Re: [Discuss-gnuradio] Question about sampling colored image with TVRX

[Peter Monta]

Hi Nan,

> 1. For TVRX board, as we know, the output signal is a 6MHz chuck. But 
> in which band is this 6MHz bandwidth located? For example, if the 
> carrier frequency is 477.25MHz, is the output signal from 
> 477.25-1.5MHz to 477.25+4.5MHz, or from 477.25-3MHz to 477.25+3MHz, or 
> in other band?

From 476 to 482 MHz for this channel.  In general, if the video carrier 
is at f, the NTSC
signal will span the 6 MHz interval [f-1.25,f+4.75].

> 2. I analyzed the frequency response of the sampled signal, and found 
> there are two peaks in spectrum. A higher peak is located in DC 
> frequency. Another one, which is a little bit lower, is located at 
> roughly f = 2.5MHz. It is different from what I expected. My original 
> thought is that the second peak should be in f=3.58MHz which is the 
> frequency of the subcarrier carrying color information. Could someone 
> help to explain this?

The largest peak will be the video carrier; there should also be a peak 
at +4.5 MHz
about 15 dB below the video, which is the audio subcarrier.  Yes, you 
should also see
a broad feature around +3.58 MHz, but chroma uses (almost) suppressed 
carrier, so
it won't be as distinct.  If you're really seeing something at +2.5, it 
could be local
interference or IMD in the front end.  Might want to play with the gains 
for best
dynamic range.

For best flatness you'll want to set your frequency so as to center the 
NTSC signal
as a whole in the USRP passband; once you have the digital complex 
baseband signal,
you can translate the video back to centered at DC for carrier 
acquisition and tracking,
VSB filtering, demodulation, and so on.

Cheers,
Peter