Tuesday, November 29, 2011

Re: [Discuss-gnuradio] Architecture of *_rx_cfile.py output vector

On Tue, Nov 29, 2011 at 11:32 AM, Marcus D. Leech <mleech@ripnet.com> wrote:
On 29-11-2011 11:12 AM, Sebastian Döring wrote:
On Mon, 28 Nov 2011 10:32:28 -0500
 "Marcus D. Leech" <mleech@ripnet.com> wrote:
On 28-11-2011 10:21 AM, Sebastian Döring wrote:
Hello List,

just wanted to know exactly how the output vector of ..._rx_cfile.py is structured.
Is the first element of the complex vector v[0] the one at the desired frequency sprecified by "-f FREQ"?

Thanks

Sebastian

'
It's just raw complex-float samples in native-binary format.

The first item is I the second is Q then I then Q, etc.

They're time-series samples, not FFT outputs.




Uh ok -  explains a lot...
Since the data is getting recorded as 32-bit complex float, is "read_complex_binary()" the right octave method to put it into a vector I can use for further processing?
I found a gnuradio page that says "read_short_binary()" is supposed to be the right method, but the output vector does not make any sense to me...

Regards
Sebastian


Yup, read_complex_binary() is the one, as far as I can tell.  I'm not a MatLab/Octave user, but it seems like that reads complex-floats.

Yes, that should work for you. The "read_short_binary()" should be usable if you use the '-s' flag in the uhd_rx_cfile.py to output shorts. I think you'll have to deinterleave the I&Q after you read it in, IIRC.

Tom

No comments:

Post a Comment