I was wondering whether there's a compile of mppenc with replaygain built-in so that one don't have to spend time on another replaygain pass. I'm sure the volume information could be extracted during the mpc encoding process, so why need another pass. Can't we have it built in?

This seems like it should be doable for the radio (per-track) gain, but the album gain obviously can't be computed by looking at only the current track.

This would be a good feature for EAC, (encode/replaygain) If we nag Andre enough we should get it in.

Hmm, I've never used album replay gain anyways. If I wanted to sit down and listen to an album all the way through, I would turn RG off, so I could listen at the intended volume. RG built into the excoder would be nice, wonder why it hasn't been implemented already?, since it is already included in the decoder. Probably because of the album gain issues.

The whole point of album-based replaygain is to preserve the relative volume of the tracks in an album. The same gain is applied to all the tracks in the album, so the most audible affect is the same as simply turning down your speakers. Of course here the volume is modified during decoding to prevent clipping, so if you then turn up your speakers to counteract the volume reduction of album replaygain you are getting the volume the artist intended, but clipping is also (in most cases) elliminated (assuming the rest of the audio pipeline is also clipping-free).

Here's an older thread where this first came up, including a comment by Frank Klemm:

http://www.hydrogenaudio.org/forums/index....t=ST&f=1&t=4183

So, what EAC should do is "replaygain after extraction"

First of all there's a way to achieve this with EAC using the EMER. You first mppenc then replaygain.

But the problem is there're two passes still which I think is unneccesary. AFAIK "replaygain --auto" calculates title gain only, album gain is handled by plug-in based on title gain values.

I think Frank Klemm's response settles it on the thread Citay put a link to...

That's wrong. MPC Replaygain writes album and title values as well as peak values. You can take one file from an album and put it into a seperate folder, and you'll notice that title and album gain are still different.

But how come? I thought it processes a single file each time. How could it depend on the folder the file is in I'm completely confused now.

It calculates the album gain (and track gain) for the files in the folder it's run.
Meaning that it treats one folder as one album.

So you saying that if you have one file in one folder and run Replaygain in that folder, then that file will have DIFFERENT values for album and track gain?

Hmm, but this effect is not good. When you give a file as an input to replaygain, replaygain should come up with a result entirely dependent on that file only. Even if you give *.mpc as input I think it should calculate the track gain only for each file.

The album gain should be left to the player (and the plug-in) based on the playlist. I am still confused though since I never use the album gain and can't understand the reasoning behind it (even though it's default in foobar2000).

torok, yes.
It will calculate a track gain value only for each individual file. And it will add an album gain value based on all the tracks in the folder.
You then choose in the player what you want to use.

@atici.
This is intended. The reason you want album gain is to keep the volume difference tracks inbetween as the artist wanted it to be on the album but still have equal overall loudness thru the album compared to the rest you your music collection.
For a player to calculated the album gain you will have to analyse all tracks in an album and add the tags. Why not do it with replaygain in one go?
I don't see the problem. This is how it is intended to work.

I thought of that also but doesn't EMER only work with the wav files? also this wouldn't work for --auto, how could it calc albumgain?