Expected audioIn to be one of these types: single, double Instead its type was cell.

can anyone help me please

clc; clear; close all;

trainDir = 'dataset';
ext = '*.wav';
maxAmp = 0.5;

dinfo = dir(fullfile(trainDir, '**', ext));

filenames = fullfile({dinfo.folder}, {dinfo.name});
nfiles = length(filenames);

audio = cell(nfiles,1);
labels = cell(nfiles,1);
feature = cell(nfiles,1);

for K = 1 : nfiles
    thisfile = filenames{K};
    [fullparent, label] = fileparts(thisfile);
    [~, foldername] = fileparts(fullparent);
    [y,fs] = audioread(thisfile);
    labels{K} = foldername;
   
    y = resample(y,15000,32000);
    
    norm = zeros(length(y),1);
    if( maxAmp > 1 || maxAmp < 0 )
        fprintf('(ampMax) out of bound.');
    else
        if max(y) > abs(min(y))
            norm = y*(maxAmp/max(y));
        else
            norm = y*((-maxAmp)/min(y));
        end
    end
    audio{K} = norm;
end

[m,~] = size(labels) ;
P = 0.80 ;
idx = randperm(m);

trainAudio = audio(idx(1:round(P*m)),:) ; 
testAudio = audio(idx(round(P*m)+1:end),:) ;

trainLabels = labels(idx(1:round(P*m)),:) ; 
testLabels = labels(idx(round(P*m)+1:end),:) ;

aFE = audioFeatureExtractor("SampleRate",fs, ...
    "SpectralDescriptorInput","melSpectrum", ...
    "spectralCentroid",true, ...
    "spectralSlope",true);

featuresTrain = extract(aFE,trainAudio);
featuresTrain = permute(featuresTrain,[2,1,3]);
featuresTrain = squeeze(num2cell(featuresTrain,[1,2]));

[numFeatures,numHopsPerSequence] = size(featuresTrain{1});

featuresTest = extract(aFE,testAudio);
featuresTest = permute(featuresTest,[2,1,3]);
featuresTest = squeeze(num2cell(featuresTest,[1,2]));

layers = [ ...
    sequenceInputLayer(numFeatures)
    lstmLayer(50,"OutputMode","last")
    fullyConnectedLayer(numel(unique(trainLabels)))
    softmaxLayer
    classificationLayer];

options = trainingOptions("adam", ...
    "Shuffle","every-epoch", ...
    "ValidationData",{featuresValidation,labelsValidation}, ...
    "Plots","training-progress", ...
    "Verbose",false);

net = trainNetwork(featuresTrain,trainLabels,layers,options);

[FOG]

here some error

Error in audioFeatureExtractor/extract (line 626)

validateattributes(x,{'single','double'},{'2d','real'},'extract','audioIn')

Error in test2 (line 53)

featuresTrain = extract(aFE,trainAudio);

Answer 1

[jibrahim]

Hi FOG,

In this call to extract,trainAudio is a cell array:

featuresTrain = extract(aFE,trainAudio);

extract does not support cell array inputs. You will have to call extract on each signal separately. Something like this:

for K = 1 : nfiles
    featuresTrain{K} = extract(aFE,trainAudio{K});
end