[opendtv] Re: Sinclair to test OFDM-DTV2 signal
- From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
- To: "opendtv@xxxxxxxxxxxxx" <opendtv@xxxxxxxxxxxxx>
- Date: Sat, 23 Feb 2013 01:13:14 +0000
Al Limberg wrote:
> IMHO though, DVB-T2 is deeply flawed in its use of rotated QAM symbol
> constellation techniques for its plural carriers. The gains noticed
> in such practice are those attributable to transmitting the same
> information twice with appreciable time-diversity. Rotation of QAM
> symbol constellations reduces the size of data-slicing bins during
> QAM de-maping procedures, however, which adversely affects
> performance at low carrier to AWGN ratios.
>
> Single-time retransmission of non-rotated QAM symbol constellations
> is preferable, since surprisingly it allows much larger digital
> payloads in practice for given performance at low carrier to AWGN
> ratios. This is because larger QAM symbol constellations are
> feasible for a given bin size for data-slicing, and the
> two-dimensionality of QAM is exploited.
Interesting tradeoff. Instead of rotating and delaying, to achieve
time-diversity, don't rotate and just send the same info twice.
Don't know if you looked at this (don't worry, it's in English):
http://public.enst-bretagne.fr/~douillar/Publis/ISSSTA2008.pdf
It says first of all that large rotation angles help most only in cases of high
SNR. Which agrees with what you say. And it says that if you take into account
erasures, as you'd see especially in SFNs, the optimal rotation angle becomes
much lower than it should be for single-transmitter in Rayleigh fading. So it
shows a compromise solution where at 256-QAM, the rotation would only be 3.6
degrees. When it would otherwise be 22.5 degrees. Which tells me, you lose most
of that time diversity.
And this doesn't even take into account your concern on the efficiency of the
de-mapper.
Some impressive looking graphs in the article, but they are only for the 16-QAM
case.
All very interesting.
Bert
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts:
