Ticket #2839 (closed bug: wontfix)

Opened 1 year ago

Last modified 1 year ago

PUID not calculated on some mp3 files

Reported by: murdos Assigned to: luks
Priority: normal Milestone: Picard Qt
Component: Picard Tagger Version: Picard 0.9.0alpha8
Keywords: Cc:
Browser: OS:

Description

I've some mp3 files for which Picard can't find PUID.

However the genpuid command returns for them a PUID.

Reproduced with Picard 0.7.2 too.

I can send you some files, just ping me when you'll work on this ;-)

Attachments

2007-05-11.log (3.9 kB) - added by murdos on 2007-05-11 12:05:26.

Change History

2007-05-11 12:05:26 changed by murdos

  • attachment 2007-05-11.log added.

2007-05-16 16:13:03 changed by zout

Isn't it the case that these PUIDs just aren't yet calculated and added to MusicIP?

2007-05-16 16:28:27 changed by murdos

No, genpuid does return the PUID and not 'unanalyzable' or 'pending' (it seems to be the state when addition to MusicIP is pending).

And of course I've waited some days and then tried again. And it still didn't work :-/

2007-06-06 13:25:12 changed by murdos

might be related to #2419

I've also tested the same files with the gst-puid command line tool provided by the GstPUID project (http://freshmeat.net/projects/gstpuid) and it finds the same PUIDs that genpuid finds (whereas Picard does not return any PUIDs).

2007-06-06 13:38:12 changed by luks

Not that I'm very insterested in fixing MusicDNS problems :) but can you send me the file? I'm guessing the track length is calculated incorrectly, and that's the reason why it doesn't return any PUID.

2007-06-06 23:11:24 changed by luks

  • status changed from new to closed.
  • resolution set to wontfix.

From the two files you sent me, "Skaferlatine-PisteAudio 02.mp3" seems ok and Picard can find the PUID (at least on Linux). "Beautés Vulgaires - 07 - suis ton etoile.mp3" has some syncing problems: 

$ lame --decode ../Beaut%C3%A9s%20Vulgaires%20-%2007%20-%20suis%20ton%20etoile.mp3
input:  ../Beaut%C3%A9s%20Vulgaires%20-%2007%20-%20suis%20ton%20etoile.mp3
        (44.1 kHz, 2 channels, MPEG-1 Layer III)
output: ../Beaut%C3%A9s%20Vulgaires%20-%2007%20-%20suis%20ton%20etoile.mp3.wav
        (16 bit, Microsoft WAVE)
skipping initial 1105 samples (encoder+decoder delay)
Frame#   971/9302   128 kbps   MS    bitstream problem: resyncing...
Frame#  1029/9302   128 kbps  LMSR   bitstream problem: resyncing...
Frame#  1538/9302   128 kbps  L  R   bitstream problem: resyncing...
Frame#  1567/9302   128 kbps  LMSR   bitstream problem: resyncing...
Frame#  1586/9302   128 kbps   MS    bitstream problem: resyncing...
Frame#  1645/9302   128 kbps   MS    bitstream problem: resyncing...
Frame#  9325/9302   128 kbps   MS            

$ madplay ../Beaut%C3%A9s%20Vulgaires%20-%2007%20-%20suis%20ton%20etoile.mp3 -o ../F.wav
MPEG Audio Decoder 0.15.2 (beta) - Copyright (C) 2000-2004 Robert Leslie et al.
error: frame 971: lost synchronization
error: frame 1029: lost synchronization
error: frame 1538: lost synchronization
error: frame 1567: lost synchronization
error: frame 1586: lost synchronization
error: frame 1645: lost synchronization
9318 frames decoded (0:04:03,4), +1,8 dB peak amplitude, 3270 clipped samples

and different decoders handle it differently. I've tested three decoders - mpglib from lame (I suspect something based on mpglib is also used in MusicIP), madlib (used by picard 0.7) and ffmpeg (used by picard 0.9) and they all produced different results on those corrupted frames. So... there is not much I can do about it.

2007-06-06 23:16:11 changed by luks

One more thing, after decoding and reencoding the file with LAME, Picard was able to find the PUID. So this really affects only corrupted files.

2007-06-06 23:49:54 changed by BrianFreud

I've run into several files with related issues, either corrupted frames or other issues. See my post to the users list today for some solutions I was able to dig out of the musicip posts and forums.

2007-06-11 22:00:05 changed by murdos

luks: Thanks for looking at this issue, and sorry for the wasted time since it was lastly an encoding problem :/

BrianFreud: I've seen your post and tried both software you mentioned. I've mixed results: one seems to fix the file so that PUID is retrieved, the other does not.

2007-06-12 17:23:58 changed by BrianFreud

I've had similar results. Seems to be, if a VBR frame had an error, fixing it will normally allow PUID generation, but there's still a subclass that can't generate - and transcoding isn't always desirable. Still trying to find a non-transcoding solution, or at least, a cause, for these.

Add/Change #2839 (PUID not calculated on some mp3 files)




Action