NoBackendError:

[sharonosahon]

Hi all,

When I run this code: import os import librosa #for audio processing import IPython.display as ipd import matplotlib.pyplot as plt import numpy as np from scipy.io import wavfile #for audio processing import warnings warnings.filterwarnings("ignore")

train_audio_path = 'train/audio/' samples, sample_rate = librosa.load(train_audio_path+'yes/0a7c2a8d_nohash_0.wav', sr = 16000)

I get these errors:

Error opening 'train/audio/yes/0a7c2a8d_nohash_0.wav': File contains data in an unknown format and NoBackendError:

I've tried downloading ffmpeg and gstreamer to fix the second error but no luck. I'm not sure what to do about the first error as I have imported libraries that should be able to handle .wav files. I've also tried re-installing librosa.

Thank you for your help in advance.

[sampsyo]

Hello! Audioread uses the built-in libraries to handle WAV files, so no external dependencies should be necessary. Is there any chance your WAV file is corrupted or otherwise nonstandard? Can you share the file?

[sharonosahon]

@sampsyo thank you for your reply. I tried attaching the wav file and zip folder with all wav files but neither worked. I got the data here: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/data The wav file is located in the folder train--> audio--> yes--> 0a7c2a8d_nohash_0.wav

Apologies for not being able to attach it directly. Would you mind checking the file out on the link please?

Thank you

[sampsyo]

Have you tried opening the file with other tools? Please try opening them with a few other things, like VLC or whatever, to make sure the files actually work.

[sharonosahon]

There was indeed something wrong with the files. For some reason they were 0 seconds long. Thank you very much for your help!

[sampsyo]

Great.