[Discuss-gnuradio] Question on output of FFT block

[Brook Lin]

Hi All,

I am not familiar with Python. I got confused on the output of FFT block. I
can record the raw data coming to the USRP. However, I also need the data
coming outof FFT block. For example, when usrp_fft.py is running, it calls
fftsink2.py and give the spectrum plot. I found that in class fft_window in
fftsink2.py file, def set_data() gives the data after FFT block, which is
points[:,1]. Am I right? How can I show the maximum value of points[:,1] in
status bar of FFT_GUI window?

If I change the def set_data() which I attached below, I can print that max
value in the terminal, but I can not update the frequency from GUI window.
Can I do both things, updating freq and updating the max(points[:,1]) value
in GUI status bar not the terminal?

I also tried adding a vector sink in class fft_sink_c, 
self.connect(self, self.s2p, self.one_in_n, self.fft, self.c2mag, self.avg,
self.log, self.sink)

        self.win = fft_window(self, parent, size=size)
        self.set_average(self.average)
self.dst2 = gr.vector_sink_f ()
self.connect(self.log, self.dst2)
However, it doesn't work.

I was stcuk here for a long time. If anyone can help me with this question,
I really appreciate!

Brook
##############
    def set_data (self, evt):
        dB = evt.data
        L = len (dB)

        if self.peak_hold:
            if self.peak_vals is None:
                self.peak_vals = dB
            else:
                self.peak_vals = numpy.maximum(dB, self.peak_vals)
                dB = self.peak_vals

        x = max(abs(self.fftsink.sample_rate),
abs(self.fftsink.baseband_freq))
        #if x >= 1e9:
        #    sf = 1e-9
        #    units = "GHz"
        if x >= 1e6:  #elif
            sf = 1e-6
            units = "MHz"
        else:
            sf = 1e-3
            units = "kHz"

        if self.fftsink.input_is_real:     # only plot 1/2 the points
            x_vals = ((numpy.arange (L/2)
                       * (self.fftsink.sample_rate * sf / L))
                      + self.fftsink.baseband_freq * sf)
            points = numpy.zeros((len(x_vals), 2), numpy.float64)
            points[:,0] = x_vals
            points[:,1] = dB[0:L/2]
        else:
            # the "negative freqs" are in the second half of the array
            x_vals = ((numpy.arange (-L/2, L/2)
                       * (self.fftsink.sample_rate * sf / L))
                      + self.fftsink.baseband_freq * sf)
            points = numpy.zeros((len(x_vals), 2), numpy.float64)
            points[:,0] = x_vals
            points[:,1] = numpy.concatenate ((dB[L/2:], dB[0:L/2]))
        lines = plot.PolyLine (points, colour='BLUE')

        graphics = plot.PlotGraphics ([lines],
                                      title=self.fftsink.title,
                                      xLabel = units, yLabel = "dB")

        self.Draw (graphics, xAxis=None, yAxis=self.y_range)
        self.update_y_range ()

        allpoints=[]    
        allpoints=points[:,1]

        #print 'x_vals',x_vals, len(x_vals)
        print 'points',allpoints, len(allpoints)
        print 'max_point',max(allpoints)

        mag_max=0
        for idx in range(len(allpoints)):
                if allpoints[idx]>mag_max:
                        mag_max = allpoints[idx]
                        idx_mag_max = idx
        print 'mag_max is ', mag_max
        print 'ndx_mag_max is ', idx_mag_max
        deltafreq = (2e6)*(256*(512**(-1)))
        if mag_max>10:
                print 'There is a signal @ delta',deltafreq/1e6, 'MHz'
###########################
-- 
View this message in context: 
http://www.nabble.com/Question-on-output-of-FFT-block-tp15951020p15951020.html
Sent from the GnuRadio mailing list archive at Nabble.com.