After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 761741 - gst-discoverer reports incorrect tag information
gst-discoverer reports incorrect tag information
Status: RESOLVED NOTABUG
Product: GStreamer
Classification: Platform
Component: gst-devtools
1.2.4
Other Linux
: Normal normal
: git master
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2016-02-09 00:38 UTC by leland_mbox
Modified: 2016-02-09 23:27 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Sample problem file (3.07 MB, application/x-compressed-tar)
2016-02-09 05:00 UTC, leland_mbox
Details

Description leland_mbox 2016-02-09 00:38:06 UTC
In Nautilus file properties and Rhythmbox, the reported metadata values for a .mp3 audio file with an ID3v2.4 tag remove spaces from the artist name and do not preserve the case of other values.

The applications get their reported values from GStreamer.  The contents of the file tag are different from the output reported by gst-discoverer.  I am looking for a reason (and a way to correct it).

gst-discoverer:

Properties:
  Tags: 
      title: Every Day I Have The Blues
      artist: B.B. King
      album: Live In Cook County Jail

mid3v2 from python-mutagen:

TIT1=Live in Cook County Jail
TIT2=Every Day I Have the Blues
TPE1=B. B. King
Comment 1 Thiago Sousa Santos 2016-02-09 03:07:03 UTC
Inspected some files locally and both gst-discoverer and mutagen-inspect report the same values for the tags.

Does this happen with every file you try? Anything special in your setup?
Comment 2 leland_mbox 2016-02-09 04:45:20 UTC
So far, I have noticed it only on this set of files, from this album.  Most files report the tag information as I expect, or I would have noticed earlier.

For this album, the reported information seems to be what was provided in the original files, rather than the tags I have edited.  The files were originally purchased and downloaded from Amazon.com.  I have removed Amazon's private tag in an attempt to resolve the problem, but it doesn't help.  I have many files for which I have edited the tags, from Amazon and other sources.  It has always seemed to work before.  The problem has led me on a search to see where the information might be cached, but I have not found that place.

Nothing is particularly special about the files.  They are part of a large collection, scanned by Rhythmbox.  The location given to Rhythmbox is full of symlinks, but Rhythmbox resolves them into the actual location of the files.  That is reported on the property pages.  The Nautilus property pages also report the information incorrectly when right-clicking on the files.  I imagine that the two applications are getting it from the same GStreamer source.

The Quod Libet media player that I also use resolves the tags correctly, but it does not use GStreamer services to read tags.

I am careful not to allow media player applications to modify the tags of audio files through updates from an online source.  That road leads to despair when a large collection is to be maintained.
Comment 3 leland_mbox 2016-02-09 05:00:54 UTC
Created attachment 320682 [details]
Sample problem file

This is one of the problem files for me.
Comment 4 Tim-Philipp Müller 2016-02-09 08:57:46 UTC
Running gst-discoverer on your file (GStreamer 1.6), I get:

  Tags: 
      title: Every Day I Have the Blues, Every Day I Have The Blues, Every Day I Have The Blues
      artist: B. B. King, B.B. King, B.B. King
      album: Live in Cook County Jail, Live In Cook County Jail, Live In Cook County Jail
      datetime: 1971
      comment: B. B. King - Every Day I Have the Blues
      track number: 2
      genre: Blues
      container format: ID3 tag
      extended comment: Songs-DB_Custom1[xxx]=B. B. King - Every Day I Have the Blues
      ID3v2 frame: buffer of 15 bytes
      image: buffer of 102052 bytes, type: image/jpeg, width=(int)600, height=(int)599, sof-marker=(int)0
      album sortname: live in cook county jail
      composer: Peter Chatman
      track count: 8
      album artist: B. B. King
      title sortname: every day i have the blues
      artist sortname: king, b. b.
      publisher: ABC
      copyright: 1971 Geffen Records
      disc count: 1
      disc number: 1
      track ID: a0a79a35-7cb3-4784-b240-cca6d552f191
      album ID: 93ddf09e-6ba8-30a9-b4b6-b45668139f27
      album artist ID: dcb03ce3-67a5-4eb3-b2d1-2a12d93a38f3
      artist ID: dcb03ce3-67a5-4eb3-b2d1-2a12d93a38f3
      has crc: false
      channel mode: joint-stereo
      audio codec: MPEG-1 Layer 3 (MP3)
      nominal bitrate: 256000
      bitrate: 256025

Couple of observations:

1) There appear to be multiple artist tags and we extract all of them:

      artist: B. B. King, B.B. King, B.B. King

   (arguably it should have spotted the duplicate here)

2) There is also an album-artist tag

      album artist: B. B. King

3) There are two(!) ID3v1 tags at the end of the file:

0033ccd0  54 41 47 45 76 65 72 79  20 44 61 79 20 49 20 48  |TAGEvery Day I H|
0033cce0  61 76 65 20 54 68 65 20  42 6c 75 65 73 00 00 00  |ave The Blues...|
0033ccf0  00 42 2e 42 2e 20 4b 69  6e 67 00 00 00 00 00 00  |.B.B. King......|
0033cd00  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 4c  |...............L|
0033cd10  69 76 65 20 49 6e 20 43  6f 6f 6b 20 43 6f 75 6e  |ive In Cook Coun|
0033cd20  74 79 20 4a 61 69 6c 00  00 00 00 00 00 00 00 00  |ty Jail.........|
0033cd30  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0033cd40  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 ff  |................|
0033cd50  54 41 47 45 76 65 72 79  20 44 61 79 20 49 20 48  |TAGEvery Day I H|
0033cd60  61 76 65 20 74 68 65 20  42 6c 75 65 73 00 00 00  |ave the Blues...|
0033cd70  00 42 2e 20 42 2e 20 4b  69 6e 67 00 00 00 00 00  |.B. B. King.....|
0033cd80  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 4c  |...............L|
0033cd90  69 76 65 20 69 6e 20 43  6f 6f 6b 20 43 6f 75 6e  |ive in Cook Coun|
0033cda0  74 79 20 4a 61 69 6c 00  00 00 00 00 00 31 39 37  |ty Jail......197|
0033cdb0  31 42 2e 20 42 2e 20 4b  69 6e 67 20 2d 20 45 76  |1B. B. King - Ev|
0033cdc0  65 72 79 20 44 61 79 20  49 20 48 61 76 00 02 00  |ery Day I Hav...|

with different artist names (space and no space btween B. and B.) and we extract info from both of them.

We also present the information from the ID3v2 tag first where present, so the 'B.B. King' is last.
Comment 5 leland_mbox 2016-02-09 23:27:54 UTC
None of my tagging tools told me that.  I will obviously have to refine my technique.  Thanks for the information.  It is not a bug on your part, so it can be closed.