Poweramp ignore replaygain 2.0 tag

[maxmp]

Yes, these files have R128_TRACK_GAIN set to 0. This means default +5dB gain is applied in addition to (decoder) handled implicit gain which is for the track test1 is FD.C0 (hex Q7.8 signed value, == -2.25) and for test2 is FE.D3 (-1.18).

So basically final gain for opus is: +5dB + header gain + R128_TRACK_GAIN (+ optional gains from Poweramp RG settings/preamp).

Poweramp doesn't display implicit decoder gain, but it displays replay gain with the +5dB offset applied.

[Masklin]

On 3/12/2024 at 1:18 PM, maxmp said:

Yes, these files have R128_TRACK_GAIN set to 0. This means default +5dB gain is applied in addition to (decoder) handled implicit gain which is for the track test1 is FD.C0 (hex Q7.8 signed value, == -2.25) and for test2 is FE.D3 (-1.18).

So basically final gain for opus is: +5dB + header gain + R128_TRACK_GAIN (+ optional gains from Poweramp RG settings/preamp).

Poweramp doesn't display implicit decoder gain, but it displays replay gain with the +5dB offset applied.

Hey @maxmp, I swear I had enabled email notifications but I still didn't get any...

I interpret your post to mean that Poweramp is working as intended. However, these opus files that I sent get their RG tags set by foobar2000 just like all other audio files I listen to... but they're either too faint or too loud, I constantly have to change the volume manually. How can we work around this? Shall I send you an example opus file that has proper RG interpretation in Poweramp??

Sincerely,

Masklin

[andrewilley]

@Masklin Yes, the problem here is that your sample files had a ReplayGain tag which had a value of zero. Thus Poweramp will apply that zero boost value as instructed (plus or minus the additional overall 'RG Preamp' setting in the RG Menu).

See my earlier post showing you the zero-value tag contents in your files, using MP3Tag, TagScanner, and as a raw binary dump:

Andre

[Masklin]

@andrewilley@maxmp Hello again, so I conclude that the root of this problem is OPUS. Poweramp chose one way of dealing with it, foobar2000 chose another, and they don't play well together. As is tradition, foobar2000 devs say the fix is in Poweramp's hands, and vice versa. I chose to circumvent all of it by re-muxing all opus audio into matroska containers.

For reference:

ffmpeg -i input.opus -c:a copy output.mka

I hope this helps someone Googling in the future. Now all my RG values are applied properly in Poweramp. Good times!

[andrewilley]

I'm not sure why foobar2000 (usually my benchmark for file handling) is creating a valid R128_TRACK_GAIN tag but is setting its value to zero. But that's definitely what it is doing - MP3Tag, TagScanner and even a raw dump of the bytes contained in the file clearly show that:

It is also creating a header OutputGain value (which is an overall gain value to be applied to a track, regardless of whether ReplyGain processing is being applied) which Poweramp correctly detects and applies to the whole track.

Remember that in addition to any quoted RG value inside tags (in this case, zero) PA also applies a user-configurable offset whenever an RG tag is present - which by default is +5dB - to counter the unexpectedly low volume that the RG spec otherwise defines. If you set that offset value to zero, then only the OutputGain value in the header will get applied, as RG will become nothing (zero plus zero).

But yes, using a different format, or a different container for the same audio data, may well solve the problem.

Andre