Prev: estimation parameter using fminsearch
Next: Error, (in int) integer too large in context
From: Kevin on 23 Mar 2010 03:13
It seems that the training symbols have to use the same modulation as
the data symbols. Is there any way that I can use different
modulations? say QPSK for training and 16QAM for data. I tried
equalizing the two parts separately with a constellation change in
between. The shape of the equalized constellation seems correct, but
it is compressed. Of course the demodulation result is wrong.
Thanks
From: Kevin on 23 Mar 2010 03:27
In other words, what should I do if the average power of the training
and data are different?

On Mar 23, 3:13 pm, Kevin <umass.amhe...(a)gmail.com> wrote:
> It seems that the training symbols have to use the same modulation as
> the data symbols.  Is there any way that I can use different
> modulations?  say QPSK for training and 16QAM for data.  I tried
> equalizing the two parts separately with a constellation change in
> between.  The shape of the equalized constellation seems correct, but
> it is compressed. Of course the demodulation result is wrong.
> Thanks

 | 
Pages: 1
Prev: estimation parameter using fminsearch
Next: Error, (in int) integer too large in context