Re: [Discuss-gnuradio] porting ATSC to 2.x and USRP

[Eric Blossom]

On Tue, Mar 21, 2006 at 05:22:04PM -0800, mao mao wrote:
> Sounds good!
> 
> I am still not convinced that USRP can handle ATSC
> demodulation. Please tell me how are you gonna recover
> the carrier with 16Ms data and 4MHz IF? If this
> problem can be solved, the rest are pretty easy.

The USRP can sample, filter and decimate to 8MS/s complex across the
USB.  Per Nyquist, 8MS/s complex is sufficient to represent our 6 MHz
bandwidth of interest.

We just restructure the PLL to accept complex input data, use a
complex NCO and then track the pilot as before.  In fact, we may be
able to use one of the gr_pll_* blocks in gnuradio-core/src/lib/general.

> The reason why I want to buy a working mc4020/mt4937
> combo is that I want to play with REAL HDTV data. This
> is way more exciting than playing with the test data.

Yes. I want to work with real data too.  However, there are lots of
things that can go wrong when using real data.  Antenna pointing,
multipath, equalization, etc.  That's why I like to start with a
controlled environment and reproducible test cases for the entire tx
and rx paths.  Once those are working, I'm all for playing with real
data.  Otherwise, what's the point ;)

> If you can recover the atsc carrier with USRP, I can
> finish the rest of atsc code real quick

OK.

I'll move the stuff into gr-atsc tonight or tomorrow.
I'll send you more info off-list to get you set up as a committer.

Eric