On Thu, Jun 01, 2006, Matthias Rosenkranz wrote: > I extraced the first 500k, but they are actually played back without > problems. So I think the problem is the ID3 tag. I resaved the tags in > BMP -- which played the file -- and now the failing file works in > quodlibet, too. So I compared the ID3 tag of a failing file from the > same album with the resaved one: > $ tail -c 128 failing.mp3 > [output: see attachement faling.id3] > $ tail -c 128 working.mp3 > [see attachement working.id3]
Yeah, it's usually a problem with tags. I'm surprized it doesn't appear in the first 500k though. Could you try producing a minimalist file which is still a playable MP3 but exposes the problem? > The main difference is that the failing ID3 tag is padded with spaces > (0x20), the working one with zeros (0x00). Also the last character is > 0x01 for the working file and 0x02 for the failing one. How can I > provide more info on this as I probably can't send you the whole mp3 > beacause of copyright restrictions? Perhaps you can point at the tool which produced the file and tell me how to reproduce instead? Bye, -- Loïc Minier <[EMAIL PROTECTED]>