Max bitrate setting is not working about immich HOT 5 OPEN

This seems correct to me. The unit is different in each of those. You requested 4000 Kilobytes and are getting results in Megabits which are a different unit. 4000 Kilobytes is 32 Megabits and you are well within that range based on what you've shown us.

Bits v Bytes

from immich.

This seems correct to me. The unit is different in each of those. You requested 4000 Kilobytes and are getting results in Megabits which are a different unit. 4000 Kilobytes is 32 Megabits and you are well within that range based on what you've shown us.

Bits v Bytes

This setting is for "max bitrate". So the unit should be kilo bit.
And this setting also appear at ffmpeg command line:
/usr/bin/ffmpeg -hwaccel qsv -hwaccel_output_format qsv -async_depth 4 -threads 1 -i /media/photos/xxx.mp4 -y -c:v h264_qsv -c:a aac -movflags faststart -fps_mode passthrough -map 0:1 -map 0:0 -bf 7 -refs 5 -g 256 -v verbose -vf scale_qsv=-1:720:async_depth=4:mode=hq:format=nv12 -preset 1 -global_quality:v 5 -maxrate 4000k -bufsize 8000k xxx.mp4
The ffmpeg manual also indicate the unit is kilo bit.
I add -b:v 1000k argument to ffmpeg command line and calculate the bitrate of output file, it shows 1000k is 1000kilo bit. It's not reasonable that the unit of -b:v is different from -maxrate in the single command line.
So I request for 4000kilo bits here and immich also treat it as 4000 kilo bits on ffmpeg command line but treat it as 4000 kilo bytes on examing the orinal video file?

from immich.

There are two things that could be causing this problem:

1. The final bitrate is including the aac which you are not transcoding into something smaller like opus
2. QSV isn't compliant with all the filters and other flags you've set and is misbehaving. I don't have access to anything with QSV enable and generally don't use quicksync so I can't test this on my machine either. Sorry.

I have always had problems with NVENC and more complex transcodes and just don't bother. Also, I understand that ffmpeg likes two pass encoding to achieve a desired max bitrate.

from immich.

I have taken a quick look at some of the documentation and there is no mention of support for ffmpeg filters in QSV but the ffmpeg wiki seems like it should include support for scaling and most of the other filters you are using.

from immich.

There are two things that could be causing this problem:

1. The final bitrate is including the aac which you are not transcoding into something smaller like opus
2. QSV isn't compliant with all the filters and other flags you've set and is misbehaving. I don't have access to anything with QSV enable and generally don't use quicksync so I can't test this on my machine either. Sorry.

I have always had problems with NVENC and more complex transcodes and just don't bother. Also, I understand that ffmpeg likes two pass encoding to achieve a desired max bitrate.

I'm not talking about transcoding it self but the decision of if transcoding or not.
I think the video should be transcoded instead of just copy the video track.
Here's the command line of ffmpeg:

/usr/bin/ffmpeg -init_hw_device qsv=hw -filter_hw_device hw -i /media/photos/小区相关/业委会/00001.MTS -y -c:v copy -c:a aac -movflags faststart -fps_mode passthrough -map 0:0 -map 0:1 -bf 7 -refs 5 -g 256 -v verbose -preset 7 -global_quality:v 23 -maxrate 4000k -bufsize 8000k upload/encoded-video/0ef79aea-c9d6-4f20-b7d2-bc3a01408cb6/57/7f/577f0c7e-cb57-45ec-85e4-bacb04c4b2e0.mp4

from immich.