CODEWITHSUNDEEP
pythonnumpyaudiodeep-learninglibrosa
Mitesh Puthran
Mitesh Puthran

Reputation: 127

ParameterError: data must be of type numpy.ndarray (While Using Librosa)

I am trying to use Librosa library(Python 3.x) for the first time to extract features from multiple audio files which are in wav format. I am facing a issue where it says the data should be of type numpy.ndarray even when it is of that type. Here is my code:

mylist is a list of all the wav audio files in my directory.

import os
import librosa
mylist= os.listdir('RawData/')

for y in mylist:
   X, sample_rate = librosa.load('RawData/'+y, res_type='kaiser_fast')
   print(type(sample_rate))
   sample_rate = np.array(sample_rate)
   print(type(X))
   print(type(sample_rate))
   mfccs = np.mean(librosa.feature.mfcc(x=X, sr=sample_rate, n_mfcc=40),axis=0)

And this is the output and the error that pops up:

class 'int'

class 'numpy.ndarray'

class 'numpy.ndarray'

ParameterError                            Traceback (most recent call last)
ipython-input-23-817ec793d6c4 in module()
  7     mfccs = np.mean(librosa.feature.mfcc(x=X, 
  8                                          sr=sample_rate,
----> 9                                          n_mfcc=40),
 10                     axis=0)
 .
 .
 .
 ParameterError: data must be of type numpy.ndarray

Any help is appreciated

Upvotes: 1

Views: 8149

Answers (1)

ShpielMeister
ShpielMeister

Reputation: 1455

you perhaps meant to say librosa.feature.mfcc(y=X,

https://librosa.github.io/librosa/generated/librosa.feature.mfcc.html

Upvotes: 1

Related Questions