What downmix algorithm does Poweramp use?

[Neo2447]

I was listening to a song that is Dolby Atmos in Poweramp and it sounded pretty good. However, when I downmixed the song with ffmpeg using the -ac2 command I noticed that the song was too low in volume. Does Poweramp use some special algorithm to downmix and that's why it sounds better when playing the song with it?

[Fitzian]

There's no open source decoder for Dolby's AC-4 or Dolby Digital Plus (EAC3) lossy atmos delivery methods.

So Poweramp is certainly not doing anything particularly special with whatever multichannel file you've extracted. If it did, it would be subject to Dolby Laboratories licensing.

The volume behavior is odd, though. Atmos mixes in general are supposed to be produced quieter overall, or just given more headroom for processing.

[Neo2447]

46 minutes ago, Fitzian said:

There's no open source decoder for Dolby's AC-4 or Dolby Digital Plus (EAC3) lossy atmos delivery methods.

So Poweramp is certainly not doing anything particularly special with whatever multichannel file you've extracted. If it did, it would be subject to Dolby Laboratories licensing.

The volume behavior is odd, though. Atmos mixes in general are supposed to be produced quieter overall, or just given more headroom for processing.

I don't think so, because then wouldn't be impossible to modify those files with ffmpeg?. Also, when playing those type of files with Poweramp it shows "Built-in ffmpeg decoder"

Edited March 3 by Neo2447

[Fitzian]

5 minutes ago, Neo2447 said:

But if so, then how is it possible to downmix those types of files using ffmpeg?

I stand corrected, as of last December there's been some activity on this front it appears.

[Neo2447]

6 minutes ago, Fitzian said:

I stand corrected, as of last December there's been some activity on this front it appears.

Yes. Apparently Poweramp uses ffmpeg to do the downmix but what I find curious is that the result ends up sounding at a fairly decent volume but when trying to do the process manually with the -ac2 code on ffmpeg the result has a fairly low volume.

Edited March 3 by Neo2447

[Fitzian]

By that, I mean, explicit reference to

> Raw AC-4 muxer and demuxer.

There's no reference to "we decode dolby atmos now" though...

That's not something Poweramp advertises. And also, since there's no legal resource for AC-4 atmos content, this is all kinda on the outskirts of what we can really discuss.

[Fitzian]

19 minutes ago, Neo2447 said:

when playing those type of files with Poweramp it shows "Built-in ffmpeg decoder"

That is the case when playing back any file, FYI.

[Neo2447]

6 minutes ago, Fitzian said:

By that, I mean, explicit reference to

> Raw AC-4 muxer and demuxer.

There's no reference to "we decode dolby atmos now" though...

That's not something Poweramp advertises. And also, since there's no legal resource for AC-4 atmos content, this is all kinda on the outskirts of what we can really discuss.

Yes, I understand what you're referring to. However, my files are in eac3 format so maybe that's why it can decode them without problems. Actually, I'm not so interested in the formats but in how it handles downmixing. The reason is that I find the surround mix of these songs massively better than the stereo mix (even already downmixed) but as I said, when trying to do the downmix manually it ends up with a very low volume and in Poweramp it has a decent volume. That makes me wonder if Poweramp uses a special downmixing algorithm (for all surround formats in general).

[Fitzian]

3 minutes ago, Neo2447 said:

Yes, I understand what you're referring to. However, my files are in eac3 format so maybe that's why it can decode them without problems. Actually, I'm not so interested in the formats but in how it handles downmixing. The reason is that I find the surround mix of these songs massively better than the stereo mix (even already downmixed) but as I said, when trying to do the downmix manually it ends up with a very low volume and in Poweramp it has a decent volume. That makes me wonder if Poweramp uses a special downmixing algorithm (for all surround formats in general).

Searching this forum for keyword 'downmix' gives only a handful of posts, and two from Max himself just say it's a "proper" downmix to stereo via ffmpeg.

[Neo2447]

4 minutes ago, Fitzian said:

Searching this forum for keyword 'downmix' gives only a handful of posts, and two from Max himself just say it's a "proper" downmix to stereo via ffmpeg.

Well, if that's the case, how strange. Assuming that the standard method is the code -ac2, I don't understand why when doing it manually the volume ends up being too low.

[andrewilley]

I don't know the specific internal parameters used for downmixing multi-channel audio files to PCM stereo, but the task would definitely be performed by ffmpeg. @maxmp would be able to confirm if he does anything special or just leaves it to the defaults.

Andre