shutter-encoder icon indicating copy to clipboard operation
shutter-encoder copied to clipboard

[Bug Report] Incorrect audio codec and bitrate when converting from videos with multiple audio tracks

Open daoxi opened this issue 3 months ago • 5 comments

Version tested: v19.4 OS: Windows 10 22H2 (64bit)

First of all, here's how you can reproduce the bug:

Let's say I have a mkv video with two audio tracks: # 1 and # 2, and I only want to export the video with audio track # 2, so in the software I set it up like the following screenshot:

Image

Notice how A1 is set to "No audio" and A2 is set to "Audio 1", and when I convert and export the video, I found something weird: The audio track in the converted video has incorrect codec and bitrate: although I set it to be AAC 256kbps, the converted audio turned out to be either AAC ~128kbps (when exporting as MP4) or Vorbis ~100kbps (when exporting as MKV), regardless of what format/bitrate I set the Audio Settings to be.

My suspicion is that something happened when the software detected A1 is set to "No audio" (this bug only happens when I do this) something unexpected happened, which caused the A2 audio settings to be overridden by internal default settings (this could be coming from either the software or FFmpeg).

Let me know if you have any questions or need any further clarifications

daoxi avatar Sep 29 '25 19:09 daoxi

You can't use A2 if A1 is set on 'No audio', if you only need the 'audio 1' you to set A1 to 'audio 1', if you need the 'audio 2' set A1 to 'audio 2'.

Paul.

paulpacifico avatar Oct 02 '25 17:10 paulpacifico

Ah I see, I originally thought "A1:" and "A2:" are for Audio 1 and Audio 2 from the original video should be mapped to which audio number in the converted video (that's why I chose "Audio 1" for "A2:" in the screenshot, as I tried to map Audio 2 from the original video to Audio 1 in the converted video), but turns out it's actually the other way around (i.e. "A1:" and "A2:" refer to the audio number in the converted video, not the original video).

That makes sense now, thanks a lot.

Since the user is supposed to add in audio numbers in a consecutive manner (as you explained: you shouldn't use A2 if A1 is set to no audio), maybe it would be beneficial to add some UI hints to reflect that? (setting "A1:" to "No audio" already will automatically set all audios below it to "No audio" too, maybe it'd better to just disable/grey-out all audios below it instead?)

daoxi avatar Oct 03 '25 06:10 daoxi

Sorry I have encountered another issue about audio selection,

So I have a mkv video with two audio tracks, audio 1 is AAC and audio 2 is FLAC, when I do this: Image the converted video contains two AAC audio tracks, as intended.

but when I do this (without changing anything else): Image the converted video somehow only contains audio 1, but not audio 2

I thought the settings in both screenshots are equivalent to each other? So why did one of them not work as intended? Am I missing something? (or possibly a bug?)

daoxi avatar Oct 03 '25 09:10 daoxi

No they are not the same, the first one use the same mapping as source. The second is downmixing track 1 & 2 to create the stereo.

If you want to use one audio codec for all, select the 'Multi' mode from the second screenshot. Paul.

paulpacifico avatar Oct 03 '25 13:10 paulpacifico

Ah I see, thanks for the clarification, for now I'll just stick with the "Custom" in the first "Convert:" dropdown menu (instead of "Multi" in the second dropdown menu), because the "Custom" mode offers more flexibility to have different codecs.

Really appreciate all your help so far!

daoxi avatar Oct 03 '25 18:10 daoxi