Problem with my modulated noise

SDFAS

[Stephen23]

Original question from Google Cache:

"Problem with my modulated noise"

I'm coding some modulated noise but it's coming out weird. I have 8 modulators - one for each of 4 frequencies for each ear (i.e. 4 presented to the left and 4 presented to the right). I used the designfilt option originally but changed to one created by my supervisor that uses fir filter. The output sounds correct but the plot of the image looks way off. This is my code. Please let me know if you can help.

%INPUT VARIABLES
clc;
pth = 'C:\MATLAB\HHL\SAMNoise\';
%carPhase = 0;
Fs=32e3; %sample rate
dur = 1000;%duration in ms
ti = 1/Fs:1/Fs:(dur/1000); %time index
modFreqs = [79 85 91 97]; %modulation frequencies for LE=79 85 91 97 and for RE=83 89 95 101
HPF1 = [421 915 1909 3903]; %Lower bound High Pass Frequency1 for LE 421 915 1909 3903 and RE is 417 911 1905 3899];
HPF2 = [579 1085 2091 4097]; %Upper bound High Pass Frequency2 for LE 579 1085 2091 4097 and RE is 583 1089 2095 4101];
FOrderValue = 300; %filter order value for designfilter
modDepth = 0; %100 percent
modPhase = 0;%modulation phase
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% CREATE FILTERS to shape the noise
%for i=1:length(HPF1)
  %  filts(i) = designfilt('bandpassfir', 'FilterOrder', FOrderValue, 'HalfPowerFrequency1', HPF1(i), 'HalfPowerFrequency2', HPF2(i), 'SampleRate', Fs);
    %fvtool(filts(i));  % Use fvtool to visualize a digitalFilter, i
%end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%CREATE THE STIMULUS
% CREATE zero array to store the mod freqs and output noise
modulation = zeros(length(modFreqs),Fs);
NoiseOut = zeros(Fs,length(modFreqs));
%sideAmp = 10.^(modDepth./20);
% % GENERATE the modulator signal
for i=1:length(modFreqs)
    modfreq = modFreqs(i);
    modSignal = 0.5*(1+cos(2*pi*modfreq*ti+pi));
      %CREATE Pink Noise for filter; p = bandpower(x,Fs,freqrange)
      pinknoise = dsp.ColoredNoise(1,Fs,1);
      rng default;
      x = pinknoise();
      % ADD filter to shape the pink noise; Use filter in the form dataOut = filter(i,dataIn) to filter a signal with a digitalFilter
      NoiseOut(:,i)  = cleanfiltfilt(x,Fs,HPF1(i),HPF2(i),1000);
      %filtfilt(filts(i),x); % apply the filter
      %ADD Modulation to filtered noise
      ShpdModNoise(i,:) = modSignal .* NoiseOut(:,i)'; %modulation plus filtered noise
      %WRITE out the wav file
      scaled_stim(i,:) = ShpdModNoise(i,:).*10^(-16/20);
      modFreqsstr = num2str(modFreqs(i));
      filenam = [pth 'SAM_Noise' modFreqsstr 'x_24.wav'];
      audiowrite(filenam,scaled_stim(i,:)',Fs,'BitsPerSample',24)
      %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
      %%PLOTS
      % VISUALIZE the signal
      nfft = Fs;
      nyq = Fs/2;
      fi = 0:nyq;
      f = fft(scaled_stim(i,:),nfft);
      fAmp = (2/nfft)*abs(f);
      h = abs(hilbert(scaled_stim(i,:)));
      fh = (2/nfft)*abs(fft(h,nfft));
      hold on;
      if i > 1
          figure;
      end
      subplot(221)
      stairs(fi(2:4501),fAmp(2:4501),'k');
      title({'FFT of Spectral Envelope'; 'and Modulation Frequency'});
      xlabel('Frequency (Hz)');
      ylabel('Amplitude');
      %axis([20 4500 0 0.5]);
      hold on
      stairs(fi(2:4501),fh(2:4501),':r');
      legend('Signal','Modulation Frequency');
      hold off;
      % PLOT the SAM Noise modulation envelope
      subplot(222);
      plot(scaled_stim(i,1:length(scaled_stim)/100),'k');
      title('Modulation Envelope');
      xlabel('Time (s)');
  %     ylabel(' ');
      %axis([0 320 -.5 .5]);
      %stairs(fi(1:1001),fh(1:1001),'k');
      %legend('Modulation Envelope','Carrier Signal');
      subplot(212);
      plot(scaled_stim(i,1:length(scaled_stim)/10),'k');
      title(['Modulated Signal at MF' modFreqsstr]);
      xlabel('Time (seconds)');
     % ylabel('');
      %axis([0 3200 -1 1]);
end
SumShpdModNoise = sum(scaled_stim)
return

Answer 1

[frankenberry]

Found out that the modulated is supposed to look this way when you combine them. All is well on this.

Answer 2

[Image Analyst]

When you do this:

MF2sum = zeros(Fs,(i)); %zero array for holding mod freqs

and you haven't defined i in advance (in effect overwriting the imaginary variable), it considers i as 0 + 0*i, which is the imaginary variable.

To fix, you should use a different name, like numPoints or whatever. And you have to define its value before you call zeros().

[Image Analyst]

What looks wrong about it? What does it (the 4 signals added together) look like? What do you think it's supposed to look like?

[Image Analyst]

And what are you passing in to that function for x,Fs,HP,LP,order?