[FFmpeg-user] Compressed audio streams (DTS, Dolby) decoding

[Carl Eugen Hoyos]

2018-01-20 11:16 GMT+01:00 Frédéric <fma at gbiloba.org>:

> I'm using a NUC to run Kodi, sending the audio signal to my HC amplifier
> through HDMI. I let the amplifier decode compressed audio streams, by
> activating the passthrough param in Kodi.
>
> But I plan to use a nanoAVR HD sound processor, to make some audio
> correction. This device uses HDMI as in/out. The problem is it can't
> correct compressed streams; it needs PCM streams (it can handle up to 8
> channels).
>
> So, the idea is to let Kodi - ffmpeg - decode the compressed audio streams,
> and send a multi-channels PCM stream to the amplifier, which nanoAVR will
> be able to correct.
>
> The problem is I have big differences in the audio image when I let ffmpeg
> decode the stream or if I let the amplifier does the job. For example, the
> balance between front and rear speakers is very strange with ffmpeg (way
> too much level on rear speakers).

Please confirm that this happens with a raw (encoded) audio sample
and the ffmpeg command line tool (use wav or raw pcm as output)
with current FFmpeg git head.
If yes, please provide the command line including complete, uncut
console output and provide the raw input audio file.

Carl Eugen