besides what I suggested? [ I don't want to shred the lifetime of my SSD just yet so that's out
]
Personally I but my source, temp and output folder on different drives since it does speed up muxing and extraction jobs.
Other than that, like I wrote using another tool which does the workload in a single process might speed up things a bit.
Seeing that you are not doing any Avisynth or Vapoursynth based filtering you might even be able to do your audio&video processing in one go using just NVEncC which also supports audio processing (software based only).
Cu Selur