When I try to run featureAndTrainRegression I get this:

File "tbtrainer_regression.py", line 3, in <module>
    att.featureAndTrainRegression('wav',1.0,1.0,att.shortTermWindow,att.shortTermStep,'svm','mesto',True)
  File "/home/oliver/Documents/pyAudioAnalysis/pyAudioAnalysis/audioTrainTest.py", line 433, in featureAndTrainRegression
    numOfFeatures = featuresFinal[0].shape[1]
IndexError: tuple index out of range

This is after it finishes analyzing the 15 files in the directory. featureAndTrain works fine for me.

I am trying to use pyAudioAnalysis but am facing an issue for long i.e. no module named eyeD3. I have all libraries ( including eyeD3) already installed and there isn't any directory issue probably. I am using Python 2.7.13 for the purpose. from pyAudioAnalysis import XYZ whatever I write here, it shows import error in eyeD3. Would be glad for your help in this regard.

This could probably do with a setup.py.

To make this work, everything should be moved into a subdirectory.

Is there a prefered way to install pyAudioAnalysis currently ?

Hi, I got the following error : Traceback (most recent call last): File "/home/jmw/Bureau/bacasable/musiques-stream/testpydub.py", line 42, in F = pafe.stFeatureExtraction(mysample[:2*w], Fs, w, pas) File "/home/jmw/Bureau/bacasable/musiques-stream/pyAudioAnalysis-master/pyAudioAnalysis/audioFeatureExtraction.py", line 591, in stFeatureExtraction chromaNames, chromaF = stChromaFeatures(X, Fs, nChroma, nFreqsPerChroma) File "/home/jmw/Bureau/bacasable/musiques-stream/pyAudioAnalysis-master/pyAudioAnalysis/audioFeatureExtraction.py", line 283, in stChromaFeatures print "nFreqsPerChroma[nChroma]",nFreqsPerChroma[nChroma] IndexError: index 56 is out of bounds for axis 1 with size 55 if the windows is small, e.g. below the max(nChroma), in my case the Win size is 110 (10ms with Fs=11000), half window is 55 then C /= nFreqsPerChroma[nChroma] crashes.

This error happens only with a small window, e.g. Fs=11000 with 10ms window so I set Win=110 samples and half window =55 for FFT.

I put a few prints in the code and saw that he nChroma table contains 92 samples) so that the code : C /= nFreqsPerChroma[nChroma] in stChromaFeatures crashes. This is because the nFreqPerChroma array is of size nChroma (=55) but nChroma contains values larger than 55. I think you should have a look at this code. best regards

Hello,

I have modified the code to be compatible with both Python2 and Python3. The following are all the changes I made:

• Ran the codes through a Pep8 Linter to make code compatible with pep8 (spaces to tabs, spaces around arithmetic operators, etc.)
• Changed "import cPickle" to "from six.moves import cPickle" to be compatible with Python 2 and 3
• Change print statements from "print 'blah blah'" to "print("blah blah")"
• Changed order of imports to alphabetical in every code
• Changed some divisions from '/' to '//' so as to get integer output, to be compatible with both python versions
• Added requirements.txt files for Python2 and Python3
• Appended .gitignore file with lines from git's standard .gitignore file for Python

I have tested the code for both Python2 and Python3, seems to be working well. I ran the same codes as mentioned in your Wiki and got the same files saved and the same graphs generated.

Let me know if you face any problems! :)

Regards, Vikram

I am geting the following error when running aT.extract_features_and_train(["<path>\Audio Classification\BabyCry","<path>\Audio Classification\GlassBreaking", "<path>\Audio Classification\Gunshot", "<path>\Audio Classification\Falling", "<path>\Audio Classification\Background"], 1.0, 1.0, aT.shortTermWindow, aT.shortTermStep, "randomforest", "randomForest2", False). All the data in the folders are .wav files

<more outputs>
Feature extraction complexity ratio: 26.6 x realtime
Analyzing file 1 of 3: <path>\Audio Classification\Background\2019-05-24-12-26-34-355000__1.wav
Analyzing file 2 of 3: <path>\Audio Classification\Background\2019-05-24-12-26-34-355000__22.wav
Analyzing file 3 of 3:<path>\Audio Classification\Background\2019-05-24-12-26-34-355000__24.wav
Feature extraction complexity ratio: 25.8 x realtime
Param = 10.00000 - classifier Evaluation Experiment 1 of 26

Traceback (most recent call last):
  File "<path>\main.py", line 3, in <module>
    aT.extract_features_and_train(["<path>\Audio Classification\BabyCry","<path>\GlassBreaking", "D:\Research\Datasets\Audio Classification\Gunshot", "<path>\Audio Classification\Falling", "<path>\Audio Classification\Background"], 1.0, 1.0, aT.shortTermWindow, aT.shortTermStep, "randomforest", "randomForest2", False)
  File "<path>t\anaconda3\lib\site-packages\pyAudioAnalysis\audioTrainTest.py", line 298, in extract_features_and_train
    best_param = evaluate_classifier(features, class_names, classifier_type,
  File "<path>\anaconda3\lib\site-packages\pyAudioAnalysis\audioTrainTest.py", line 623, in evaluate_classifier
    cm = cm + cmt
ValueError: operands could not be broadcast together with shapes (5,5) (4,4) 

I'm trying to make a classifier with pyAudioAnalysis to detect non-music sections of songs. When I set everything up and run it, the trainer blows through every step in about 5 seconds total. It neither learns, nor changes with different parameters. If 40% of the data is music, it gets 40% wrong. It simply chooses to predict either music or non-music for every piece of data. Any insight into what might cause this?

this project is amazing! I have tried the speaker diarization feature and it really works! Super Cool! However it is also worthwhile to specify the version of eyeD3 being used.. While I was installing this package, there was an error on eyeD3 module being used. The package will not work with the latest eyeD3 of versions 0.7.x but with versions 0.6.x.

Hope this project will be added in PyPi. The project is promising!

The beat_extraction function is covered in the MidTermFeatures.py. It requires two arguments: the short_features and the window size in seconds. Maybe it is since I am new to audio analysis, but I've been struggling to figure out what the right window size should be... Usually I encounter this error: ValueError: attempt to get argmax of an empty sequence or in other cases the estimated bpm always ends up as 60 The term "window size" seems ambiguous to me. I thought it could be either of two things: (1) The length of the audio snippet for which I want to estimate the bpm or (2) the short-term window size I have used previously to get the short term features. For reference, this is how I am extracting the short term features. From my understanding the window size in this case would be 1 second.

[Fs, x] = audioBasicIO.read_audio_file(path)
x = audioBasicIO.stereo_to_mono(x)
# for reference: feature_extraction(signal, sampling_rate, window, step, deltas=True)
f_names = ShortTermFeatures.feature_extraction(x, Fs, 1 * Fs, 1 * Fs)

On the segmentation wiki page, there are code examples which seem to be from an earlier version (pre-refactoring).

To bring these up to date, references to the following functions should be updated

1. mtFileClassification() -> mid_term_file_classification()
2. mtFeatureExtraction() -> mid_feature_extraction()
3. trainHMM_fromFile -> train_hmm_from_file()
4. trainHMM_fromDir -> train_hmm_from_dir()
5. hmmSegmentation() -> hmm_segmentation()
6. evaluateSegmentationClassificationDir() -> evaluate_segmentation_classification_dir()
7. aS.silenceRemoval(x, Fs, 0.020, 0.020, smoothWindow = 1.0, Weight = 0.3, plot = True) -> `segments = aS.silence_removal(x, Fs, 0.020, 0.020, smooth_window = 1.0, weight = 0.3, plot = True)

I may have missed a couple. I would have PRed but it seems like I can't for wiki pages. Great library!

#54 addressed this briefly, to "use flags2segs" but I'm not sure where/when/how is the best way to put that function call. Could someone walk through what that process would look like?

Hello, I am having the error bellow while running my program. I believe there have been some updates on_ PyAudioAnalysis

AttributeError: module 'pyAudioAnalysis.audioTrainTest' has no attribute 'featureAndTrain'

I am wondering if there is a possibility to update this script:

from pyAudioAnalysis import audioTrainTest as aT aT.featureAndTrain("...")

Thank you :)

Try this wave sound

https://drive.google.com/file/d/1IMnX38HGDSl7aweJoMygqIgIOizdhXF0/view?usp=share_link

Reference https://github.com/tyiannak/pyAudioAnalysis/blob/master/pyAudioAnalysis/audioSegmentation.py Line 798

Please help

This pull request fixes issue #376 . I know you should use read_audio_file and it will call read_audio_generic if needed and then flatten the signal, but one might end up using the generic function by discovering its existence from hints in an environment such as jupyter notebooks.
On the other hand this does not break the current behavior of the read_audio_file -> read_audio_generic option.

What is the problem

The feature_extraction function requires the signal to be of shape (m, ) and fails when given a signal of shape (m, 1).
Try running the following code as a test:

from pyAudioAnalysis import ShortTermFeatures as aF
from pyAudioAnalysis import audioBasicIO as aIO
import numpy as np
Fs, s = aIO.read_audio_file("data/mio_audio.wav")
print(s.shape)
Fs2, s2 = aIO.read_audio_generic("data/audio_test.mp3")
print(s2.shape)
# extracting features directly from the first signal is ok
_, _ = aF.feature_extraction(s, Fs, 500, 500, deltas=False)
# extracting features from the second requires reshaping
# causes
# ValueError: shapes (250,1) and (250,40) not aligned: 1 (dim 1) != 250 (dim 0)
_, _ = aF.feature_extraction(s2, Fs2, 500, 500, deltas=False)
# reshaping to (m, ) fixes the issue
s2 = s2.reshape((s2.shape[0], ))
_, _ = aF.feature_extraction(s2, Fs2, 500, 500, deltas=False)

Why this matters

This is not a critical flaw but the inconsistency is annoying as it might be hard to spot when working with read_audio_generic.

Proposed fix

Add a check at the beginning of feature_extraction to detect signals in the form (m, 1) and reshape them to (m, ).
OR
Return a signal with the same shape as read_audio_file from read_audio_generic.

There are small typos in:

• pyAudioAnalysis/ShortTermFeatures.py
• pyAudioAnalysis/audioSegmentation.py
• pyAudioAnalysis/audioTrainTest.py

Fixes:

• Should read plotted rather than ploted.
• Should read optional rather than optinal.
• Should read exactly rather than exacty.

Semi-automated pull request generated by https://github.com/timgates42/meticulous/blob/master/docs/NOTE.md