On Fri, Jan 24, 2014 at 2:37 PM, Henry J. [email protected]
I think the underlying assumptions to the algorithms are specific to
PSKs rather than the code. You might look at papers describing the
algorithms to get an idea of whether they are suitable for QAM or not.
If not, you might find another algorithm that you can implement. This
post has a little more detail (although it’s earlier than the one you
link to) http://www.trondeau.com/blog/tag/snr . It might be a good
idea to read through the paper Tom cites in that article since it
appears to be a good overview.
A quick search turned up this paper
didn’t read through the whole paper and I’m not really familiar with
M2M4 other than it calculates the 2nd and 4th moments, but the
conclusion suggests M2M4 is probably good for QAM.