[opendtv] Re: CableLabs Proposes Specifications for OTA Reception in Cable Box

  • From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Sun, 11 Mar 2007 18:53:23 -0400

Tom Barry wrote:

> I think you have to measure spectral efficiency as being AFTER
> any correction for errors. That is, subtract both the FEC
> overhead plus any resulting average number of error bits that
> survive FEC. It is only the resulting throughput of good
> original data bits that counts (per Mhz).

Yes, of course, but that's why I said:

"Any minor differences are only caused by error correction mechanisms
layered over the symbol stream. If anything, this almost always results
in the COFDM stream being somewhat lower in spectral efficiency, to gain
a measure of robustness."

If the COFDM example uses 2/3 convolutional FEC and 64-QAM, then any
amount of GI is going to bring its spectral efficiency lower than that
of 8T-VSB. Even a GI of 1/32, which is the minimum possible. The reason
is simple: 8T-VSB uses no GI, and the Reed-Solomon code is essentially
identical between ATSC and DVB-T (208,188 vs 204,188). And other
overhead, such as training sequences, is small enough that it doesn't
change things appreciably.

For example, the baseline used for the French TNT is 8K, 64-QAM, 2/3
FEC, 1/32 GI, although that won't work in the few SFNs they have in
France. Still, as a baseline, big stick only, that works out to 24.13
Mb/s in their 8 MHz channels.

To eliminate any bias caused by the more rapid dropoff of the spectrum
with 8K COFDM compared with 8-VSB, I'll use the entire channel width in
making the comparison. (The actual numbers are a little higher, because
you can't use the entire channel width in practice.)

24.13 / 8 = 3.02 b/s/Hz

19.39 / 6 = 3.23 b/s/Hz

The C/N margin required, given that the FEC is virtually identical
between the two systems, should be very close, and it is. In a gaussian
channel, COFDM set this way needs 16.5 dB of margin, and 8T-VSB needs
15.2 dB. I think the difference is probably caused by the greater
reliance on active carriers in COFDM, compared with 8-VSB.

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:

  • » [opendtv] Re: CableLabs Proposes Specifications for OTA Reception in Cable Box