Re: [Discuss-gnuradio] fft_vcc & file_sink

[Brian Padalino]

On Wed, Mar 11, 2009 at 10:02 AM, E. Ornelas <address@hidden> wrote:
>
> Hello!
>
> I'm new to GNU Radio and USRP.
>
> I've done this simple code below.
>
> class topBlock(gr.top_block):
>        def __init__(self):
>                gr.top_block.__init__(self)
>
>                fftsize = 256
>                udecim = 256
>
>                mywin = window.blackmanharris(fftsize)
>                fft = gr.fft_vcc(fftsize, True, mywin)
>
>                signal = usrp.source_c(0,udecim)
>
>                ss2v = gr.stream_to_vector(gr.sizeof_gr_complex, fftsize)
>                v2ss = gr.vector_to_stream(gr.sizeof_gr_complex, fftsize)
>
>                c2f = gr.complex_to_float(fftsize)
>
>                v_sink = gr.vector_sink_c()
>                f_sink = gr.file_sink(gr.sizeof_gr_complex, "fft.data")
>
>                self.connect(signal, ss2v, fft, v2ss, f_sink)
>
> def main():
>        print "Initilizing.."
>        tb = topBlock()
>        print "Flowgraph start.."
>        tb.run()
>        print "Exiting.."
>
> if __name__ == "__main__":
>        main()
>
>
> The thing is that I don't know what kind of values is the fft giving to the
> f_sink,
> because if a load them on octave using "y=read_complex_binary(...)", a run
> plot(abs(y(1:256))
> it does come close to what I saw using the GUI version.
> Without looking at the scales there should be a spike in the middle of the
> graph.
>
> http://www.nabble.com/file/p22457062/scrsht.png
>
> Shouldn't the first 256 values be the first fft computed from the flowgraph?

The display (GUI) is probably shifting the output such that 0Hz is in
the center, and you span from -Fs/2 to Fs/2 as opposed to a 0Hz to Fs
span.

Take a gander at fftshift() in octave.

> Thank you in advance!

You're welcome and good luck.

Brian