RE: [Discuss-gnuradio] Sample rate for source_noise and source_vector

[De Lima Julian]

So I tested but I have another question.
I read that the data through the USB cable (from computer to USRP) are in the 
format I/Q complex. (Dawei Shen tutorial)

My interpolation factor is 512, so the sample rate is 128M/512 = 250kSample/s

I have a data list with 500e3 data. 250e3 are of 1000uV and 250e3 are of 0uV. I 
see with my spectrum analyzer during 1 second some power and during another 
second nothing.

With a vector_source in complex, no problem. But if I specify a vector_source 
in short, during 0.5 second I see power and during another 0.5 second nothing. 
So there is a factor of 2 between the complex and short format.

Therefore with a vector_list in short the sampling rate is doubled? Because in 
I/Q format there are 16bits * 2 (I & Q) and in short only 16bits?

Is it true that when we have a source in short the sampling rate is double 
because we do not use the I/Q format?

Thanks

Julian De Lima

-----Message d'origine-----
De : Eric Blossom [mailto:address@hidden 
Envoyé : jeudi, 26. octobre 2006 15:36
À : De Lima Julian
Cc : address@hidden
Objet : Re: [Discuss-gnuradio] Sample rate for source_noise and source_vector

On Thu, Oct 26, 2006 at 05:33:26PM +0200, De Lima Julian wrote:
> Hello,
> 
> When you specify a signal_source, you have to set up the sample rate.

This is just for convenience and allows you to specify the output
frequency in absolute instead of normalized terms.  If you prefer, you
can of course set the sample_rate to 1.

> But with the noise and vector sources you do not have to specify this
> parameter. 
> 
> What is the sample rate in the computer -> USB when we generate a noise
> or generate data from a list/file?

The sample rate across the USB is determined by the interpolation or
decimation rate programmed into the usrp.

On receive, USB sample rate = 64e6 / decimation_rate
On transmit, USB sample rate = 128e6 / interpolation_rate

Eric