Undefined function 'VAD' for input arguments of type 'double'.

Question

maheesha madhubashini 2017-5-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/341127-undefined-function-vad-for-input-arguments-of-type-double

回答： Walter Roberson 2017-5-20

I have use following downloaded code for convert speech to text.

% %  traning phase
clc; clear;close all
w = warning ('off','all');
Fs=8000;
% %  For the word "ONE"
filepart1='myrecordone';
filepart2='.wav';
% %  check for signal length
% %  we will append zeros after VAD to make sure all the signals are of
% % equal length
% % We need same number of MFCC of each signal
length_all_sig=[];
for i=1:10
    filename=strcat(filepart1,num2str(i),filepart2);
    y1=wavread(filename);
      results = VAD(y1,0.1,0.025,0.0125,20,1);
      ind_st=(0:size(results)-1)*200+1;
      ind_en=(1:size(results))*200;
      ind1=ind_st(logical(results));
      ind2=ind_en(logical(results));
      all_ind=cell2mat(arrayfun(@colon,ind1,ind2,'uni',0));
      filt_signal=y1(all_ind);
      length_all_sig=[length_all_sig length(filt_signal)];
  end
max_length=max(length_all_sig);
all_data_one=[];
for i = 1:10
    filename=strcat(filepart1,num2str(i),filepart2);
    y1=wavread(filename);
      results = VAD(y1,0.1,0.025,0.0125,20,1);
      ind_st=(0:size(results)-1)*200+1;
      ind_en=(1:size(results))*200;
      ind1=ind_st(logical(results));
      ind2=ind_en(logical(results));
      all_ind=cell2mat(arrayfun(@colon,ind1,ind2,'uni',0));
      filt_signal=y1(all_ind);
      if length(filt_signal)==max_length
          [cepstra1,aspectrum,pspectrum] = melfcc(y1,Fs,'wintime',0.025,'hoptime',0.010);
      else
          filt_signal=[filt_signal' zeros(1,max_length-length(filt_signal))];
          [cepstra1,aspectrum,pspectrum] = melfcc(filt_signal,Fs,'wintime',0.025,'hoptime',0.010);           
      end   
      all_data_one=[all_data_one cepstra1];
  end
% %  For the word "TWO"
all_data_two=[];
filepart1='myrecordtwo';
length_all_sig=[];
for i=1:10
    filename=strcat(filepart1,num2str(i),filepart2);
    y1=wavread(filename);
      results = VAD(y1,0.1,0.025,0.0125,20,1);
      ind_st=(0:size(results)-1)*200+1;
      ind_en=(1:size(results))*200;
      ind1=ind_st(logical(results));
      ind2=ind_en(logical(results));
      all_ind=cell2mat(arrayfun(@colon,ind1,ind2,'uni',0));
      filt_signal=y1(all_ind);
      length_all_sig=[length_all_sig length(filt_signal)];
  end
max_length=max(length_all_sig);
for i = 1:10
    filename=strcat(filepart1,num2str(i),filepart2);
    y1=wavread(filename);
      results = VAD(y1,0.1,0.025,0.0125,20,1);
      ind_st=(0:size(results)-1)*200+1;
      ind_en=(1:size(results))*200;
      ind1=ind_st(logical(results));
      ind2=ind_en(logical(results));
      all_ind=cell2mat(arrayfun(@colon,ind1,ind2,'uni',0));
      filt_signal=y1(all_ind);
      if length(filt_signal)==max_length
          [cepstra2,aspectrum,pspectrum] = melfcc(y1,Fs,'wintime',0.025,'hoptime',0.010);
      else
          filt_signal=[filt_signal' zeros(1,max_length-length(filt_signal))];
          [cepstra2,aspectrum,pspectrum] = melfcc(filt_signal,Fs,'wintime',0.025,'hoptime',0.010);           
      end   
      all_data_two=[all_data_two cepstra2];
  end
% %  For the word "THREE"
all_data_three=[];
filepart1='myrecordthree';
length_all_sig=[];
for i=1:10
    filename=strcat(filepart1,num2str(i),filepart2);
    y1=wavread(filename);
      results = VAD(y1,0.1,0.025,0.0125,20,1);
      ind_st=(0:size(results)-1)*200+1;
      ind_en=(1:size(results))*200;
      ind1=ind_st(logical(results));
      ind2=ind_en(logical(results));
      all_ind=cell2mat(arrayfun(@colon,ind1,ind2,'uni',0));
      filt_signal=y1(all_ind);
      length_all_sig=[length_all_sig length(filt_signal)];
  end
max_length=max(length_all_sig);
for i = 1:10
    filename=strcat(filepart1,num2str(i),filepart2);
    y1=wavread(filename);
      results = VAD(y1,0.1,0.025,0.0125,20,1);
      ind_st=(0:size(results)-1)*200+1;
      ind_en=(1:size(results))*200;
      ind1=ind_st(logical(results));
      ind2=ind_en(logical(results));
      all_ind=cell2mat(arrayfun(@colon,ind1,ind2,'uni',0));
      filt_signal=y1(all_ind);
      if length(filt_signal)==max_length
          [cepstra3,aspectrum,pspectrum] = melfcc(y1,Fs,'wintime',0.025,'hoptime',0.010);
      else
          filt_signal=[filt_signal' zeros(1,max_length-length(filt_signal))];
          [cepstra3,aspectrum,pspectrum] = melfcc(filt_signal,Fs,'wintime',0.025,'hoptime',0.010);           
      end   
      all_data_three=[all_data_three cepstra3];
  end
% % Building model
X=[all_data_one'];
options = statset('MaxIter',500,'Display','final');
obj1 = gmdistribution.fit(X,8,'CovType',...
    'diagonal','Options',options);
X=[all_data_two'];
obj2 = gmdistribution.fit(X,8,'CovType',...
    'diagonal','Options',options);
X=[all_data_three'];
obj3 = gmdistribution.fit(X,8,'CovType',...
    'diagonal','Options',options);
% %  Test data
test_data=cepstra1';
% test_data=cepstra1';
% test_data=cepstra2';
% % Word recognition
[~,nlogl1] = posterior(obj1,test_data);
[~,nlogl2] = posterior(obj2,test_data);
[~,nlogl3] = posterior(obj3,test_data);
log_like=[nlogl1 nlogl2 nlogl3];
[~,Spoken_word]=min(log_like)

but it gives

    Error in main4 (line 26)
    results = VAD(y1,0.1,0.025,0.0125,20,1);

what should I do for this

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Image Analyst 2017-5-20

You can either format your code like explained here or you can attach it with the paper clip icon.

Did you write VAD yourself? If not, do you know where it's located on your computer? If not, then why do you think MATLAB should know anything at all about this function?

请先登录，再进行评论。

请先登录，再回答此问题。