lame mp3's not completely gapless on 5G ipod

[dunno]

Mp3 encoded with Lame is gapless with rockbox, however, mp3 encoded by means of a non Lame encoder is generally not gapless.

[Jag86]

Another thing I forgot to say is that all of my Mp3s are encoded with lame of various versions, but all are post-3.90.3, so there are extra informations such as encoder delay and padding.
I tried the last version of Rockbox 060912-2235 and the problem is still there.
I know there is an option which sets the anti-skip buffer. I left it at 5 seconds. May I have to increase that value? What's the most correct value for you? Thanks!

[lalittle]

Another thing I forgot to say is that all of my Mp3s are encoded with lame of various versions, but all are post-3.90.3, so there are extra informations such as encoder delay and padding.
I tried the last version of Rockbox 060912-2235 and the problem is still there.
I know there is an option which sets the anti-skip buffer. I left it at 5 seconds. May I have to increase that value? What's the most correct value for you? Thanks!

Go ahead and try increasing it to 15 or 30 seconds and see what happens.  I have no idea if this will effect the issue you're having, but it can't hurt to try it.  I don't think there is really a "correct" value, so feel free to play with it.

Can you notice any other patterns to the behavior -- i.e. is it after a certain amount of "time" that it malfuncitons, or a certain number of songs, etc.  What about other albums that need gapless playback -- do these have the same problem?

Larry

[Jag86]

I'll try to increase that value, although I remember that before upgrading to last CVS build it was at 30 seconds, maybe even at 1 minute. I tried various albums, just to be sure the problem wasn't the songs themselves, but the problem is always present. The number of the songs after the problem starts isn't always the same, I haven't checked about total elapsed time yet.

[Jag86]

Nope. Different values don't solve the problem. But I found out that the same tracks, without any ID3 tag, play fine. All of my mp3s have ID3v1.1 & ID3v2.4 (UTF-8), tagged with Mp3Tag v2.36. RockBox should supports these, doesn't it?

EDIT: Uhm, I tried to remove ID3v1.1, leaving only the ID3v2.4 tag and it plays fine now. Quite strange!

[Jag86]

"Little" UP. There is any known bug (and maybe the fix) which can refer to my problem? I tried to search in the section "Bug Reports", but I haven't found anything.

[lowlight]

Make sure Mp3tag is NOT writing APEv2 tags to the files in addition to id3v1 & id3v2 tags. Currently APEv2 tags will break gapless playback in Rockbox.

[michael.conner]

Make sure Mp3tag is NOT writing APEv2 tags to the files in addition to id3v1 & id3v2 tags. Currently APEv2 tags will break gapless playback in Rockbox.

So using mp3gain (which writes APEv2 tags, I believe, for undo data) will break gapless in Rockbox?  Would removing the tags from the files, which MP3Gain allows, fix the problem?  (Assuming one doesn't care about having permanently changed the volume of the files...)

[AlexP]

Yeah, the APE undo tags will  break gapless.

Why not use replaygain?   It does the same thing as MP3gain, but via ID3 tags, so no permanent changes, no APE tags, and rockbox compatibility!  Foobar will generate them for you, and there are plugins for Winamp etc for playback.  (If you aren't on Windows, sorry I'm not sure about Linux/Mac).

[michael.conner]

Yeah, the APE undo tags will  break gapless.

Why not use replaygain?   It does the same thing as MP3gain, but via ID3 tags, so no permanent changes, no APE tags, and rockbox compatibility!  Foobar will generate them for you, and there are plugins for Winamp etc for playback.  (If you aren't on Windows, sorry I'm not sure about Linux/Mac).

Basically because my wife's MP3 player is a Creative one (not Rockboxable) and most of our collection was ripped prior to having a Rockboxed iPod, so replaygain wasn't an option for me.  Nothing against replaygain itself, mind you.

[lowlight]

At some point, I'll make the playback code strip the ape tags from the files during buffering (much like id3v1 tags are stripped). So, hold on a little while longer.

[Deano]

Either that, or get rid of mp3gain's undo tags (there are options in mp3gain to do this).

[lowlight]

I've committed the change to strip ape tags so lame-encoded mp3's with apetags should be gapless. Let me know if there are issues.