Error using audioDatastore/read (line 25) Range requested is greater than the total number of samples in the file. TotalSamples is 0.

When Im trying to classify gender of child, I extracted MFCC and f0 and trained the classifier with these features, but MY PROBLEM is when I want to use this trained model for test data sets its showing this error

"Error using audioDatastore/read (line 25)

Range requested is greater than the total number of samples in the file. TotalSamples is 0."

line 25: [audioIn,dsInfo] = read(adsTest);

Please help me out I was stucked here from few days.

[jibrahim]

Hi Kodali,

Can you try reading your file with audioread? It might be useful to see if audioread throws the same error.

[Kodali Radha]

Hi ibrahim, This is the error I have got when I use audioread for same test dataset

Error using audioread (line 105)

Expected input to be one of these types:

char, string

Instead its type was audioDatastore.

[Kodali Radha]

Actually the problem is,

• When I use small data sets the code is working fine, but when I use the same code larger data sets its throwing an error in test data set as """"Range requested is greater than the total number of samples in the file. TotalSamples is 0. """"

features = [];

labels = [];

numVectorsPerFile = [];

while hasdata(adsTest)

[audioIn,dsInfo] = read(adsTest);

-----

• values of audioIn: 1921923
• fs=44100
• adsTest:

adsTest =

 audioDatastore with properties:

                       Files: {
                              ' ...\MATLAB\DummyAudioFiles\DummyAudio\01_M_native\the frog is gone.wav';
                              ' ...\DummyAudioFiles\DummyAudio\01_M_native\uhh he's looking at it.wav';
                              ' ...\DummyAudio\03_F_native\uuuhhh the dog hm jumps out of the window .wav'
                               ... and 29 more
                              }
                     Folders: {
                              'C:\Users\RADHA\Documents\MATLAB\DummyAudioFiles\DummyAudio'
                              }
                      Labels: [01_M_native; 01_M_native; 03_F_native ... and 29 more categorical]
    AlternateFileSystemRoots: {}
              OutputDataType: 'double'
      SupportedOutputFormats: ["wav"    "flac"    "ogg"    "mp4"    "m4a"]
         DefaultOutputFormat: "wav"

[jibrahim]

Let me clarify what I mean by using audioread:

Call audioread with the file name: audioread('myfile.wav'). Do not pass the datastore to audioread. I suspect something is wrong with one of your files.

[Kodali Radha]

there are 700 recordings and cant identify which .wav file is in tesrdatastore. But the thing is, all audiofiles are clearly audible when listened thorugh PC,

i found a solution for this as, since there are too many internal folder and in few folders testdata samples go to 0 may in this case Im getting error. So I removed internal folder and and madr a single folder. then it worked accurately.

thank you ibrahim for your response