Beets: chroma tells me that audio could not be decoded

[xsteadfastx]

i cant get the chroma plugin working on all my ubuntu unicorn boxes. i have the feeling that i didnt had the problem with the previous beets version. i checked all installed packages named in the doc.

• python-gst0.10-dev
• gstreamer0.10-plugins-good
• gstreamer0.10-plugins-bad
• gstreamer0.10-plugins-ugly

and i even installed libchromaprint-tools to have the fpcalc command. i still get the error:

fingerprinting of '/home/myuser/Musik/Kill Holiday/Somewhere Between the Wrong Is Right/01 Somewhere Between the Wrong Is Right.flac' failed: audio could not be decoded

if i use the fpcalc command on that file... it gives me the fingerprint. so i think that it can be decoded somehow. im pretty stuck with finding the error. anything i could try?

my config:

directory: ~/Musik/beets                                                          
paths:                                                                            
        singleton: Non-Album/$artist/$title                                       
plugins: chroma                                                        

[brunal]

This error comes from pyacoustid, in _fingerprint_audio_read(): it catches a DecodeError from audioread, which happens because no backend could read the music. Please try to determine which backends are available: try the following
import gst # accourding to your post, this should work...
import mad
And do you have ffmpeg available on the commad-line?

[brunal]

If using GST, you need 3 elements: uridecodebin, audioconvert and appsink. You may be missing one of those 3. Are there other plugin packages available on ubuntu? Is something like gstreamer0.10-plugins-base available?

[xsteadfastx]

i "investigated" a little and i think the problem is that i try to run beets in a virtualenv. import gst doesnt work in the venv. now i installed everything system-wide and it works. mmhhh not the best way but it works now.

[brunal]

Right! That's a recurring problem in the python ecosystem (cf. matplotlib, opencv...). You should be able to get it working in a virtual environment by creating/updating it with the --system-site-packages option.

[xsteadfastx]

so... im sorry for generating just hot air with this bugreport and thanks for all the informations and help :)

[brunal]

No problem, the doc should probably mention that point (I'll update it). Enjoy beets!

—
Tagada tsouin tsouin

On Wed, Jan 21, 2015 at 4:38 PM, xsteadfastx [email protected]
wrote:

Closed #1250.

Reply to this email directly or view it on GitHub:
https://github.com/sampsyo/beets/issues/1250#event-222387751

[bolisaf]

I've the same problem but I'm not as python-savvy as xsteadfastx
I've installed beets using pip.

1. how do i know if i'm running beets in a virtual env
2. if so, how to i not do this or how do i update with --system-site-packages? i've tried installing virtualenv

thanks!

[sampsyo]

1. You aren't unless you know you are (you need to do specific non-standard work to set up a virtualenv).
2. No need unless you're using a virtualenv.

Maybe you're having a different but related problem?

[bolisaf]

i even went as far as to install virtualenv :D removed it now. do i create another ticket or shall we keep it in here?

[xsteadfastx]

from my experience with beets, you shouldnt use virtualenv for beets. just install it as normal user with "pip install --user beets". but i dont know if this solves the problem you have with chroma.

[RemyEsclapeau]

Hello everyone, I am in the same case as bolisaf:

• I have installed beets and chroma plugin on ubuntu
• chroma tells me that audio could not be decoded
• using fpcalc works by giving me a fingerprint...

Does anyone has a clue? Thanks a lot

[RemyEsclapeau]

Finally I get it working by modifying the def fingerprint_file(path, maxlength=MAX_AUDIO_LENGTH): function of the acoustid.py file:

def fingerprint_file(path, maxlength=MAX_AUDIO_LENGTH):
    """Fingerprint a file either using the Chromaprint dynamic library
    or the fpcalc command-line tool, whichever is available. Returns the
    duration and the fingerprint.
    """
    path = os.path.abspath(os.path.expanduser(path))
    '''if have_audioread and have_chromaprint:
        return _fingerprint_file_audioread(path, maxlength)
    else:
        return _fingerprint_file_fpcalc(path, maxlength)'''
    return _fingerprint_file_fpcalc(path, maxlength)

I just forced acoustid to use the fpcalc commands... Not a real solution but a least it works...

[linkdaniel]

Hi everyone, fingerprinting fails on my machine with a similar error message. Importing gst and mad works. ffmpeg and fpcalc commands are available. fpcalc works fine by itself, but beets fails with the following output.

$ beet fingerprint /home/daniel/Musik/Collection/A Perfect Circle/Mer De Noms/Disc 00/The Hollow.m4a: fingerprinting fingerprint generation failed: audio could not be decoded /home/daniel/Musik/Collection/A Perfect Circle/Mer De Noms/Disc 00/Magdalena.m4a: fingerprinting Traceback (most recent call last): File "/usr/bin/beet", line 9, in <module> load_entry_point('beets==1.3.8', 'console_scripts', 'beet')() File "/usr/share/beets/beets/ui/__init__.py", line 964, in main _raw_main(args) File "/usr/share/beets/beets/ui/__init__.py", line 954, in _raw_main subcommand.func(lib, suboptions, subargs) File "/usr/share/beets/beetsplug/chroma.py", line 181, in fingerprint_cmd_func write=config['import']['write'].get(bool)) File "/usr/share/beets/beetsplug/chroma.py", line 287, in fingerprint_item _, fp = acoustid.fingerprint_file(item.path) File "/usr/local/lib/python2.7/dist-packages/acoustid.py", line 321, in fingerprint_file return _fingerprint_file_audioread(path, maxlength) File "/usr/local/lib/python2.7/dist-packages/acoustid.py", line 264, in _fingerprint_file_audioread fp = fingerprint(f.samplerate, f.channels, iter(f), maxlength) File "/usr/local/lib/python2.7/dist-packages/acoustid.py", line 206, in fingerprint fper.feed(block) File "/usr/local/lib/python2.7/dist-packages/chromaprint.py", line 119, in feed raise TypeError('data must be bytes, buffer, or memoryview') TypeError: data must be bytes, buffer, or memoryview

Any ideas?

[sampsyo]

Hi! A fix has been committed for that issue in pyacoustid: https://github.com/beetbox/pyacoustid/pull/31

[linkdaniel]

Great, thanks! Do you know when it might be available via pip? I guess that's faster than waiting for the corresponding Mint update.

[sampsyo]

Pretty soon! I'll release it in a few days.

[linkdaniel]

It worked! :) There was one error in the middle of fingerprinting ~1,700 songs, but that's all. Thanks again!

Traceback (most recent call last):
  File "/usr/bin/beet", line 9, in <module>
    load_entry_point('beets==1.3.8', 'console_scripts', 'beet')()
  File "/usr/share/beets/beets/ui/__init__.py", line 964, in main
    _raw_main(args)
  File "/usr/share/beets/beets/ui/__init__.py", line 954, in _raw_main
    subcommand.func(lib, suboptions, subargs)
  File "/usr/share/beets/beetsplug/chroma.py", line 181, in fingerprint_cmd_func
    write=config['import']['write'].get(bool))
  File "/usr/share/beets/beetsplug/chroma.py", line 287, in fingerprint_item
    _, fp = acoustid.fingerprint_file(item.path)
  File "/usr/local/lib/python2.7/dist-packages/acoustid.py", line 321, in fingerprint_file
    return _fingerprint_file_audioread(path, maxlength)
  File "/usr/local/lib/python2.7/dist-packages/acoustid.py", line 262, in _fingerprint_file_audioread
    with audioread.audio_open(path) as f:
  File "/usr/local/lib/python2.7/dist-packages/audioread/__init__.py", line 83, in audio_open
    if _ca_available():
  File "/usr/local/lib/python2.7/dist-packages/audioread/__init__.py", line 57, in _ca_available
    lib = ctypes.util.find_library('AudioToolbox')
  File "/usr/lib/python2.7/ctypes/util.py", line 248, in find_library
    return _findSoname_ldconfig(name) or _get_soname(_findLib_gcc(name))
  File "/usr/lib/python2.7/ctypes/util.py", line 237, in _findSoname_ldconfig
    f = os.popen('LC_ALL=C LANG=C /sbin/ldconfig -p 2>/dev/null')
OSError: [Errno 12] Cannot allocate memory

[sampsyo]

"Cannot allocate memory!" The sounds serious. Would you mind filing a bug in the audioread repository?

[stevehemingway]

I found that installing ffmpeg fixed this for me on WSL (Windows Subsystem for Linux). beet fingerprint now works fine.

[Holzhaus]

Had the same issue on a Debian system, installing ffmpeg fixes this.