[opendtv] Re: Mobile ATSC reception paper

  • From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Mon, 12 Dec 2005 11:21:30 -0500

Al Limberg wrote:

> U.S. patent No. 6,975,689 titled DIGITAL MODULATION
> SIGNAL RECEIVER WITH ADAPTIVE CHANNEL EQUALIZATION
> EMPLOYING DISCRETE FOURIER TRANSFORMS will issue
> tomorrow, December 13.  This patent describes Doug
> McDonald's use of DFT on several-thousand-symbol
> sequences of samples of 8VSB signal to characterize
> the reception channel on a continuing basis.
> Long-signature analysis works well because of data
> randomization in the 8VSB signal.  A second patent
> will issue early next year describing later work.
> The several thousand samples do not have to be a
> training signal and can be taken from a sliding
> window, so dynamic multipath can be tracked quite
> closely.  In contradistinction to the prior art this
> is not an incremental approximation procedure, so
> revealed-ray or concealed-ray dynamic multipath does
> not upset the apple cart with regard to channel
> equalization.
>
> Characterization times are about 100 microseconds, as
> I recall.

So what I think you're saying is that a series of 8-VSB
symbols can themselves be used as if they were training
symbols, because the data randomization makes them look
so flat in the frequency domain (in the absence of
multipath distortion).

So, symbols being 0.09 usec apart, you can look at a
sliding window of thousands of symbols and obtain a
good characterization of the dynamic changes in the
channel within a few usec of time. Much more frequently
then possible with the actual training sequence, which
only appears every 24.2 msec.

There is a certain symmetry to this. In COFDM, where
symbols are very slow, you need to depend on many and
frequent sync sequences. But you can also accommodate
these easily, with the multiple active carriers. In
8-VSB, you can't accommodate enough training symbols
without seriously degrading the usable bit rate, but the
short symbol duration can be used to achieve the same
purpose.

Cool.

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: