Hi,

I just read on the dev blog that Amarok2 will not support replaygain for musepack because taglib cannot read gain value in mpc files, as musepack stores gain info in a proper header, not a metadata tag.

It is also written that Amarok will read gain values written in mp3 by the mutagen lib.
I do not see then why Musepack is causing a problem, because mutagen also provides access to gain value for musepack files.

This is as simple as:

    f = open(file,'rb')
    info = mutagen.musepack.MusepackInfo(f)
    f.close()
    return(info.title_gain,info.album_gain)

This would definitely be sad to loose replaygain support in Musepack for such a small bunch of code

Feel free to submit a proper patch for Amarok via the appropriate channels. No guarantees it'll be implemented, but you will have at least done your bit.

@nonolemono we don't depend on the Mutagen library though, so we cannot use it.

Well I think I read too fast,

gain tag will be read in MP3 even though it was written by mutagen, but amarok won't make use of mutagen for this...

Anyway why not depend on mutagen for this then ? It's not a "huge" library and amarok is already making us of python am I right ?

Nono

Amarok doesn't use python.

The only way to go seems to be copying replaygain to APEv2 fields.
I'll post a script once done.

nonolemono wrote:The only way to go seems to be copying replaygain to APEv2 fields.
I'll post a script once done.

Would be great. BTW: The ReplayGain support in Amarok2 still is very buggy here.

thenktor wrote:

nonolemono wrote:The only way to go seems to be copying replaygain to APEv2 fields.
I'll post a script once done.

Would be great. BTW: The ReplayGain support in Amarok2 still is very buggy here.

In which ways is it buggy? (provided that you are using the Phonon-xine backend) Please submit precise bug reports to us, before it's "too late", that is, before 2.1 is released. We may still be able to fix it.

Mark Kretschmann wrote:In which ways is it buggy? (provided that you are using the Phonon-xine backend) Please submit precise bug reports to us, before it's "too late", that is, before 2.1 is released. We may still be able to fix it.

I'm using the gstreamer backend. Replaygain doesn't work with xine backed at all (bug report exists).
I couldn't really figure out what the problem is, but for some files in my collection replaygain just doesn't work. Sometimes even files from the same album are affected. Perhaps I should try with a complete rescan of the collection and see if this fixes something...
Of course I'll report as soon as I know more.

thenktor wrote:

Mark Kretschmann wrote:In which ways is it buggy? (provided that you are using the Phonon-xine backend) Please submit precise bug reports to us, before it's "too late", that is, before 2.1 is released. We may still be able to fix it.

I'm using the gstreamer backend. Replaygain doesn't work with xine backed at all (bug report exists).
I couldn't really figure out what the problem is, but for some files in my collection replaygain just doesn't work. Sometimes even files from the same album are affected. Perhaps I should try with a complete rescan of the collection and see if this fixes something...
Of course I'll report as soon as I know more.

Woops, yes: I did actually mean Phonon-GStreamer.

I did the script.
Trying to figure out if Amarok is using the info now. However I'm not sure that v 2.0.2 supports replaygain at all ?

nonolemono wrote:Trying to figure out if Amarok is using the info now. However I'm not sure that v 2.0.2 supports replaygain at all ?

Nope, it does not. This is a new feature in 2.1.

nonolemono wrote:The only way to go seems to be copying replaygain to APEv2 fields.
I'll post a script once done.

Will other software still work with these retagged Musepack files?

thenktor wrote:

nonolemono wrote:The only way to go seems to be copying replaygain to APEv2 fields.
I'll post a script once done.

Will other software still work with these retagged Musepack files?

I'm quite sure yes, I do not alter the real tag field as defined by the Musepack format.
Needs a bit of testing though but I'm almost sure the response is positive.

What about the script?