AM modulation wav file tremolo effect problem

% % tremolo1.m % M Script that uses a low frequency AM to provide a tremolo effect
filename='acoustic.wav';
% read the sample waveform [x,Fs,bits] = wavread(filename);
index = 1:length(x);
% using a sin wave at the moment, in practice is usually a % triangle wave or a square wave or a cross between the two
Fc = 5; alpha = 0.5;
trem=(1+ alpha*sin(2*pi*index*(Fc/Fs)))'; % sin(2pi*fa/fs)
y = trem.*x;
% write output wavwrite(y,Fs,bits,'out_tremolo1.wav');

i found this code here http://www.cs.cf.ac.uk/Dave/CM0268/Lecture_Examples/ . Does implementation works for you? I always got this error :

??? Error using ==> times Matrix dimensions must agree.
Error in ==> tremolo1 at 26 y = trem.*x;

how to make matrixs dimensions agree, otherwise how to make it to work? thanks in advance

[filip novacek]

Hi, I also tried the script from the cs.cf.ac.uk - with 2 different wav files. The first worked perfectly, the second got me the same error as you described.

Answer 1

[Walter Roberson]

I do not have a problem with the file that is provided in the lecture notes. However, at the time it was written, the code assumed that the input file was mono, and for non-mono inputs, the code will fail in versions up to R2016a. As of R2016b, MATLAB added "implicit expansion" in a way that would permit the code to work.

The particular demonstration input file has 148799 samples. length() is 148799 and so index becomes 1 x 148799 and trem which uses input becomes 1 x 148799 but at the last moment is transposed to 148799 x 1. That is then .* the original input which is 148799 x 1 . And for the original input and any mono source that would work fine.

Suppose the input were stereo, 148799 x 2. Then length() is 148799, index is 1 x 148799, trem comes out as 148799 x 1, and that 148799 x 1 is .* the 148799 x 2. Up to R2016a the mismatch in number of columns would lead to a failure. Starting in R2016a it would be treated as if bsxfun(@times, trem, x) had been written, giving a 148799 x 2 output.