Re: [Bug-gnupod] Patch to support ReplayGain / mp3gain

[H. Langos]

Hi Richard,

On Mon, Apr 20, 2009 at 04:45:03PM +0200, Richard van den Berg wrote:
> Hi Henrik,
> 
> I agree with your reasoning in principle (and will provide a patch for the
> --noAPEtag option) but if you had ever tried to use iTunes' SoundCheck and
> compare it against ReplayGain, you wouldn't be proposing this at all. :-)
> The SoundCheck algorithm is peak based, while ReplayGain uses the RMS
> energy. The difference between the two is that ReplayGain actually works
> (makes all tracks sound equally loud) while SoundCheck doesn't. (There is
> even a commercial product named iVolume that (badly) integrates ReplayGain
> into iTunes.) For more info see http://replaygain.hydrogenaudio.org/

I see :-)

I hope the iPod does take care of clipping issues. (Or aren't there any?) 
Ever encountered any problems there? Taking the simple peek value would 
avoid that altogether but a peek value indeed does not represent loudness 
in any but the most crude way.

> ;-) Besides (after hours of googling for the syntax of the iTunNORM tag) I
> borrowed most of the code from
> http://projects.robinbowes.com/flac2mp3/trac/ticket/30

Nice summary there. You could add that link to the documentation of
_parse_ReplayGain(). :-) But if you were only after the iTunNORM syntax, 
you might have saved yourself some google time... 

$ perldoc src/ext/FileMagic.pm
...
       _parse_iTunNORM(STRING)
           Searches STRING for a sequence of 8 hex numbers of 8 digits each
           used by iTunes to describe the dynamic range.  see
           http://www.id3.org/iTunes_Normalization_settings
...

BTW: I guess we should get rid of the static man pages and get that content
into the perl files and only generate them on when releasing. Otherwise it's
a maintenance nightmare ... different file, different format, different 
directory ... I wasn't even aware of their existence until well after 
sending my 5th feature patch upstream. :-/

Any volunteers?

cheers
-henrik