flac returning ValueError: array is too big

[AnkS4]

trafficstars

I'm receiving ValueError on reading 'flac'.

with open("123.flac", 'rb') as f:
    _, _ = sf.read(f)

ErrorLog:

ValueError                                Traceback (most recent call last)
/tmp/ipykernel_17098/1860618875.py in <module>
      1 with open("123.flac", 'rb') as f:
----> 2     data, samplerate = sf.read(f)
      3     print(data)
      4     print(samplerate)

~/miniconda3/envs/stealth/lib/python3.8/site-packages/soundfile.py in read(file, frames, start, stop, dtype, always_2d, fill_value, out, samplerate, channels, format, subtype, endian, closefd)
    283                    subtype, endian, format, closefd) as f:
    284         frames = f._prepare_read(start, stop, frames)
--> 285         data = f.read(frames, dtype, always_2d, fill_value, out)
    286     return data, f.samplerate
    287 

~/miniconda3/envs/stealth/lib/python3.8/site-packages/soundfile.py in read(self, frames, dtype, always_2d, fill_value, out)
    886         if out is None:
    887             frames = self._check_frames(frames, fill_value)
--> 888             out = self._create_empty_array(frames, always_2d, dtype)
    889         else:
    890             if frames < 0 or frames > len(out):

~/miniconda3/envs/stealth/lib/python3.8/site-packages/soundfile.py in _create_empty_array(self, frames, always_2d, dtype)
   1318         else:
   1319             shape = frames,
-> 1320         return np.empty(shape, dtype, order='C')
   1321 
   1322     def _check_dtype(self, dtype):

ValueError: array is too big; `arr.size * arr.dtype.itemsize` is larger than the maximum possible size.

[bastibe]

What does f.frames contain? It sounds like it contains an unreasonable number.

[Fretless14]

also getting this error for a flac file

[bastibe]

What does f.frames contain? It sounds like it contains an unreasonable number.

[Fretless14]

For me it looks like I just made a bad flac file. I used VNC to convert a wav into flac and I must have done something wrong, because I tried another converter and it worked.

[bastibe]

Interesting! I have seen another issue where a malformed file caused similar confusion. That's perhaps good to keep in mind.

[incidentist]

I'm seeing this error on FLACs generated by tuberipper.com . I don't know what TubeRipper uses to generate FLACs, unfortunately. I can't speak to whether these files are technically malformed, but this file opens in Audacity and VLC but fails with soundfile on Python 3.11:

>>> with open("/Users/incidentist/Desktop/temp/Now Get Busy.flac", "rb") as f:
...     _,_ = soundfile.read(f)
... 
Traceback (most recent call last):
  File "<console>", line 2, in <module>
  File "/Users/incidentist/Library/Caches/pypoetry/virtualenvs/the-tuul-NYeGoq9V-py3.11/lib/python3.11/site-packages/soundfile.py", line 285, in read
    data = f.read(frames, dtype, always_2d, fill_value, out)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/incidentist/Library/Caches/pypoetry/virtualenvs/the-tuul-NYeGoq9V-py3.11/lib/python3.11/site-packages/soundfile.py", line 888, in read
    out = self._create_empty_array(frames, always_2d, dtype)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/incidentist/Library/Caches/pypoetry/virtualenvs/the-tuul-NYeGoq9V-py3.11/lib/python3.11/site-packages/soundfile.py", line 1320, in _create_empty_array
    return np.empty(shape, dtype, order='C')
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: array is too big; `arr.size * arr.dtype.itemsize` is larger than the maximum possible size.

[qmeeus]

Hi,

I got the same issue with a flac file recorded with gnome-sound-recorder. Converting it to wav with ffmpeg works though

ffmpeg -i <flac> <wav>

Looks like a bug in pysoundfile

[EDIT]: After some digging, I found out that when a Soundfile instance is created from a flac file, self.frames is set to 9223372036854775807, although I have no clue why

[bastibe]

This is a backend issue with libsndfile, the library soundfile uses to open audio files. Please check there if this is a known bug. Soundfile currently lags behind libsndfile by one release, so it is possible that it is already fixed upstream.

Since you mentioned pysoundfile. I hope that's a typo. Pysoundfile is severely outdated. Use soundfile instead.

[JainVidit12]

Hi,

I got the same issue on soundfile.