Replay gain issue

[sordup]

I have synched my music from my PC (via MediaMonkey) with my iRiver H340/Rockbox. Before adding the music to my device I had MM painstakingly tag all tracks for Track Volume and I have Rockbox set for Replay gain/track type for volume leveling. All tracks are either mp3 or OGG.
 This function is not working, apparently, at all. There continue to be drastic volume differences, especially when playing random tracks, about %30 of the time.
What should I be looking for to correct this issue?

[Buschel]

Please check whether rockbox identifies the replaygain data:
- Play a file which should contain replaygain data, then longpress SELECT in the WPS
- Go to "Show Track Info..."
- Check if the expected "Track Gain" and/or "Album Gain" is visible and the value is correct

[sordup]

OK. Did that. All tracks I have checked show Track and Album gain data.

[sideral]

You haven't mentioned which version of Rockbox you're using. I'm assuming it's 3.9?

[Buschel]

OK. Did that. All tracks I have checked show Track and Album gain data.

So far so good. Now please verify the next step:
- Choose a track with a large dB offset in replaygain
- Play the track
- During playback switch Replagain on and off (wait several seconds between each change to allow replaygain to step in)

You should hear a difference. What are you experiencing now?

[sordup]

OK. I was using ver. 3.8 but have upgraded to 3.9 and am testing again.

Following Buschel's advice I am checking file info for Gain data (and comparing that info with the info on my PC/MediaMonkey) and then comparing playback on Rockbox with Replaygain on and off. It does seem to be working.

What I have noticed with the above process, though, is that there are tracks whose Track/Album gain tags did not transfer over to the device. It is true that there is some transcoding of FLAC and higher bit rate OGG and mp3 files to lower ones going on when my files are moved over from MM in order to economize space on the device. However, although most of the tracks that were transcoded still contain their Gain info, some do not. The only similarity between the unsuccessful ones seems to be with tracks of a low bit rate to begin with.

Is there a way to have Rockbox analyze these questionable tracks again for Gain data once they are on the device?

[Buschel]

If I got you right you claim to loose Replaygain data when transcoding a file from A to B (B is of same format, but lower bitrate). Did you check the Replaygain data of those transcoded files in MediaMonkey?

For further analysis you might upload one of those files and the corresponding original somewhere. I could then take a closer look.

[sideral]

Maybe 3.9 already fixed your problem? There are a number of gain-related fixes in 3.9, including a correct fallback to track gain in case of missing album gain.

[sordup]

If I got you right you claim to loose Replaygain data when transcoding a file from A to B (B is of same format, but lower bitrate). Did you check the Replaygain data of those transcoded files in MediaMonkey?

For further analysis you might upload one of those files and the corresponding original somewhere. I could then take a closer look.

No. Transcoding from A to B, where A is a low bitrate to begin with. B may/may not be different format. Other transcodes (i.e. from higher bitrate and/or changing formats) works fine.

[saratoga]

If I got you right you claim to loose Replaygain data when transcoding a file from A to B (B is of same format, but lower bitrate). Did you check the Replaygain data of those transcoded files in MediaMonkey?

For further analysis you might upload one of those files and the corresponding original somewhere. I could then take a closer look.

No. Transcoding from A to B, where A is a low bitrate to begin with. B may/may not be different format. Other transcodes (i.e. from higher bitrate and/or changing formats) works fine.

What about Buschel's main question?  Do the transcoded tracks actually have replaygain information?  If not, then no surprise Rockbox doesn't find it   The bitrate part probably doesn't matter too much.

[sordup]

If I got you right you claim to loose Replaygain data when transcoding a file from A to B (B is of same format, but lower bitrate). Did you check the Replaygain data of those transcoded files in MediaMonkey?

For further analysis you might upload one of those files and the corresponding original somewhere. I could then take a closer look.

No. Transcoding from A to B, where A is a low bitrate to begin with. B may/may not be different format. Other transcodes (i.e. from higher bitrate and/or changing formats) works fine.

What about Buschel's main question?  Do the transcoded tracks actually have replaygain information?  If not, then no surprise Rockbox doesn't find it   The bitrate part probably doesn't matter too much.

The transcoded tracks in question (i.e. the ones with the low bit rate issue outlined in my last post) do not have their gain info transferred to the device. It also seems to be that gain info from lower bit rate files that aren't transcoded doesn't transfer either(as in example at link). All other files, apparantly, transcoded or not,  do have gain info transferred over.

Here is a link to a zip file containing two files, the original from my PC and its copy just transferred over to the DEVICE. The PC file has gain info, the DEVICE copy does not. This file is at a bit rate of 128. Most of my files are 160 or above and it is those that do not have this issue.

Link: https://docs.google.com/leaf?id=0B-AKDYBd2jIjMGM5YjI1YWMtNTM4Ny00ZTI3LWE2MDQtMDUxZDI3NzllYjZl&hl=en_US

[Buschel]

Your "DEVICE" file does definately not contain any replay gain information, the "PC" files does. Nothing rockbox can do about this...

[saratoga]

The transcoded tracks in question (i.e. the ones with the low bit rate issue outlined in my last post) do not have their gain info transferred to the device. It also seems to be that gain info from lower bit rate files that aren't transcoded doesn't transfer either(as in example at link).

Thats fairly common.  Transcoding changes the gain slightly, so some programs don't transfer over the old replaygain values.

Here is a link to a zip file containing two files, the original from my PC and its copy just transferred over to the DEVICE.

The second file has no replaygain tag, but has had its volume adjusted by scaling the intermediate PCM values to the replaygain value, hence its actual replaygain value is very nearly 0dB (the difference probably being due to transcoding).

[sordup]

OK. I think I see that maybe this is more of an issue with the transfer process from MediaMonkey, not RockBox. However, is it worth it to run another gain analyses once the tracks are on the device in order to try and further the attempt to make a fairly consistent playback level, or is this the best it gets at this point?
I am still having to maintain a close hand at the volume knob when playing back from RockBox due to the still prevalent volume differences (most noticeable when playing a random or shuffled playlist). This is not so when playing from PC with MediaMonkey.

[saratoga]

Could you tell me if it's worth it to run another gain analyses once the tracks are on the device

You could do that, or get software that transfers the replaygain tags, or software that adjusts the volume while transcoding like the file you linked above.