[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Gnu Radio NTSC processing

From: Daniel Garcia
Subject: Re: [Discuss-gnuradio] Gnu Radio NTSC processing
Date: Sat, 19 May 2007 20:40:09 -0700 (PDT)

>   I worked with the usrp_tv_rcv.py code, using TVRX
> and BasicRX daughter boards. From what I could see,
> I concluded about the project current status, and I
> would like to confirm this with you. I wonder if
> someone could help me.

I don't understand what you are asking. What did you
conclude? What do you need help with? What do you need
confirmation on? Perhaps we can talk off list (in

>   The implementation, as it is concerned with the
> main properties of NTSC signal (chrominance,
> luminance and synchronization), just works with the
> luminance. Is it right? 

The script you are talking about is trying to
implement a black and white television. It is by
definition luminance only.

> From what I found out,
> Daniel Garcia is working in an implementation of
> synchronization

I did write a couple of different versions of the sync
detector. Each of them had its own problems.

> but I couldn´t run his application.

Why could you not run it? Did it compile? Does they
example run and you get no output (ie black screen)?
What was the compile error or program output on the

What did you use for a signal? Was it an antenna? Did
you use a cable television signal? Did you use a
modulated output from a VCR or similar device?

> And Prateek have been working with chrominance, in
> matlab, but doesn´t have a python version. Is it
> right? Does anyone know any other work related to
> implementation of chrominance and synchronization?

You may have some trouble using the MATLAB based
approach Prateek used for chrominance detection; it is
not easily ported to GNU Radio as a block. You must
reliably detect the phase of the chroma burst for
every line. There may already be a PLL block in GNU
Radio but I don't know what it is called.

>   From the luminance, that is working, I observed
> that the main blocks implemented in Gnu Radio are:
>   - gr.agc_cc: it modifies the input signal gain
> depending on the daughter-board that is used
> (BasicRX and TVRX). When BasicRX is selected, the
> gain ranges from 0 to 20 dB. Using TVRX the gain
> ranges from 0 to 115 dB.
>   - gr.complex_to_mag: the AM demodulator deals with
> a stream of complex samples giving an 1-to-1 output
> stream of floating points. This demodulated signal
> is received by filters varying from 0 to 1,5 Mhz and
> from 0 to 0,5 Mhz, with the aim to eliminate
> sub-products of the demodulation of the I signal
> above of 1,5 MHz and demodulation sub-products of
> the Q signal above of 0,5 MHz.  
>   - gr.add_const_ff: it forces the increase of the
> black levels in the image.  
>   - gr.multiply_const_ff: the I inversor supplies
> two I signals dephased of 180º, that is +I and -I.
> The Q inversor also supplies two Q signals with a
> difference between them of 180º.

You've got the main idea down well except for the
filtering part. Your explanation is confusing. 


 the free Yahoo! toolbar and rest assured with the added security of spyware 

reply via email to

[Prev in Thread] Current Thread [Next in Thread]