DemoAF_PowerSpecModel.m

%
%     Demonstrations for introducting auditory filters
%     DemoAF_PowerSpecModel
%     Error from Power Spectral model of masking
%     Irino, T.
%     Created:  16 Mar 2010
%     Modified: 16 Mar 2010
%     Modified: 12 May 2010
%
%     
function ErrorVal = DemoAF_PowerSpecModel(ParamOpt,ParamNN,ProbeLevel,SwRslt);

    if nargin < 4, SwRslt = 0; end;

    b = ParamOpt(1);
    K = ParamOpt(2); % efficiency K (dB)
    [frsp, freq] = ...
      GammaChirpFrsp(ParamNN.fp,ParamNN.fs,4,b,0,0,ParamNN.Nrsl);
    WeightFunc = abs(frsp).^2;
    WeightFunc = WeightFunc/max(WeightFunc);
    dF = mean(diff(freq));
  
   ErrorVal = [];
   for npl = 1:length(ProbeLevel)
       nLowBPN = find(freq >= ParamNN.FreqLowMin(npl) & ...
                      freq <= ParamNN.FreqLowMax(npl));
       nUppBPN = find(freq >= ParamNN.FreqUppMin(npl) & ...
                      freq <= ParamNN.FreqUppMax(npl));
       IntgrMask  = sum(WeightFunc([nLowBPN nUppBPN]))*dF;
       ErrorVal(npl) = ProbeLevel(npl) - (10*log10(IntgrMask)+ParamNN.N0+K);
   end;

   if SwRslt == 0,  % when optimization
     ErrorVal = sqrt(mean(ErrorVal.^2));
   end;

%%