The question is how to replaygain in a proper way.

When mp3 file is encoded, Lame writes some replaygain information in Lame tag. (By the way, does it calculates only track value, or album value too? I guess, that album only)

Afterwards, when MP3Gain is used, it changes volume of mp3 file and information in Lame tag became unconsistent. By the way, it writes APE2 tag, and when using winamp with im_mpg123ot67, it's unpossible to read ID3v1 information. By the way, storing 3 tag in one file is some kind of madness, but for the reasons of comatibility we need three of them:)

So, the question is - how to make TAG data consistent? Is there any tool that could read and change Lame tag information, and MP3 APE2 information?

No, Lame only calculates track gain.

After applying MP3gain I let foobar update the replaygain values.
But I don't know if they are written to the LAME tag or anywhere else

Then im_mpg123ot67 is broken. APE tags come before ID3v1 so they should have no effect on them.

A few programs can read LAME-tag replaygain fields, but I don't think any of them (other than LAME) write that data - all writing is done to APE tags.

We can always hope that snelg will release an updated version of MP3gain with the (optional) ability to read/write the LAMEtag fields

Just verifying that this is somewhere on my to-do list

-Glen

Maybe, the formal logic of im_mpg123 is display ID3v1, then APE (if exist), then ID3v2. If fields of APE2 tag are blanc, but TAG exist, it could thing that right name of a song is "". You could program anything you want, you know:)

Ah, yes, that makes sense of course. MP3gain adds an APE tag with replaygain information but doesn't copy any existing ID3v* information over. The perhaps better approach for a tag reader is to display APE in preference to ID3v1 only if APE actually contains something useful - if APE title & artist are empty but the ID3v1 versions are non-empty it'd make more sense to use that information.