Errors on DTS audio tracks

[Herzfrequenz21]

Errors on DTS audio tracks

[Purfview]

Not same, now the error is:

  File "av\\audio\\fifo.pyx", line 86, in av.audio.fifo.AudioFifo.write
ValueError: Frame does not match AudioFifo parameters.

Before it was:

  File "av\\codec\\context.pyx", line 416, in av.codec.context.CodecContext._send_packet_and_recv
  File "av\\error.pyx", line 336, in av.error.err_check
av.error.ValueError: [Errno 22] Invalid argument; last error log: [dca] Residual encoded channels are present without core

[Purfview]

Errors shouldn't be different between Faster-Whisper and Faster-Whisper-XXL, probably you run now on the different audio.
Btw, both these are standalone.

Remux problematic audio removing other streams and upload it to https://wetransfer.com or mega, so I can look what is wrong there.

[Herzfrequenz21]

Yeah, I guess so. I tried all three files - they all failed. I don't remember which ones and when exactly.
The files are too big. I made a torrent. Can you download it from me?
By the way, I didn't try to recognize audio files separately. So, if you suddenly succeed, maybe this is what will affect it.

[Purfview]

Looks like it downloads.

[Purfview]

Reproduced error on ~700MB dts file, it comes from internal PyAV lib which is based on ffmpeg.
No error if external ffmpeg is used, you can trigger external ffmpeg use by any --ff... arg, for example --ff_dump, this command by itself doesn't make any filtered change on audio.
Or with Faster-Whisper-XXL you can use the vocal extraction with --ff_mdx_kim2.

3GB file I'll check later.

[Purfview]

Same for 3GB DTS-HD file, if you trigger external ffmpeg use then it works.

I tried directly those files on ffmpeg, same errors reported there but it doesn't stop the process when PyAV breaks it.
Maybe those files are damaged as I tried some other DTS files and there were no such problems.