librosa: conda librosa 0.8.0: cffi library Attribute Error using librosa.load() for .wav file

I installed librosa 0.8.0 into a conda environment having Python 3.8.6 and conda 4.9.1:

> conda install librosa
Collecting package metadata (current_repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: C:\Users\me\Miniconda3\envs\forge38

  added / updated specs:
    - librosa


The following NEW packages will be INSTALLED:

  appdirs            pkgs/main/noarch::appdirs-1.4.4-py_0
  audioread          conda-forge/win-64::audioread-2.1.8-py38h32f6830_3
  libflac            conda-forge/win-64::libflac-1.3.3-h6538335_0
  libogg             pkgs/main/win-64::libogg-1.3.2-he774522_0
  librosa            conda-forge/noarch::librosa-0.8.0-pyh9f0ad1d_0
  libsndfile         conda-forge/win-64::libsndfile-1.0.29-ha925a31_0
  libvorbis          pkgs/main/win-64::libvorbis-1.3.7-he774522_0
  llvmlite           pkgs/main/win-64::llvmlite-0.34.0-py38h1a82afc_4
  numba              pkgs/main/win-64::numba-0.51.2-py38hf9181ef_1
  pooch              conda-forge/noarch::pooch-1.2.0-py_0
  pysoundfile        conda-forge/noarch::pysoundfile-0.10.2-py_1001
  resampy            conda-forge/noarch::resampy-0.2.2-py_0

Output of librosa.show_versions():

INSTALLED VERSIONS
------------------
python: 3.8.6 | packaged by conda-forge | (default, Oct  7 2020, 18:22:52) [MSC v.1916 64 bit (AMD64)]

librosa: 0.8.0

audioread: 2.1.8
numpy: 1.19.2
scipy: 1.5.0
sklearn: 0.23.2
joblib: 0.17.0
decorator: 4.4.2
soundfile: 0.10.2
resampy: 0.2.2
numba: 0.51.2

numpydoc: 1.1.0
sphinx: 3.3.0
sphinx_rtd_theme: None
sphinxcontrib.versioning: None
sphinx-gallery: None
pytest: None
pytest-mpl: None
pytest-cov: None
matplotlib: 3.3.2
presets: None

When I tried using librosa.load() on a .wav file, I get this error:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-33-070e4879b935> in <module>
      1 # Read in the first audio file, create the time array
----> 2 audio, sfreq = lr.load('Datasets/set_b/normal__278_1311163365896_A.wav')
      3 time = np.arange(0, len(audio)) / sfreq
      4 
      5 # Plot audio over time

~\Miniconda3\envs\forge38\lib\site-packages\librosa\core\audio.py in load(path, sr, mono, offset, duration, dtype, res_type)
    144 
    145     try:
--> 146         with sf.SoundFile(path) as sf_desc:
    147             sr_native = sf_desc.samplerate
    148             if offset:

~\Miniconda3\envs\forge38\lib\site-packages\soundfile.py in __init__(self, file, mode, samplerate, channels, subtype, endian, format, closefd)
    625         self._info = _create_info_struct(file, mode, samplerate, channels,
    626                                          format, subtype, endian)
--> 627         self._file = self._open(file, mode_int, closefd)
    628         if set(mode).issuperset('r+') and self.seekable():
    629             # Move write position to 0 (like in Python file objects)

~\Miniconda3\envs\forge38\lib\site-packages\soundfile.py in _open(self, file, mode_int, closefd)
   1168             if isinstance(file, _unicode):
   1169                 if _sys.platform == 'win32':
-> 1170                     openfunction = _snd.sf_wchar_open
   1171                 else:
   1172                     file = file.encode(_sys.getfilesystemencoding())

AttributeError: cffi library 'C:\Users\me\Miniconda3\envs\forge38\Library\bin\sndfile.dll' has no function, constant or global variable named 'sf_wchar_open'

Others have had this problem:

https://stackoverflow.com/questions/59915598/attributeerror-cffi-library-pymodulespath-soundfile-data-libsndfile64bit-dl

I then completely removed the conda librosa and replaced it with the version from pip, and the error went away, all is good.

> pip install librosa
...
Successfully built librosa audioread resampy
Installing collected packages: audioread, llvmlite, numba, resampy, soundfile, appdirs, pooch, librosa
Successfully installed appdirs-1.4.4 audioread-2.1.9 librosa-0.8.0 llvmlite-0.34.0 numba-0.51.2 pooch-1.2.0 resampy-0.2.2 soundfile-0.10.3.post1

However, I would much rather use the conda version. Please help me solve the problem with the conda version.

Thank you…