function x = decode(X, winLen, hopSize)
% x = decode(X, winLen, hopSize)
% X - input stft signal
% winLen - length of the hann window
% hopSize - number of samples in the hop from bin to bin
% preforms the inverse short time fourier transform of a signal.  The
% signal x is reconstructed from X by taking the fft

signalLen = size(X,2); % number of columns in the Hanning windowed signal
x = zeros(1,winLen+signalLen*hopSize); % Initialize the output vector

% Create a periodic hanning window the length of f
hannWin = hann(winLen, 'periodic');
hannWin = hannWin'; %transpose of hannning window

winIndex = 1;

for i = 0:hopSize:(hopSize*(signalLen-1))
    Xfft = X(:,winIndex)';
    Xfft = [Xfft, conj(Xfft([(winLen/2):-1:2]))]; %Conjugate of the input signal
    xMag = real(ifft(Xfft,winLen)); % Real portion of the ifft of the window vector
    x((i+1):(i+winLen)) = x((i+1):(i+winLen)) + xMag.*hannWin;
    winIndex = winIndex + 1;
end;

%takes the fft samples of the bin (since the bin is size N/2) and then the
%conjugate of those samples in reverse order.  This is to get the entire
%length N vector (earlier adjusted for the Nyquist rate sample reduction of the encode process) to preform the ifft operation on.  this is equivalent to
%to getting