One thing you could do is this: before scanning album_gain, save the values track_gain values to a text file. Scan album_gain, then reload track_gain from the file.
MP3Tag can do this.

That's the only way i know, but someone else here might know better.

One thing you could do is this: before scanning album_gain, save the values track_gain values to a text file. Scan album_gain, then reload track_gain from the file.
MP3Tag can do this.

That's the only way i know, but someone else here might know better.

I'd love to offer a helpful suggestion, but I'm not sure I understand what you mean.

In my understanding, there are four potentially separate ReplayGain-related things...
1. calculating the ReplayGain (i.e. running the algorithm to automatically figure out how loud the audio file is)
2. setting the ReplayGain = putting that calculated value into a ReplayGain tag (or manually entering your own value into a ReplayGain tag)
3. using the ReplayGain = getting the player to a) read the value from the file's ReplayGain tag and b) use that value to change the volume during playback
4. applying the ReplayGain = reading the value from the file's ReplayGain tag, and then changing the actual audio data in the file so that its volume is reduced or increased by the amount specified in the tag. Good software would either re-set or remove the tag because the original value would no longer be accurate or appropriate.

There's also
5. scanning, which normally means doing items 1 and 2 on one or more files.
It's also possible to do steps 1 and 4 to apply ReplayGain volume changes without writing any tags (like wavegain does).
(It's also possible to do steps 1 and 3 on a live stream to apply ReplayGain volume changes on playback without writing any tags, but without the whole file to process up-front results can be unpredictable.)

The track and album values are separate. They're both calculated from the audio data. The album values aren't calculated from the track values.

When you say you "apply" album gain to your FLAC files, do you mean calculate, set, use, or apply - or some ombination of them?

Cheers,
David.

I'd love to offer a helpful suggestion, but I'm not sure I understand what you mean.

In my understanding, there are four potentially separate ReplayGain-related things...
1. calculating the ReplayGain (i.e. running the algorithm to automatically figure out how loud the audio file is)
2. setting the ReplayGain = putting that calculated value into a ReplayGain tag (or manually entering your own value into a ReplayGain tag)
3. using the ReplayGain = getting the player to a) read the value from the file's ReplayGain tag and b) use that value to change the volume during playback
4. applying the ReplayGain = reading the value from the file's ReplayGain tag, and then changing the actual audio data in the file so that its volume is reduced or increased by the amount specified in the tag. Good software would either re-set or remove the tag because the original value would no longer be accurate or appropriate.

There's also
5. scanning, which normally means doing items 1 and 2 on one or more files.
It's also possible to do steps 1 and 4 to apply ReplayGain volume changes without writing any tags (like wavegain does).
(It's also possible to do steps 1 and 3 on a live stream to apply ReplayGain volume changes on playback without writing any tags, but without the whole file to process up-front results can be unpredictable.)

The track and album values are separate. They're both calculated from the audio data. The album values aren't calculated from the track values.

When you say you "apply" album gain to your FLAC files, do you mean calculate, set, use, or apply - or some ombination of them?

Cheers,
David.

Track gain and Album gain values are calculated independently. And foobar2000 doesn't use track gain values to calculate album gain.

Quote from: Lossless Lady on 2013-11-08 01:44:03

The idea here is that I like to create my own albums of songs ranging from rock to classical ...within an album. I need to manually adjust the track gain of songs within an album.

You have some original albums A B and C, and you compile tracks from those into a new album Z.

Are you using ReplayGain as it's designed (i.e. to try to match the loudness of various tracks/albums), or trying to abuse it for some other purpose (e.g. making tracks intentionally much louder than others)?

When tracks from album A were being listened to as part of album A (before you made the compilation), did you think that the track gains were wrong, and needed manual adjustment to meet your requirements?

If you set foobar2k to use the album gains, and then you jump between tracks on albums A B and C (forget about compiling album Z for a moment), is that meeting your requirements?


btw, it makes absolutely no sense to ReplayGain scan album Z as a single album and use album gain. I'll try to explain that properly, and how you might work around it, when I understand the context of what you're doing.

Cheers,
David.

If ReplayGain was perfect (i.e. perfectly matched your perception and taste in judging loudness), you wouldn't need to adjust the track gains to make your compilation "work" - trackgain would already do that job for you. Then you could use track gain when playing the albums. All the track gains are aiming for a common loudness level, so using track gain will still match the loudness of all the tracks on albums X, Y and Z. You wouldn't need or want album gain at all.

Of course ReplayGain isn't perfect, and can't predict any subjective adjustments you've made to intentionally boost or cut a track's volume - but as long as your adjustments to the track gains don't drag everything much louder or much quieter over all, then those track gains should still match in loudness across albums X, Y and Z. You don't need album gain. (and you don't have a proper album gain value to use).

I assume though that you're editing the track gains to put the loudness where you want it. It's equivalent to using an editor editor with the amplify function, but probably easier, and lossless. It could easily make one album much louder overall than another, and you want to use ReplayGain to fix that.

As I think you've figured out, you could use Apply track gain after you have manually adjusted the track gains, and then re-scan X, Y and Z as three distinct albums to get correct album gains for X Y and Z. Then, even if your track gain adjustments managed to make one album much louder over all, and another much quieter over all, you could still use album gain. You couldn't play albums A, B and C as they were intended though, so you should probably apply the changes to copies.

FWIW that's what I do. Obviously I use playlists sometimes to create compilations (sometimes very long ones), but if I end up creating new compilations with most tracks having more changes to them than you could preserve in a playlist (editing tracks, EQ-ing them, changing the fade-out, cross fading them, and I guess changing the volume) then those are saved as a whole new set of audio tracks in a whole new album, with new artwork, new album name tags, and new ReplayGain data. It's not efficient having two copies of essentially the same tracks, but it makes things easier and clearer. If only a few tracks are changed in this way, then I usually save them separately and use them in the playlist in place of the originals, while using the originals for all the unchanged tracks.

Hope this helps, though I think you figured it out yourself.

Cheers,
David.