Hello everyone,
Does anyone know if theres a utility, or a frontend to replaygain.exe (for MPC) which will recursively traverse a directory structure doing RG analysis/tagging on all the MPCs, considering each directory to be an album? I have about 10gigs of MPC'd albums, most of them lacking replygain tags. I know the vorbis gain utility has a --recursive switch to do exactly what I'm talking about, but it doesn't seem to exist in the MPC replaygain util.

Any ideas on the best way to do this?

Case's ReplayGain version can do this, he fixed the --smart switch. Download: http://www.saunalahti.fi/~cse/replaygain.zip

Use it like this: replaygain --smart --auto C:\All_your_MPC_folders\ *.mpc

Already replaygained albums will be skipped.

(This would be correct if ReplayGain --smart was properly working, but it isn't. Read below how to do it the right way for now.)

Great! I knew there had to be an easy way to do this. One nit-picky thing though: is there a way to make it NOT skip already tagged albums? As I have no way of knowing that exiting tags are accurate (could have been done with an older version utility that calculates the values wrong). Or a way to strip RG tags from a group of MPCs?

Good question, i don't know. Case?



Use this command: replaygain 0 *.mpc

Replaygain can go through directories recursively, but it doesn't consider directories as albums (except if files are named in a specific way). You can use sweep for this job while waiting for better replaygain tool.
Using sweep is simple, just extract the zip to a directory under path, for example to c:\windows, and run command "sweep replaygain --auto *.mpc" from console in a directory where you want to start replaygaining. Sweep will execute replaygain command in all subdirectories. Btw, I uploaded recompiled sweep tool while writing this post, for some weird reason GCC compiled sweep had some weird issues with pipe functions.

This --smart switch that was mentioned is only used to skip already tagged albums, but official version has bad habit of clearing replaygain values from files when it is used. This is fixed in my modified version.
Btw2, I just uploaded new fixed replaygain. Previous version crashed if it was run with --smart in a directory without files.

OK thanks for the help all, I've kind of got it working now, using sweep. I say kind of because I'm having weird problems with getting consistent crashes from mppdec.exe during the process. At first it seemed it would crash when processing directories which contained non MPC files (the directory structure I was processing was mixed mpc/mp3/ogg) although I don't see why that would be a problem since I was processing "*.mpc"... but anyway I segregated the directory structure by file type and ran sweep over the mpc-only heirarchy, at which point it seemed to be working, so I left it to finish work on the many many gigs of files.... coming back to it hours later though, expecting it to be finished, I was greeted by yet another crash in mppdec. This time, it was a folder of MPCs which were reproducably causing the crash, even when I tried to just run replygain directly on the directory (not as part of a 'sweep')... in fact, any individual file in this directory would cause the crash... stripping the APE 2.0 tags from them fixed the problem. However, all the other files in this collection are tagged with APE 2.0, and none of them caused problems (at least none of them which the run of sweep had managed to get to before the crash). There didn't seem to be anything unusual about the contents of these APE tags, and they, like the rest, had been applied, by me, with tag.exe. Anybody have any ideas to explain and/or fix this weird behavior?

This is a batch file I wrote to solve the same problem.. I didn't know about the existance of sweep:)

Is there a way to do a recursive text dump for each folder?

That is, to put a replaygain.txt file in each directory .. these all being replaygained on an individual basis.

(Apologies for the old bump).