I have several tracks that sound quieter than other ones. Is there a possibility to edit replaigain value that plex has saved? That way I could adjust incorrect loudness value.
No, this is not possible. Mainly because Plex doesn’t compute just one loudness value per track, but ten values per second. This enables Sweet Fades.
So is there any way to correct wrong loudness?
There is not. But loudness is rarely “wrong”.
You must keep in mind that Plex is primarily using the loudness of the whole album as reference.
If you have combined files from different sources into an album, then this loudness can be wrong, because the loudness of the different tracks on this album doesn’t match.
It is compilation, but as you mentioned tracks are from different sources.
Before I put it in plex, I removed all tags and added just artist, title, album title, year and cover image. There should not have left any replay-gain tags.
Maybe you can code into plex, that users can mark that it is compilation and then plex is using individual track loudness not the loudness of whole album/compilation.
Or if I play it as album, then plex uses album loudness, if I play random tracks, then use track loudness. I think that would be a good solution.
So for now the only way to correct is to decode all mp3 to wav, normalize and re-encode back to mp3? That’s really not the best solution. 
This issue cannot be solved even with normalization.
If the tracks are from different sources, they may have been mastered (read: compressed) differently. This results in differing loudness, even though the peak levels are the same.
You simply don’t combine tracks from different sources into one album.
Thats weird you say that. When i use simple replay-gain on them, they play as expected with equal loudness despite being from different sources. And they play nice on Linux (Audacious) and Windows (Foobar2000) players.
Why did you (plex developers) create new loudness normalizer function that cannot fully do that?
Simple replay-gain works great.
If you use replay gain per-track, it works.
But plex uses replay gain per album.
I do understand that, but when I play random tracks from different albums, Plex should use replay-gain per-track. Most replay-gain software makes two tags: per-album and per-track.
Plex should do same. Calculate two values: one per-album and one per-track. Then it uses appropriate value when playing album or playing just random tracks.
It shouldn’t be that difficult to implement and that would make Plexamp even awesomer 
player.
Plex did that first. But users complained that it raised the loudness of (intentionally) quiet tracks too much. So the approach of Tidal was adopted, which use the album value even in mixed playlists. Their method is backed by solid research.
oh, well. That’s disappointing, but I will try to experiment with those tracks that are quieter, maybe I will find a method that works for me.
Anyway, thank you for your time.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
