communications Merge Request #1: Implementation of genqamdemod.m (open)

Hello,

This is my first commit here. I included genqamdemod.m to the project.

Thanks!

[Andreas Weber]

Hi Yassin, thank you for your contribution.

I had a look at your code and have some comments:
* If you want to make a column vector, just use y(:), this is basically the same as reshape (y, numel (y), 1);
* Be careful if using x' this is complex conjugate transpose . If you just want to transpose use x.'
* there is no need to bring c and y to the same size, see https://docs.octave.org/latest/Broadcasting.html and it is automatically done since Octave 3.6.x
* There is no need to call min in a loop, see my modified code attached
* The second output of min is the index of the found minimum, no need to first calculate the minimum and then find its index.

-- Andy

[Achengli]

Hi Andy, thank you for the review and so good tips.

when I measure the EVM with the variable evm, I did the absolute value of the difference to find the minimum error but with all symbols of c, y will be a vector with input samples and c is all the possible symbols, so the function must find the minimum of each one of y between all of c. In any case, the code has many aspects to improve and I will push another commit with changes, not only to avoid the for loop, but also to change the complex conj transpose to a literal transpose.

Many thanks :)
Yassin.

[Achengli]

I already pushed a commit with your changes. I see now that doing the remainder of column vector and row vector gives a matrix.

Thanks for the help!