I have a few mpc files (less than 1 % of my mpc collection) that replaygain has problems with. Replaygain just uses 100% cpu and it runs forever so I have to press crtl-C to stop it. Tags should not be the problem, I always remove id3v2. This happens in both linux and windows. In windows I have this replaygain: http://www.saunalahti.fi/~cse/replaygain.zip.

All the problem files have a lot of silence somewhere in the middle (e.g. Offspring - Smash), maybe this is causing the problem. They all play fine and mpcscan says there are no errors.

I did some search and didn't find anything relating to this.

Try uploading a problematic track.

Last week i had the same problem with a track. It was the last track of an album with 20 min of silence after the song, and a short talk afterwards (all in one track). Replaygain took about 10 minutes to complete with 100% cpu time. It did finish correct though.

Seems like replaygain doesn't like long passages of silence for some reason.

You are right! It does finish, but it takes a really loooong time.

Jan S.: maybe uploading is not such a good idea because of © stuff If you want to see this problem, you could just insert e.g. 10 minutes of digital silence into middle of a track, encode it and use replaygain.

Maybe it's causing many FPU exceptions ? Those can slow a program down by a factor of 10..50

If I may add something...

I tried using ReplayGain on a FLAC file with a long period of silence, and it did fine for me. The track was Alanis Morissette ~ You Oughta Know - Your House from Jagged Little Pill (track #13). The song has a long period of silence after about 4 mins or so of music, the silence lasts a couple of more mins, then the track finishes with an a capella sequence. I don't have the file on me, unfortunately, because I've had to change my gain-adjustment approach to accommodate my PhatBox, which doesn't read ReplayGain tags (i.e., I had to start WaveGaining everything). But I can recreate the WAV, convert to FLAC and apply ReplayGain in short order if so requested for a comparison.

But anyway, I just thought I'd share my experience using ReplayGain on a file with extended silence, but encoded in a different format. If it is a bug, maybe it's specific to ReplayGaining MPC files?

Edit: After reading the two posts that occurred in the time it took me to post this one (got distracted doing background stuff), I should add that I think it *did* take longer on the track I mentioned, but not a tremendous amount of time if my memory serves. For reference, my machine is a P4m 2.0 GHz with 512MB of RAM...pretty average I think by todays standards.

I thought replaygain is for musepack sv7 files only (or at least the version I have demands a SV7 file). Are you sure you used replaygain with flac? I use metaflac --add-replay-gain for flac.

@zokik: replaygain is a technique used, it's not the software itself or format. almost every format has some proggie that does replaygain on it, flac, ogg, mp3, mpc, etc. have it.

Could be.

I'm on a P4 so it could be some silly bug.

@zokik

What kind of cpu do you have ?

I too have P4, 1.6 GHz to be more precise.

lexor: I know what general meaning of term "replaygain" is. I was refering to the program for musepack files which is also called "replaygain" (unlike vorbisgain for ogg vorbis, mp3gain for mp3, ...)

Must be that crazy P2/P4 FPU bug then.

IIRC the crossfeed plugin had a bug similar like this, cpu time going koko when playing digital silence.