Replay Gain: A how to informational

Well, I can’t help you here. Maybe you can do some trial and errors (change encoding, tag version…) to see how it works, or upload some problematic samples?

Well I have exactly the same problem as coreying. Fuze doesn’t recognize RG values for some files (values don’t appear in Track info). And it’s not just one or two tracks, but a lot of them. Usually whole albums. Have checked on MP3Tag as well and everything seems fine, and e.g. foobar2000 can see the values just fine too. Tried rescanning the RG values but it doesn’t help.

I’d really like a fix for this bug because it renders the whole ReplayGain function pretty much useless. I use shuffle 95% of the time for all my songs and every once in a while a song almost explodes my head blasting out at full volume.

Testing with file-per-track FLAC files, the Fuze can read the RG tags - but changing between album and track mode or even the pre-gain makes no audible difference to the output.

With FLAC images with embedded cue sheet, the Fuze does not pick up *any* Replay Gain information from the file.

[edit] Ah, you have to stop the player and start again to get the RG to affect the output - pausing will not change it. [/edit]

Message Edited by NickC on 05-07-2009 01:31 PM

The Fuze applies the replay gain adjustment only when the  track is started.  Settings are not monitored or updated until another track is started or the track is restarted.

Well, I believe I have got something. It appears that if mp3 has v1+v2 tags (doesn’t matter if it is 2.3 or 2.4) then replaygain is not seen by the fuze. But, if the v1 tag is removed (say, in mp3tag), then replaygain is magically here.

So, may be fuze can see replaygain in mp3 if there is _only_ a v2 tag?

Message Edited by borka on 06-04-2009 02:47 AM

Hmm - it could be something like that. Although, for the few MP3 files of mine that the Fuze does recognise RG tags, they do have v1 tags too.

I’m rather disappointed with this RG implementation. This was supposed to be the “big reason to not use rockbox” - and yet it has been very sloppily implemented. Judging by the fact that Sansa have recommended specific software for applying RG to the MP3s, it seems that they haven’t tested compatability with a range of applications.

The fact is, I’ve been using RG now for over 4 years. My MP3s work fine with Rockbox, Winamp, Foobar, XBMC, MP3Tag, and any other program I’ve used over the years with RG support - but not the Fuze. I’m not going to change my 60GB plus of personally encoded MP3s…

Not to worry, this latest firmware incorporated many individual code changes.  I’m sure that the glitches will be worked out in a future firmware release.

This is one advantage of such an active forum!  There’s plenty of feedback from the field.

Bob  :smileyvery-happy:

Well, _if_ there will be future firmware. Let’s hope.

For me replaygain works Ok - that is, if there is id3v1 it sometimes works, sometimes no. But if the id3v1 is removed, it works 100%.

But, the main annoyance now is 2.4 incompatibility - and problems with foobar.

What problems with Foobar?

@borka wrote:

Well, _if_ there will be future firmware. Let’s hope.

 

For me replaygain works Ok - that is, if there is id3v1 it sometimes works, sometimes no. But if the id3v1 is removed, it works 100%.

 

But, the main annoyance now is 2.4 incompatibility - and problems with foobar.

They were using MediaMonkey and ReplayGain when they worked on it…could it be that MM and foobar use it differently?

@roj wrote:
What problems with Foobar?

Foobar’s replaygain ignored or the genre and/or other tags are seen as a garbage (depending on foobar preferences). You cannot get both tags and replaygian working :frowning:

@marvin_martian wrote:

…could it be that MM and foobar use it differently?

 Yes, it could. But MM and replaygain do see each other replaygain data. And I notice that id3v2-2.3 without id3v1 in foobar works not 100%, but sometime fails. And even more - sometimes MM doesn’t work either (for example, if tags are edited with mp3tag afterwards).

OK, I’m not having the problem and potentially here is why:

I use MP3Tag for tagging (kicks Foobar’s butt royally in that department).  I use Foobar for Replaygain only - not tagging.  MP3Tag is configured as follows:

ID3 V1 support - ON

ID3Tag 2.3 support - ON (UTF-16)

Read:

ID3V1

ID3V2

APE Tags

Write:

Read:

ID3V1

Delete:

ID3V1

ID3V2

APE Tags

It’s all good.

If I download MP3s, I remoe all tags using MP3Tag and re-tag them.  I get all tag fields recognized and ReplayGain tags also seen.  I don’t like Media Monkey because I see no reason to load yet another piece of software on my system for no good reason.

You’re talking about the Tags > Mpeg option screen?
If you’ve only got ID3v1 selected in the write area, then are you only writing ID3v1 tags?

Apart from that I already use those settings (it’s basically the standard config in MP3Tag I believe). I always put in the lyrics into the ID3v2 tags (using the ‘UNSYNCEDLYRICS’ field, and I also put the CD cover (max 50KB) into the ID3v2 tags. I’m not sure whether you do these things.

My feeling is that the changes made by Sansa in this latest firmware have basically broken the MP3 ID3v2 tag handling with ‘large’ tags (being that I have the lyrics and the CD cover in there). Not only does ReplayGain not work for me for most tracks (as its not even detected that they’re present), but many albums no longer display the CD cover, and even in albums which do display the CD cover, various tracks do not. This was not the case on the previous firmware (but I don’t wish to go into firmware downleveling to do further testing on that). 

First of all, when importing things into windows media player (which is what I use to sync my fuze) there’s an option there to “apply volume leveling labels to all files.” Isn’t that exactly what I’m looking for? Why do I have to use media monkey?

@coreying wrote:

@ mmike. “Song” means it will use the track gain values instead of the album gain values. You’d select “Song” normally only if playing mixes and playlists etc. Listening to albums with “Song” selected can give weird results - which is what I should get with my After Forever example. The fact I don’t get a weird result is what indicates an issue with the Fuze in that example.

 

If I understood this correctly, then why should you get weird results when playing albums on “song” mode? Wouldn’t that just make all the songs on the album the same level? Usually they already are pretty much on the same level so you don’t really need it, but why would you get a weird result?

I try to use replay gain on my Fuze, and it goes to a white screen and locks up??? I have version 2.02.28A of the firmware installed…Everything else seems to run smooth…I flashed the firmware twice to make sure it was not a bad write…Any ideas?

@landstar wrote:

I try to use replay gain on my Fuze, and it goes to a white screen and locks up??? I have version 2.02.28A of the firmware installed…Everything else seems to run smooth…I flashed the firmware twice to make sure it was not a bad write…Any ideas?

 

 

You’re not the 1st to report this problem with the latest firmware update. Could be a bug with the most recent release. Until they (SanDisk) gets it sorted out, it might be prudent to ‘back-date’ to the previous f/w version.

Another ‘Just updated firmware…replay gain choice locks up my player.’

…if anyone is keeping track of it.

The replay gain of my MP3 files which were analyzed by MediaMonkey works fine until I edit

the tags with TagScanner. The fuze just does not detect and use the replay gain value

any more. When I try to re-analyze the MP3 using MediaMonkey, it refuses to do so as

the “track volume” column already has values.

Is there a way to clear the replay gain values so that MediaMonkey will re-analyze the files?

In addition to the “+” bug, it seems to be dropping zeros.  That is, a gain of “-2.03” turns into “-2.3”.  Is anyone else seeing this?

My Fuze also has the replay gain problem of going to the blanc screen. Also if I select 10 seconds on the backlight timer it goes to blanc screen.