Olivier Houot wrote: > According to this article: http://www.dvb.org/news_events/news/first-dvb-t2-plug-fest-su/index.xml > It is not said this is done in an 8 Mhz channel , but I > guess it is implicit. > > Recalling recent discussions, this would translate to a spectrum > efficiency of 5.6 b/s/hz . Yes, and the figure that should go along with that is the C/N margin required for reception for it to have any real meaning. Because DVB-T2 can achieve even more than that, "en principe." For example, when DVB-T is set to the mode used by the French TNT, it achieves just over 3 b/s/Hz with a C/N margin of 16.5 dB in a gaussian channel. That's 64-QAM, 2/3 FEC, 1/32 GI. Instead, if DVB-T2 is used at 256-QAM, 3/4 FEC, and the GI is assumed to be 0, all of which conditions suggest less robust than the current French TNT standard (discounting the benefits of LDPC compared with the Viterbi code used now), then you'll achieve 5.6 b/s/Hz. The graph shown here: http://www.fub.it/public/Morello020408B.pdf slide 35, shows that with DVB-T2, assuming no loss of efficiency due to GI *or* to the active pilots, you need a C/N margin of 20 dB in a gaussian channel to achieve that level of efficiency. So in fact, the figure would be somewhat higher. However, this is better than the best you can do with DVB-T, according to that graph. Again discounting the loss of efficiency due to GI and active pilots, the best you can do with DVB-T at 20 dB of C/N gaussian is just under 5 b/s/Hz. Enfin bref, if you want to use DVB-T2 at 5.6 b/s/Hz, you need to increase the power density criterion, to achieve equal coverage to the existing TNT DVB-T. Or you accept more delicate reception conditions. 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.