For my program on BCH decoding I had to polynomially divide the corrected BCH code with the generator matrix for the final decoded input. I did the same as follows:
for j=1:num_blocks,
op4(:,j)=mod(deconv(op3(:,j)',gen),2)';
end
where op3 is my corrected BCH code matrix (each block is a column) and gen is my generator polynomial.
This tends to be slow though as my input bit stream is pretty large (~10^6).
I've read that deconvolution can also be performed through the ifft of the element wise division of the ffts of op3 (its columns) and gen. However for a 15 point fft of both op3 and gen (gen is padded), the deconvolution is also 15 points.
What I need is for a (15,11) code to recover my block size 11 input on deconvolving the 15 point BCH code with the 5 point generator.
Any suggestions?
