get_duration fails on .wav when not loaded into memory v0.8

Describe the bug get_duration should be able to get the length of an audio file without loading it into memory, according this solved issue: https://github.com/librosa/librosa/issues/345 However, when I try it on a .wav from the CREMA-D dataset, it fails with: ParameterError: Audio data must be floating-point

To Reproduce

import librosa

librosa.get_duration("1001_IEO_NEU_XX.wav")

Audio file(s):

• git clone CREMA-D with only .wav files (not video/MP3): git clone --config lfs.fetchexclude='AudioMP3,VideoFlash' https://github.com/CheyneyComputerScience/CREMA-D.git
• Download this sample file: 1001_IEO_NEU_XX.zip

Expected behavior Output the length of the audio file in seconds: 1.2832199546485261

Error If applicable, add screenshots to help explain your problem.

Software versions*

Linux-4.15.0-112-generic-x86_64-with-debian-buster-sid
Python 3.7.6 | packaged by conda-forge | (default, Jan  7 2020, 20:28:53) 
[GCC 7.3.0]
NumPy 1.18.5
SciPy 1.5.0
librosa 0.8.0
INSTALLED VERSIONS
------------------
python: 3.7.6 | packaged by conda-forge | (default, Jan  7 2020, 20:28:53) 
[GCC 7.3.0]

librosa: 0.8.0

audioread: 2.1.8
numpy: 1.18.5
scipy: 1.5.0
sklearn: 0.23.1
joblib: 0.16.0
decorator: 4.4.2
soundfile: 0.10.2
resampy: 0.2.2
numba: 0.50.1

numpydoc: None
sphinx: 2.3.1
sphinx_rtd_theme: 0.4.3
sphinxcontrib.versioning: None
sphinx-gallery: None
pytest: 5.4.3
pytest-mpl: None
pytest-cov: None
matplotlib: 3.1.1
presets: None

Additional context The following code does work, showing it shouldn’t be an issue with the .wav itself:

signal, sr = librosa.load("1001_IEO_HAP_HI.wav", sr=None)
librosa.get_duration(signal)

Might it be related to the change in default audio parser since v0.7.0?

[NumesSanguis]

I guess if I checked the documentation more carefully, I would have noticed the filename keyword. Therefore, the possible breaking of people’s code to reduce one argument might not be worth it. If you want to keep it like this, I’ll close the issue (or you can close it).

[lostanlen]

@NumesSanguis the recommended prototypes are: get_duration(y=signal, sr=sr) get_duration(filename=filename)

It’s less about the type of the first argument that about the number of arguments.