[opendtv] Re: DVB considers SVC for 1080p and hierarchical modulation

  • From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Wed, 10 Dec 2008 17:18:56 -0500

Olivier Houot wrote:

> Also, the idea of broadcasting a robust stream for difficult
> reception conditions, with additionnal layered information
> in a more fragile but higher bitrate stream permitted by
> hierarchical modulation makes a come back. Tests performed
> with older MPEG2 system were unable to show a clear
> advantage with this method in the past, but it seems the
> efficiency of the new SVC scheme makes it worth to
> reconsider.

I was wondering why that idea had been dropped, years ago. I figured it
was not for the reason you state, but rather that the real emphasis was
on creating as many multicasts as possible, in each multiplex? Any
thought of adding HM costs you, of course, in spoectral efficiency.

> In the same DVB scene issue, they mention an OTA demo of
> DVB-T2 at 36 Mbps.

At a price.

If you check out this:

http://www.fub.it/public/Morello020408B.pdf

Viewgraph 35,

and you consider that 36 Mb/s is in a 7.6 MHz channel (4.7 b/s/Hz), you
can see what mode is best for achieving that spectral efficiency.

Turns out, 64-QAM beats out 256-QAM. And the graph ignores guard
intervals entirely, so actual results will be somewhat less dramatic.

Ignoring the extra cost of a GI, 36 Mb/s with DVB-T2 can be achieved
with about 16.5 dB of C/N in a gaussian channel. That's the same
robustness as you get in the DVB-T 24 Mb/s channels you have with
64-QAM, 2/3 FEC, and the smallest 1/32 GI.

The biggest improvement is therefore attributable to the improved error
correction (LDPC), and maybe also the twisted constellation, and not to
a change to 256-QAM.

Which is why I believe that the best bang for the buck, even in ATSC, is
to make the overall FEC scheme as effective as possible. It may turn out
that the 256-QAM options aren't all that helpful, in practice, for OTA.
Over here, I'd say our biggest concern now is to eek out a lower C/N,
for the existing 3.3 b/s/Hz, to come closer to matching NTSC even when
it's at its worst.

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: