def __init__(self,freq,N_samples):
QtGui.QMainWindow.__init__(self)
self.ui = Ui_MainWindow()
self.ui.setupUi(self)
self.createQtConnections()
self.sample_rate = 2.4e5 ###1e6
#self.decim_r1 = 1e6/2e5 # for wideband fm
self.decim_r2 = 2.4e5/48000 # for baseband recovery
self.center_freq = freq #+250e3
self.gain = 38
self.N_samples = N_samples
self.is_sampling = False
self.spectrogram = np.zeros((328,200))
self.chspectrogram = np.zeros((328,200))
self.plspectrogram = np.zeros((164,200))
self.sdr = RtlSdr()
#self.sdr.direct_sampling = 1
self.sdr.sample_rate = self.sample_rate
self.sdr.center_freq = self.center_freq
self.sdr.gain = self.gain
self.pa = pyaudio.PyAudio()
self.stream = self.pa.open( format = pyaudio.paFloat32,
channels = 2,
rate = 48000,
output = True)
adj = 0
hamming = np.kaiser(self.N_samples/4 + adj,1)
lpf = np.append( np.zeros(self.N_samples*3/8),hamming)
self.lpf = np.fft.fftshift(np.append(lpf,np.zeros(self.N_samples*3/8))) #,int(-.25*self.N_samples))
hamming = 10*signal.hamming(self.N_samples/16)
lpf = np.append(np.zeros(self.N_samples*15/32),hamming)
self.lpf_s1 = (np.append(lpf,np.zeros(int(self.N_samples*15/32))))
#self.lpf_s1 = np.roll(temp,int(.5*self.N_samples*67/120))
#self.lpf_s1 += np.roll(temp,int(-.5*self.N_samples*67/120))
self.lpf_s1 = np.fft.fftshift(self.lpf_s1)
#self.lpf_s1 += np.fft.fftshift(self.lpf_s1)
# fig = plt.figure()
# ax = fig.add_subplot(111)
# ax.plot(range(self.lpf_s1.size),self.lpf_s1)
# fig.show()
hamming = 10*signal.hamming(self.N_samples/32)
lpf = np.append(np.zeros(self.N_samples*31/64),hamming)
self.lpf_s2 = (np.append(lpf,np.zeros(int(self.N_samples*31/64))))
#self.lpf_s2 = np.roll(temp,int(.5*self.N_samples*92/120))
#self.lpf_s2 += np.roll(temp,int(-.5*self.N_samples*92/120))
self.lpf_s2 = np.fft.fftshift(self.lpf_s2)
def __init__(self,freq,N_samples):
self.spectrogram = np.zeros((512,400))
self.chspectrogram = np.zeros((512,400))
self.plspectrogram = np.zeros((256,400))
self.cur_spectrogram = self.spectrogram
QtGui.QMainWindow.__init__(self)
self.ui = Ui_MainWindow()
self.ui.setupUi(self)
self.createQtConnections()
ftxt = "%.1f MHz" % (freq/1e6)
self.ui.curFreq.setText(ftxt)
self.sample_rate = 2.4e5 # tried 1.024e6 - not so great
self.decim_r1 = 1 # 1.024e6/2.4e5 # for wideband fm down from sample_rate
self.decim_r2 = 2.4e5/48000 # for baseband recovery
self.center_freq = freq #+250e3
self.gain = 16
self.N_samples = N_samples
self.is_sampling = False
self.sdr = RtlSdr()
#self.sdr.direct_sampling = 1
self.sdr.sample_rate = self.sample_rate
self.sdr.center_freq = self.center_freq
self.sdr.gain = self.gain
self.pa = pyaudio.PyAudio()
self.stream = self.pa.open( format = pyaudio.paFloat32,
channels = 2,
rate = 48000,
output = True)
self.PLL = PhaseLockedLoop(self.N_samples,19000,self.sample_rate)
self.initOpenCV()
self.noisefilt = np.ones(6554)
b,a = signal.butter(1, 2122/48000*2*np.pi, btype='low')
self.demph_zf = signal.lfilter_zi(b, a)
adj = 0
hamming = np.kaiser(self.N_samples/4 + adj,1)
lpf = np.append( np.zeros(self.N_samples*3/8),hamming)
self.lpf = np.fft.fftshift(np.append(lpf,np.zeros(self.N_samples*3/8))) #,int(-.25*self.N_samples))
hamming = 10*signal.hamming(self.N_samples/16)
lpf = np.append(np.zeros(self.N_samples*15/32),hamming)
self.lpf_s1 = (np.append(lpf,np.zeros(int(self.N_samples*15/32))))
#self.lpf_s1 = np.roll(temp,int(.5*self.N_samples*67/120))
#self.lpf_s1 += np.roll(temp,int(-.5*self.N_samples*67/120))
self.lpf_s1 = np.fft.fftshift(self.lpf_s1)
#self.lpf_s1 += np.fft.fftshift(self.lpf_s1)
# fig = plt.figure()
# ax = fig.add_subplot(111)
# ax.plot(range(self.lpf_s1.size),self.lpf_s1)
# fig.show()
hamming = 10*signal.hamming(self.N_samples/32)
lpf = np.append(np.zeros(self.N_samples*31/64),hamming)
self.lpf_s2 = (np.append(lpf,np.zeros(int(self.N_samples*31/64))))
#self.lpf_s2 = np.roll(temp,int(.5*self.N_samples*92/120))
#self.lpf_s2 += np.roll(temp,int(-.5*self.N_samples*92/120))
self.lpf_s2 = np.fft.fftshift(self.lpf_s2)
class FMRadio(QtGui.QMainWindow,Ui_MainWindow):
sample_buffer = Queue.Queue(maxsize=10)
base_spectrum = np.ones(5780)
plotOverall = True
plotChannel = False
plotPlaying = False
plotWaveform = False
useStereo = False
stereoWidth = 10
useMedianFilt = True
useLPFilt = True
demodFiltSize = 100000
useAudioFilter = True
audioFilterSize = 16
toDraw = True
demodMain = True
demodSub1 = False
demodSub2 = False
toDrawWaterfalls = True
toDrawPlots = False
prevCutoff = 0
prevSpec1 = np.zeros(128)
prevSpec2 = np.zeros(128)
prevSpec3 = np.zeros(128)
prevConvo1 = np.zeros(128)
prevConvo2 = np.zeros(128,dtype='complex')
limiterMax = np.zeros(32)
toPlot = (np.cumsum(np.ones(5780)),np.cumsum(np.ones(5780)))
def __init__(self,freq,N_samples):
self.spectrogram = np.zeros((512,400))
self.chspectrogram = np.zeros((512,400))
self.plspectrogram = np.zeros((256,400))
self.cur_spectrogram = self.spectrogram
QtGui.QMainWindow.__init__(self)
self.ui = Ui_MainWindow()
self.ui.setupUi(self)
self.createQtConnections()
ftxt = "%.1f MHz" % (freq/1e6)
self.ui.curFreq.setText(ftxt)
self.sample_rate = 2.4e5 # tried 1.024e6 - not so great
self.decim_r1 = 1 # 1.024e6/2.4e5 # for wideband fm down from sample_rate
self.decim_r2 = 2.4e5/48000 # for baseband recovery
self.center_freq = freq #+250e3
self.gain = 16
self.N_samples = N_samples
self.is_sampling = False
self.sdr = RtlSdr()
#self.sdr.direct_sampling = 1
self.sdr.sample_rate = self.sample_rate
self.sdr.center_freq = self.center_freq
self.sdr.gain = self.gain
self.pa = pyaudio.PyAudio()
self.stream = self.pa.open( format = pyaudio.paFloat32,
channels = 2,
rate = 48000,
output = True)
self.PLL = PhaseLockedLoop(self.N_samples,19000,self.sample_rate)
self.initOpenCV()
self.noisefilt = np.ones(6554)
b,a = signal.butter(1, 2122/48000*2*np.pi, btype='low')
self.demph_zf = signal.lfilter_zi(b, a)
adj = 0
hamming = np.kaiser(self.N_samples/4 + adj,1)
lpf = np.append( np.zeros(self.N_samples*3/8),hamming)
self.lpf = np.fft.fftshift(np.append(lpf,np.zeros(self.N_samples*3/8))) #,int(-.25*self.N_samples))
hamming = 10*signal.hamming(self.N_samples/16)
lpf = np.append(np.zeros(self.N_samples*15/32),hamming)
self.lpf_s1 = (np.append(lpf,np.zeros(int(self.N_samples*15/32))))
#self.lpf_s1 = np.roll(temp,int(.5*self.N_samples*67/120))
#self.lpf_s1 += np.roll(temp,int(-.5*self.N_samples*67/120))
self.lpf_s1 = np.fft.fftshift(self.lpf_s1)
#self.lpf_s1 += np.fft.fftshift(self.lpf_s1)
# fig = plt.figure()
# ax = fig.add_subplot(111)
# ax.plot(range(self.lpf_s1.size),self.lpf_s1)
# fig.show()
hamming = 10*signal.hamming(self.N_samples/32)
lpf = np.append(np.zeros(self.N_samples*31/64),hamming)
self.lpf_s2 = (np.append(lpf,np.zeros(int(self.N_samples*31/64))))
#self.lpf_s2 = np.roll(temp,int(.5*self.N_samples*92/120))
#self.lpf_s2 += np.roll(temp,int(-.5*self.N_samples*92/120))
self.lpf_s2 = np.fft.fftshift(self.lpf_s2)
# Not currently used
def getSamples(self):
return self.sdr.read_samples(self.N_samples);
def getSamplesAsync(self):
#Asynchronous call. Initiates a continuous loop with the callback fn
self.is_sampling = True
samples = self.sdr.read_samples_async(self.sampleCallback,self.N_samples,context=self)
def sampleCallback(self,samples,sself):
self.is_sampling = False
self.sample_buffer.put(samples)
#print 'put some samples in the jar'
# recursive loop
#sself.getSamplesAsync()
def demodulate_th(self):
# Initiates a loop to process all the incoming blocks of samples from the Queue
# This should be run in its own thread, or the program will become unresponsive
while(1):
try:
samples = self.sample_buffer.get()
#samples2 = self.sample_buffer.get()
except:
#print "wtf idk no samples?" # even though this can't happen... (although I'm not sure why not)
#print 'gonna try to finish off the to-do list'
#self.sample_buffer.join()
break
out1 = self.demodulate(samples)
self.sample_buffer.task_done()
#out2 = self.demodulate(samples2)
#self.sample_buffer.task_done()
audio_out = out1 #np.append(out1,out2)
self.play(audio_out)
def gen_spectrogram(self,x,m,prevSpec):
itsreal = np.isreal(x[0])
m = int(m)
lx = x.size
nt = (lx) // m
#NT = (lx +m -1)//m
#padsize = NT*m -lx
cutsize = lx -nt*m
padsize = int(cutsize + m/2)
if not prevSpec.size == padsize:
prevSpec = np.zeros(padsize)
xp = np.append(x[cutsize:],prevSpec)
prevSpec = x[:(padsize)]
xh = np.zeros((m,nt*2), dtype='complex')
for n in range(int(nt*2)):
block = xp[m*n//2:m*(n+2)//2]
xh[:,n] = block*np.hanning(block.size)
#if self.prevSpec.size == padsize:
# xb = np.append(self.prevSpec[:prevSpec.size-m/2],x)
# xc = np.append(self.prevSpec,x[:lx-m/2])
# self.prevSpec = x[lx-m/2:]
#else:
# xb = np.append(x,np.zeros(-lx+nt*m))
# xc = np.append(x[m/2:],np.zeros(nt*m - lx + m/2))
#xr = np.reshape(xb, (m,nt), order='F') * np.outer(np.hanning(m),np.ones(nt))
#xs = np.reshape(xc, (m,nt), order='F') * np.outer(np.hanning(m),np.ones(nt))
#xm = np.zeros((m,2*nt),dtype='complex')
#xm[:,::2] = xr
#xm[:,1::2] = xs
if itsreal:
spec = np.fft.fft(xh,m,axis=0)
spec = spec[:m//2,:]
else:
spec = np.fft.fftshift(np.fft.fft(xh,m,axis=0))
#mx = np.max(spec)
pwr = np.log(np.abs(spec) + 1e-6)
return (np.real(pwr),prevSpec)
def initOpenCV(self):
cv2.namedWindow("Spectrogram")
def demodulate(self,samples):
# DEMODULATION CODE - And the core function
# samples must be passed in by the caller
self.count += 1
#spectral_window = signal.hanning
#spectrum = np.fft.fftshift(np.fft.fft(samples*spectral_window(samples.size)))
self.spectrogram = np.roll(self.spectrogram, 16,axis=1)
stft,self.prevSpec1 = self.gen_spectrogram(samples,samples.size//8,self.prevSpec1)
self.spectrogram[:,:16] = stft[::8,:] # np.log(np.abs(spectrum[::100]))
if(self.plotOverall): # and self.count % 10 == 9):
#self.drawSpectrum()
self.cur_spectrogram = self.spectrogram
self.drawCurSpectrum()
# cutoff = self.demodFiltSize
# h = signal.firwin(128, cutoff,nyq=self.sample_rate/2)
# lp = signal.fftconvolve(samples[::self.decim_r1],h,mode='full')
# lps = lp.size
# hs = h.size
# prev = lp[lps-hs:]
# lp[:hs/2] += self.prevConvo2[hs/2:]
# lp = np.append(self.prevConvo2[:hs/2],lp)
# self.prevConvo2 = prev
# lp_samples = lp[:lps-hs+1]
lp_samples = samples
power = np.abs(self.mad(lp_samples))
self.ui.signalMeter.setValue(20*(np.log10(power)))
# polar discriminator
dphase = np.zeros(lp_samples.size, dtype='complex')
A = lp_samples[1:lp_samples.size]
B = lp_samples[0:lp_samples.size-1]
dphase[1:] = ( A * np.conj(B) )
dphase[0] = lp_samples[0] * np.conj(self.prevCutoff) #dphase[dphase.size-2]
self.prevCutoff = lp_samples[lp_samples.size-1]
# limiting
dphase /= np.abs(dphase)
# if self.useMedianFilt:
# rebuilt = signal.medfilt(np.angle(dphase)/np.pi,self.demodFiltSize) # np.cos(dphase)
# else:
# rebuilt = self.lowpass(np.angle(dphase),self.demodFiltSize)
rebuilt = np.real(np.angle(dphase) / np.pi)
demodMain = False
demodSub1 = False
demodSub2 = False
isStereo = False
if self.demodMain:
demodMain = True
if self.useStereo:
isStereo = True
elif self.demodSub1:
demodSub1 = True
elif self.demodSub2:
demodSub2 = True
#spectrum = np.fft.fft(rebuilt* spectral_window(rebuilt.size))
#print rebuilt.size/8
self.chspectrogram = np.roll(self.chspectrogram, 16,axis=1)
stft,self.prevSpec2 = self.gen_spectrogram(rebuilt,rebuilt.size/8,self.prevSpec2)
self.chspectrogram[:,:16] = stft[::-4,:] # np.log(np.abs(spectrum[spectrum.size/2:spectrum.size:50]))
if(self.plotChannel):# and self.count % 10 == 9):
self.cur_spectrogram = self.chspectrogram
self.drawCurSpectrum()
#plotspectrum = np.abs(channel_spectrum[::100])
#self.toPlot = (np.linspace(-np.pi,np.pi,plotspectrum.size),plotspectrum)
#self.replot()
n_z = rebuilt.size
if demodMain:
h = signal.firwin(128,16000,nyq=1.2e5)
output = signal.fftconvolve(rebuilt,h,mode='full')
outputa = output # could be done in place but I'm not concerned with memory
outputa[:h.size/2] += self.prevConvo1[h.size/2:] #add the latter half of tail end of the previous convolution
outputa = np.append(self.prevConvo1[:h.size/2], outputa) # also delayed by half size of h so append the first half
self.prevConvo1 = output[output.size-h.size:] # set the tail for next iteration
output = outputa[:output.size-h.size:self.decim_r2] # chop off the tail and decimate
#stereo_spectrum = spectrum
if isStereo:
#pilot = rebuilt * np.cos(2*np.pi*19/240*(np.r_[0:rebuilt.size]))
h = signal.firwin(512,[18000,20000],pass_zero=False,nyq=1.2e5)
pilot_actual = signal.fftconvolve(rebuilt,h,mode='same')
self.PLL.adjust(pilot_actual)
moddif = rebuilt * np.real(np.square(self.PLL.pll)) #np.cos(2*np.pi*38/240*(np.r_[0:ss] - phase_shift))
h = signal.firwin(128,16000,nyq=1.2e5)
moddif = signal.fftconvolve(moddif,h,mode='same')
h = signal.firwin(64,16000,nyq=48000/2)
diff = signal.fftconvolve(moddif[::self.decim_r2],h,mode='same')
diff = np.real(diff)
# rdbs = rebuilt * np.power(self.PLL.pll,3)
# h = signal.hanning(1024)
# rdbs = signal.fftconvolve(rdbs,h)
# if np.mean(rdbs) > 0:
# # bit ONE
# else:
# # bit ZERO
elif demodSub1:
demod = rebuilt * np.exp(-2j*np.pi*67650/2.4e5*np.r_[0:rebuilt.size])
h = signal.firwin(128,7500,nyq=2.4e5/2)
lp_demod = signal.fftconvolve(demod,h,mode='same')
decim = lp_demod[::self.decim_r2]
dphase = np.zeros(decim.size, dtype='complex')
#
A = decim[1:decim.size]
B = decim[0:decim.size-1]
#
dphase[1:] = np.real(np.angle( A * np.conj(B) ))
h = signal.firwin(128,7500,nyq=24000)
output = signal.fftconvolve(dphase,h,mode='same')
elif demodSub2:
demod = rebuilt * np.exp(-2j*np.pi*92000/2.4e5*np.r_[0:rebuilt.size])
h = signal.firwin(128,7500,nyq=2.4e5/2)
lp_demod = signal.fftconvolve(demod,h,mode='same')
decim = lp_demod[::self.decim_r2]
dphase = np.zeros(decim.size, dtype='complex')
#
A = decim[1:decim.size]
B = decim[0:decim.size-1]
#
dphase[1:] = np.real(np.angle( A * np.conj(B) ))
h = signal.firwin(128,7500,nyq=24000)
output = signal.fftconvolve(dphase,h,mode='same')
# DC block filter, lol, srsly
output = np.real(output) - np.mean(np.real(output))
if np.isnan(output[0]):
#print "error" # for some reason, output is NaN for the first 2 loops
return np.zeros(6554)
# deemphasis - # 2122/samplerate*2 *pi - butterworth filter
b,a = signal.butter(1, 2122/48000*2*np.pi, btype='low')
output, zf = signal.lfilter(b, a, output,zi=self.demph_zf)
self.demph_zf = zf
stereo = np.zeros(output.size*2)
if (isStereo):
diff = signal.lfilter(b,a,diff)
w = self.stereoWidth # adjust to change stereo wideness
left = output + w/10 * diff
right = output - w/10 * diff
if(self.useAudioFilter):
left = self.lowpass(left,self.audioFilterSize)
right = self.lowpass(right,self.audioFilterSize)
stereo[0:stereo.size:2] = left
stereo[1:stereo.size:2] = right
else:
if self.useAudioFilter:
output = self.lowpass(output,self.audioFilterSize) # just the tip (kills the 19k pilot)
stereo[0:stereo.size:2] = output
stereo[1:stereo.size:2] = output
#normalize to avoid any possible clipping when playing
stereo /= 2*np.max(stereo)
#spectrum = np.fft.fft(stereo[::2])
output = .5*(stereo[::2]+stereo[1::2])
#spectrum = np.fft.fft(.5*(stereo[::2]+stereo[1::2])*spectral_window(output.size))
self.plspectrogram = np.roll(self.plspectrogram, 24,axis=1)
stft,self.prevSpec3 = self.gen_spectrogram(output,512,self.prevSpec3)
self.plspectrogram[:,:24] = stft[::-1,:] # np.log(np.abs(spectrum[spectrum.size/2:spectrum.size:20]))
if(self.plotPlaying): # and self.count % 2 == 0):
#if self.toDrawWaterfalls:
self.cur_spectrogram = self.plspectrogram
self.drawCurSpectrum(invert=True)
#self.drawPlspectrum()
#else:
# sm = np.abs(np.fft.fftshift(spectrum[::20]))
# toPlot = (np.linspace(-2.4e4,2.4e4,sm.size),sm)
# self.replot(toPlot)
#if(self.toDraw and self.plotWaveform):
# if self.toDrawWaterfalls:
# sm = np.real(output[::20])
# toPlot = (np.linspace(0,output.size/48000,sm.size),sm)
# self.replot(toPlot)
# else:
# sm = np.real(self.PLL.pll[::200])
# toPlot = (np.linspace(0,output.size/48000,sm.size),sm)
# self.replot(toPlot)
return np.real(stereo)
# Alternate demodulator. Not used, but extremely simple
def demodulate2(self,samples):
# DEMODULATION CODE
# LIMITER goes here
# low pass & down sampling
h = signal.firwin(128,80000,nyq=1.2e5)
lp_samples = signal.fftconvolve(samples, h)
# polar discriminator
A = lp_samples[1:lp_samples.size]
B = lp_samples[0:lp_samples.size-1]
dphase = ( A * np.conj(B) ) / np.pi
dphase.resize(dphase.size+1)
dphase[dphase.size-1] = dphase[dphase.size-2]
h = signal.firwin(128,16000,nyq=1.2e5)
rebuilt = signal.fftconvolve(dphase,h)
output = rebuilt[::self.decim_r2]
output = self.lowpass(output, self.audioFilterSize)
return np.real(output)
# utility functions #
def lowpass(self,x,width):
#wndw = np.sinc(np.r_[-15:16]/np.pi)/np.pi
#wndw = np.kaiser(width,6)
wndw = signal.firwin(16,width*999,nyq=24000)
#wndw /= np.sum(wndw)
new_array = signal.fftconvolve(x, wndw, mode='same')
return new_array
# calculate mean average deviation #
def mad(self,samples):
ave = np.mean(samples)
return np.mean(np.abs(samples-ave))
# calculate rms for power #
def rms(self,samples):
meansq = np.mean(np.square(samples))
return np.sqrt(meansq)
def play(self,samples):
self.stream.write( samples.astype(np.float32).tostring() )
# starting point
def start(self):
# Initiates running things
self.streamer = MakeDaemon(self.demodulate_th) # run demodulation in the 'background'
self.streamer.start()
self.count = 0
self.sampler_t = threading.Thread(target=self.getSamplesAsync) # sampler loop
self.sampler_t.start()
def createQtConnections(self):
QtCore.QObject.connect(self.ui.freqSelect, QtCore.SIGNAL(_fromUtf8("valueChanged(int)")), self.setFreq)
QtCore.QObject.connect(self.ui.checkBox, QtCore.SIGNAL(_fromUtf8("toggled(bool)")), self.setUseStereo)
QtCore.QObject.connect(self.ui.mainchannel, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodMain)
QtCore.QObject.connect(self.ui.subband1, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodSub1)
QtCore.QObject.connect(self.ui.subband2, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodSub2)
QtCore.QObject.connect(self.ui.stereoWidthSlider, QtCore.SIGNAL(_fromUtf8("sliderMoved(int)")), self.setStereoWidth)
QtCore.QObject.connect(self.ui.spectrum_overall, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumOverall)
QtCore.QObject.connect(self.ui.spectrum_channel, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumChannel)
QtCore.QObject.connect(self.ui.spectrum_playing, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumPlaying)
QtCore.QObject.connect(self.ui.spectrum_waveform, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumWaveform)
QtCore.QObject.connect(self.ui.demodFiltMedian, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodFiltMedian)
QtCore.QObject.connect(self.ui.demodFiltLP, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodFiltLP)
QtCore.QObject.connect(self.ui.demodFilterSize, QtCore.SIGNAL(_fromUtf8("sliderMoved(int)")), self.setDemodFiltSize)
QtCore.QObject.connect(self.ui.audioFilterActive, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setAudioFiltUse)
QtCore.QObject.connect(self.ui.audioFilterSizeSlider, QtCore.SIGNAL(_fromUtf8("sliderMoved(int)")), self.setAudioFiltSize)
QtCore.QObject.connect(self.ui.exitButton, QtCore.SIGNAL(_fromUtf8("clicked()")), self.terminate)
QtCore.QObject.connect(self.ui.drawPlot, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDrawSpec)
QtCore.QObject.connect(self.ui.waterfallButton, QtCore.SIGNAL(_fromUtf8('toggled(bool)')), self.setDrawWaterfalls)
# QtCore.QObject.connect(self.ui.plotButton, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDrawPlot)
self.bindPlot()
def bindPlot(self):
self.dpi = 100
self.fig = Figure((4.31,2.0), dpi=self.dpi)
self.canvas = FigureCanvas(self.fig)
self.canvas.setParent(self.ui.plotFrame)
self.initplot()
def initplot(self):
self.axes = self.fig.add_subplot(111, aspect=200/431)
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.fig.tight_layout()
#self.axes.invert_yaxis()
#self.anim = animation.FuncAnimation(self.fig,self.drawCurSpectrum,interval=750)
def replot(self,toPlot):
self.axes.clear()
self.axes.plot(toPlot[0],toPlot[1])
self.axes.set_aspect('auto',anchor='C')
self.canvas.draw()
def setDrawSpec(self,s):
self.toDraw = s
def drawCurSpectrum(self,invert=False):
#self.axes.clear()
#self.axes.imshow(self.cur_spectrogram, cmap='spectral')
#self.axes.xaxis.set_major_locator(ticker.NullLocator())
#self.axes.yaxis.set_major_locator(ticker.NullLocator())
#self.axes.set_aspect('auto',adjustable='box',anchor='NW')
#self.canvas.draw()
mx = np.max(self.cur_spectrogram)
mn = np.min(self.cur_spectrogram)
if invert:
self.cur_spectrogram = -cv2.convertScaleAbs(self.cur_spectrogram,alpha=255 / (mx))
else:
self.cur_spectrogram = cv2.convertScaleAbs(self.cur_spectrogram,alpha=255 / (mn))
self.cur_spectrogram = cv2.GaussianBlur(self.cur_spectrogram,(5,5),.6)
cmapped =cv2.applyColorMap(self.cur_spectrogram,cv2.COLORMAP_JET)
cv2.imshow('Spectrogram',cmapped)
cv2.waitKey(1);
def drawSpectrum(self):
self.axes.clear()
self.axes.imshow(self.spectrogram, cmap='spectral')
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.axes.set_aspect('auto',adjustable='box',anchor='NW')
self.canvas.draw()
def drawChspectrum(self):
self.axes.clear()
self.axes.imshow(self.chspectrogram, cmap='spectral')
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.axes.set_aspect('auto',adjustable='box',anchor='NW')
self.canvas.draw()
def drawPlspectrum(self):
self.axes.clear()
self.axes.imshow(self.plspectrogram, cmap='spectral')
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.axes.set_aspect('auto',adjustable='box',anchor='NW')
self.canvas.draw()
def setDrawPlots(self,s):
self.toDrawPlots = s
self.toDrawWaterfalls = not s
def setDrawWaterfalls(self,s):
self.toDrawWaterfalls = s
self.toDrawPlots = not s
def setFreq(self,freq):
if freq % 2 == 0:
freq += 1
freq /= 10.0
text = "%.1f MHz" % freq
self.ui.curFreq.setText(text)
self.center_freq = freq*1e6 #+ 250e3
setf_t = threading.Thread(target=self.setF_th, args=[self.center_freq,])
setf_t.start()
setf_t.join()
# This function is what is used to adjust the tuner on the RTL
# Currently, it causes the program to crash if used after an unspecified period of inactivity
# commented lines are attempts that didn't work
def setF_th(self,f):
while(self.is_sampling == True):
pass
#self.sdr.cancel_read_async()
time.sleep(.1)
self.sdr.center_freq = f
#self.getSamplesAsync()
def setUseStereo(self,u):
self.useStereo = u
def setStereoWidth(self,w):
self.stereoWidth = w/5
def setDemodMain(self,s):
self.demodMain = s
self.demodSub1 = not s
self.demodSub2 = not s
#self.useStereo = True
def setDemodSub1(self,s):
self.demodMain = not s
self.demodSub1 = s
self.demodSub2 = not s
#self.useStereo = False
def setDemodSub2(self,s):
self.demodMain = not s
self.demodSub1 = not s
self.demodSub2 = s
#self.useStereo = False
def setSpectrumOverall(self,s):
#self.initplot()
#self.cur_spectrogram = self.spectrogram
self.plotOverall = s
self.plotChannel = not s
self.plotPlaying = not s
self.plotWaveform = not s
def setSpectrumChannel(self,s):
#self.initplot()
self.plotChannel = s
self.plotOverall = not s
self.plotPlaying = not s
self.plotWaveform = not s
def setSpectrumPlaying(self,s):
#self.initplot()
self.plotPlaying = s
self.plotChannel = not s
self.plotOverall= not s
self.plotWaveform = not s
def setSpectrumWaveform(self,s):
self.plotWaveform = s
self.plotPlaying = not s
self.plotChannel = not s
self.plotOverall= not s
def setDemodFiltMedian(self,s):
self.useMedianFilt = s
self.useLPFilt = not s
def setDemodFiltLP(self,s):
self.useLPFilt = s
self.useMedianFilt = not s
def setDemodFiltSize(self,s):
#if(s % 2 == 0):
# s+=1
self.demodFiltSize = s
def setAudioFiltUse(self,s):
self.useAudioFilter = s
def setAudioFiltSize(self,s):
self.audioFilterSize = s
def terminate(self):
self.__del__()
# Destructor - also used to exit the program when user clicks "Quit"
def __del__(self):
# Program will continue running in the background unless the RTL is told to stop sampling
self.sdr.cancel_read_async()
print "sdr closed"
self.sdr.close()
print "pyaudio terminated"
self.pa.terminate()
cv2.destroyAllWindows()
class FMRadio(QtGui.QMainWindow,Ui_MainWindow):
sample_buffer = Queue.Queue(maxsize=10)
base_spectrum = np.ones(5780)
plotOverall = True
plotChannel = False
plotPlaying = False
plotWaveform = False
useStereo = False
stereoWidth = 10
useMedianFilt = True
useLPFilt = False
demodFiltSize = 11
useAudioFilter = True
audioFilterSize = 16
toDraw = True
demodMain = True
demodSub1 = False
demodSub2 = False
toDrawWaterfalls = True
toDrawPlots = False
prevCutoff = 0
toPlot = (np.cumsum(np.ones(5780)),np.cumsum(np.ones(5780)))
def __init__(self,freq,N_samples):
QtGui.QMainWindow.__init__(self)
self.ui = Ui_MainWindow()
self.ui.setupUi(self)
self.createQtConnections()
self.sample_rate = 2.4e5 ###1e6
#self.decim_r1 = 1e6/2e5 # for wideband fm
self.decim_r2 = 2.4e5/48000 # for baseband recovery
self.center_freq = freq #+250e3
self.gain = 38
self.N_samples = N_samples
self.is_sampling = False
self.spectrogram = np.zeros((328,200))
self.chspectrogram = np.zeros((328,200))
self.plspectrogram = np.zeros((164,200))
self.sdr = RtlSdr()
#self.sdr.direct_sampling = 1
self.sdr.sample_rate = self.sample_rate
self.sdr.center_freq = self.center_freq
self.sdr.gain = self.gain
self.pa = pyaudio.PyAudio()
self.stream = self.pa.open( format = pyaudio.paFloat32,
channels = 2,
rate = 48000,
output = True)
adj = 0
hamming = np.kaiser(self.N_samples/4 + adj,1)
lpf = np.append( np.zeros(self.N_samples*3/8),hamming)
self.lpf = np.fft.fftshift(np.append(lpf,np.zeros(self.N_samples*3/8))) #,int(-.25*self.N_samples))
hamming = 10*signal.hamming(self.N_samples/16)
lpf = np.append(np.zeros(self.N_samples*15/32),hamming)
self.lpf_s1 = (np.append(lpf,np.zeros(int(self.N_samples*15/32))))
#self.lpf_s1 = np.roll(temp,int(.5*self.N_samples*67/120))
#self.lpf_s1 += np.roll(temp,int(-.5*self.N_samples*67/120))
self.lpf_s1 = np.fft.fftshift(self.lpf_s1)
#self.lpf_s1 += np.fft.fftshift(self.lpf_s1)
# fig = plt.figure()
# ax = fig.add_subplot(111)
# ax.plot(range(self.lpf_s1.size),self.lpf_s1)
# fig.show()
hamming = 10*signal.hamming(self.N_samples/32)
lpf = np.append(np.zeros(self.N_samples*31/64),hamming)
self.lpf_s2 = (np.append(lpf,np.zeros(int(self.N_samples*31/64))))
#self.lpf_s2 = np.roll(temp,int(.5*self.N_samples*92/120))
#self.lpf_s2 += np.roll(temp,int(-.5*self.N_samples*92/120))
self.lpf_s2 = np.fft.fftshift(self.lpf_s2)
def createQtConnections(self):
QtCore.QObject.connect(self.ui.freqSelect, QtCore.SIGNAL(_fromUtf8("valueChanged(int)")), self.setFreq)
QtCore.QObject.connect(self.ui.checkBox, QtCore.SIGNAL(_fromUtf8("toggled(bool)")), self.setUseStereo)
QtCore.QObject.connect(self.ui.mainchannel, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodMain)
QtCore.QObject.connect(self.ui.subband1, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodSub1)
QtCore.QObject.connect(self.ui.subband2, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodSub2)
QtCore.QObject.connect(self.ui.stereoWidthSlider, QtCore.SIGNAL(_fromUtf8("sliderMoved(int)")), self.setStereoWidth)
QtCore.QObject.connect(self.ui.spectrum_overall, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumOverall)
QtCore.QObject.connect(self.ui.spectrum_channel, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumChannel)
QtCore.QObject.connect(self.ui.spectrum_playing, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumPlaying)
QtCore.QObject.connect(self.ui.spectrum_waveform, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setSpectrumWaveform)
QtCore.QObject.connect(self.ui.demodFiltMedian, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodFiltMedian)
QtCore.QObject.connect(self.ui.demodFiltLP, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDemodFiltLP)
QtCore.QObject.connect(self.ui.demodFilterSize, QtCore.SIGNAL(_fromUtf8("sliderMoved(int)")), self.setDemodFiltSize)
QtCore.QObject.connect(self.ui.audioFilterActive, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setAudioFiltUse)
QtCore.QObject.connect(self.ui.audioFilterSizeSlider, QtCore.SIGNAL(_fromUtf8("sliderMoved(int)")), self.setAudioFiltSize)
QtCore.QObject.connect(self.ui.exitButton, QtCore.SIGNAL(_fromUtf8("clicked()")), self.terminate)
QtCore.QObject.connect(self.ui.drawPlot, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDrawSpec)
QtCore.QObject.connect(self.ui.waterfallButton, QtCore.SIGNAL(_fromUtf8('toggled(bool)')), self.setDrawWaterfalls)
# QtCore.QObject.connect(self.ui.plotButton, QtCore.SIGNAL(_fromUtf8("clicked(bool)")), self.setDrawPlot)
self.bindPlot()
def bindPlot(self):
self.dpi = 100
self.fig = Figure((4.31,2.0), dpi=self.dpi)
self.canvas = FigureCanvas(self.fig)
self.canvas.setParent(self.ui.plotFrame)
self.initplot()
#self.anim = animation.FuncAnimation(self.fig,self.replot,interval=2000,blit=False)
#self.canvas.mpl_connect('click_event',self.on_click_plot)
# self.curve = QwtPlotCurve("Frequencies")
# self.curve.setData(self.toPlot[0],self.toPlot[1]) #np.r_[-5e5:5e5:1e6/self.toPlot.size],np.abs(self.toPlot))
# self.curve.attach(self.ui.spectrumPlot)
# self.ui.spectrumPlot.replot()
# #threading.Timer(1,self.replot).start()
def initplot(self):
self.axes = self.fig.add_subplot(111, aspect=200/431)
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
#self.axes.invert_yaxis()
def replot(self):
self.axes.clear()
#self.axes.set_s
self.axes.plot(self.toPlot[0],self.toPlot[1])
self.axes.set_aspect('auto',anchor='C')
self.canvas.draw()
def setDrawSpec(self,s):
self.toDraw = s
def drawSpectrum(self):
self.axes.clear()
self.axes.imshow(self.spectrogram, cmap='spectral')
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.axes.set_aspect('auto',adjustable='box',anchor='NW')
self.canvas.draw()
def drawChspectrum(self):
self.axes.clear()
self.axes.imshow(self.chspectrogram, cmap='spectral')
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.axes.set_aspect('auto',adjustable='box',anchor='NW')
self.canvas.draw()
def drawPlspectrum(self):
self.axes.clear()
self.axes.imshow(self.plspectrogram, cmap='spectral')
self.axes.xaxis.set_major_locator(ticker.NullLocator())
self.axes.yaxis.set_major_locator(ticker.NullLocator())
self.axes.set_aspect('auto',adjustable='box',anchor='NW')
self.canvas.draw()
def setDrawPlots(self,s):
self.toDrawPlots = s
self.toDrawWaterfalls = not s
def setDrawWaterfalls(self,s):
self.toDrawWaterfalls = s
self.toDrawPlots = not s
def setFreq(self,freq):
freq /= 10.0
text = "%.1f MHz" % freq
self.ui.curFreq.setText(text)
self.center_freq = freq*1e6 #+ 250e3
setf_t = threading.Thread(target=self.setF_th, args=[self.center_freq,])
setf_t.start()
setf_t.join()
def setF_th(self,f):
while(self.is_sampling == True):
pass
self.sdr.center_freq = f
def setUseStereo(self,u):
self.useStereo = u
def setStereoWidth(self,w):
self.stereoWidth = np.sqrt(10*w)
def setDemodMain(self,s):
self.demodMain = s
self.demodSub1 = not s
self.demodSub2 = not s
#self.useStereo = True
def setDemodSub1(self,s):
self.demodMain = not s
self.demodSub1 = s
self.demodSub2 = not s
#self.useStereo = False
def setDemodSub2(self,s):
self.demodMain = not s
self.demodSub1 = not s
self.demodSub2 = s
#self.useStereo = False
def setSpectrumOverall(self,s):
#self.initplot()
self.plotOverall = s
self.plotChannel = not s
self.plotPlaying = not s
self.plotWaveform = not s
def setSpectrumChannel(self,s):
#self.initplot()
self.plotChannel = s
self.plotOverall = not s
self.plotPlaying = not s
self.plotWaveform = not s
def setSpectrumPlaying(self,s):
#self.initplot()
self.plotPlaying = s
self.plotChannel = not s
self.plotOverall= not s
self.plotWaveform = not s
def setSpectrumWaveform(self,s):
self.plotWaveform = s
self.plotPlaying = not s
self.plotChannel = not s
self.plotOverall= not s
def setDemodFiltMedian(self,s):
self.useMedianFilt = s
self.useLPFilt = not s
def setDemodFiltLP(self,s):
self.useLPFilt = s
self.useMedianFilt = not s
def setDemodFiltSize(self,s):
if(s % 2 == 0):
s+=1
self.demodFiltSize = s
def setAudioFiltUse(self,s):
self.useAudioFilter = s
def setAudioFiltSize(self,s):
self.audioFilterSize = s
def terminate(self):
self.__del__()
def __del__(self):
#QtGui.QMainWindow.__del__()
#Ui_MainWindow.__del__()
#self.streamer.stop()
#self.sampler_t.stop()
print "sdr closed"
self.sdr.close()
print "pyaudio terminated"
self.pa.terminate()
sys.exit()
def getSamples(self):
#N_samples = self.N_samples # 1/24.4 seconds ~46336 #approximately a 2048/44100 amount's time
return self.sdr.read_samples(self.N_samples)
def getSamplesAsync(self):
#Asynchronous call. Meant to be put in a loop w/ a calback fn
#print 'gonna sample'
self.is_sampling = True
samples = self.sdr.read_samples_async(self.sampleCallback,self.N_samples,context=self)
def sampleCallback(self,samples,sself):
self.is_sampling = False
self.sample_buffer.put(samples)
#print 'put some samples in the jar'
# recursive loop
#sself.getSamplesAsync()
def demodulate_th(self):
while(1):
try:
samples = self.sample_buffer.get()
#samples2 = self.sample_buffer.get()
# print 'gottum'
except:
print "wtf idk no samples?"
break
out1 = self.demodulate(samples)
self.sample_buffer.task_done()
#out2 = self.demodulate(samples2)
#self.sample_buffer.task_done()
audio_out = out1 #np.append(out1,out2)
self.play(audio_out)
print 'gonna try to finish off the to-do list'
sample_buffer.join()
def demodulate(self,samples):
# DEMODULATION CODE
#samples = #self.sample_buffer.get()
# LIMITER goes here
# low pass & down sampling via fft
spectrum = np.fft.fftshift(np.fft.fft(samples))
self.spectrogram = np.roll(self.spectrogram, 1,axis=1)
self.spectrogram[:,0] = np.log(np.abs(spectrum[::100]))
if(self.toDraw and self.plotOverall and self.count % 10 == 9):
# self.toPlot = (np.linspace(-5e5,5e5,spectrum.size),np.abs(spectrum))
# self.replot()
self.drawSpectrum()
self.count += 1
#fig = plt.figure()
#plt.plot(np.abs(spectrum))
#plt.show()
#
# spectrum *= self.lpf
# # Decimate in two rounds. One to 200k, another to 44.1k
# # DECIMATE HERE. Note that we're looking at 1MHz bandwidth.
# n_s = spectrum.size
#
# channel_spectrum = spectrum #.25*spectrum[int(n_s*.75-.5*n_s/self.decim_r1):int(.75*n_s+.5*n_s/self.decim_r1)] #np.append(spectrum[0:int(n_s/self.decim_r1*.5)],spectrum[n_s-int(n_s/self.decim_r1*.5):n_s])
#
# #radio_spectrum -= np.mean(radio_spectrum) #attempt to remove dc bias
# #print channel_spectrum.size
# # fig = plt.figure()
# # plt.plot(np.abs(np.fft.ifftshift(channel_spectrum)))
# # plt.show()
# #self.fig.plot(np.fft.ifftshift(channel_spectrum))
#
#
# lp_samples = np.fft.ifft(np.fft.ifftshift(channel_spectrum))
#
lp_samples = samples
# lp_samples /= power
# polar discriminator
dphase = np.zeros(lp_samples.size, dtype='complex')
A = lp_samples[1:lp_samples.size]
B = lp_samples[0:lp_samples.size-1]
dphase[1:] = ( A * np.conj(B) )
dphase[0] = lp_samples[0] * np.conj(self.prevCutoff) #dphase[dphase.size-2]
self.prevCutoff = lp_samples[lp_samples.size-1]
# if self.useMedianFilt:
# rebuilt = signal.medfilt(np.angle(dphase)/np.pi,self.demodFiltSize) # np.cos(dphase)
# else:
# rebuilt = self.lowpass(np.angle(dphase),self.demodFiltSize)
rebuilt = np.angle(dphase) / np.pi
# toplot = False
# if toplot:
# fig = plt.figure()
# ax = fig.add_subplot(111)
# ax.plot(rebuilt)
power = np.abs(self.mad(lp_samples))
self.ui.signalMeter.setValue(20*(np.log10(power)))
demodMain = False
demodSub1 = False
demodSub2 = False
if self.demodMain:
demodMain = True
# lp_samples = samples
elif self.demodSub1:
demodSub1 = True
#lp_samples = samples * np.exp(-1j*2*np.pi*67650/2.4e5*np.r_[0:samples.size])
elif self.demodSub2:
demodSub2 = True
#lp_samples = samples * np.exp(-1j*2*np.pi*92000/2.4e5*np.r_[0:samples.size])
spectrum = np.fft.fft(rebuilt)
#toplot = self.plotChannel
self.chspectrogram = np.roll(self.chspectrogram, 1,axis=1)
self.chspectrogram[:,0] = np.log(np.abs(spectrum[spectrum.size/2:spectrum.size:50]))
if(self.toDraw and self.plotChannel and self.count % 10 == 9):
self.drawChspectrum()
#plotspectrum = np.abs(channel_spectrum[::100])
#self.toPlot = (np.linspace(-np.pi,np.pi,plotspectrum.size),plotspectrum)
#self.replot()
isStereo = False
#demodSub1 = False
#demodSub2 = False
n_z = rebuilt.size
if demodMain:
#stereo_spectrum = spectrum
if self.useStereo:
isStereo = True
#modulated = rebuilt * np.exp(-2j*np.pi*38000/2.4e5*np.r_[0:rebuilt.size])
#h = signal.firwin(128,22000,nyq=1.2e5)
#lp_mod = signal.fftconvolve(modulated,h,mode='same')
#decim = lp_mod[::self.decim_r2]
#dphase = np.zeros(decim.size, dtype='complex')
#
# A = decim[1:decim.size]
# B = decim[0:decim.size-1]
#
# dphase[1:] = np.angle( A * np.conj(B) )
#h = signal.firwin(128,22000,nyq=24000)
#diff = dphase# [::self.decim_r2]
mod_spectrum = np.roll(spectrum,int(.5*n_z*38000/120000))
mod_spectrum += np.roll(spectrum,int(-.5*n_z*38000/120000)) #np.fft.fft(modulated)
mod_spectrum *= self.lpf #np.fft.ifftshift(np.hamming(stereo_spectrum.size))
stereo_spectrum = np.append(mod_spectrum[0:np.ceil(n_z/self.decim_r2*.5)],mod_spectrum[n_z-np.ceil(n_z/self.decim_r2*.5):n_z])
diff = np.fft.ifft(stereo_spectrum)
#self.base_spectrum = np.append(spectrum[0:int(n_z/self.decim_r2*.5)],spectrum[n_z-int(n_z/self.decim_r2*.5):n_z])
#output = np.fft.ifft(self.base_spectrum)
h = signal.firwin(128,16000,nyq=1.2e5)
output = signal.fftconvolve(rebuilt,h,mode='same')
output = rebuilt[::self.decim_r2]
# h = signal.firwin(128,16000,nyq=2.4e4)
# output = signal.fftconvolve(output,h,mode='same')
elif demodSub1:
demod = rebuilt * np.exp(-2j*np.pi*67650/2.4e5*np.r_[0:rebuilt.size])
# spectrum = np.fft.fft(demod)*self.lpf_s1
h = signal.firwin(128,7500,nyq=2.4e5/2)
lp_demod = signal.fftconvolve(demod,h,mode='same')
decim = lp_demod[::self.decim_r2]
# base_spectrum = np.append(spectrum[0:int(.5*n_z*7500/2.4e5)],spectrum[n_z-int(.5*n_z*7500/2.4e5):n_z])
# decim = np.fft.ifft(base_spectrum)
dphase = np.zeros(decim.size, dtype='complex')
#
A = decim[1:decim.size]
B = decim[0:decim.size-1]
#
dphase[1:] = np.angle( A * np.conj(B) )
h = signal.firwin(128,7500,nyq=24000)
output = signal.fftconvolve(dphase,h,mode='same')
#retoutput)
elif demodSub2:
demod = rebuilt * np.exp(-2j*np.pi*92000/2.4e5*np.r_[0:rebuilt.size])
# spectrum = np.fft.fft(demod)*self.lpf_s2
h = signal.firwin(128,7500,nyq=2.4e5/2)
lp_demod = signal.fftconvolve(demod,h,mode='same')
decim = lp_demod[::self.decim_r2]
# base_spectrum = np.append(spectrum[0:int(.5*n_z*7500/2.4e5)],spectrum[n_z-int(.5*n_z*7500/2.4e5):n_z])
# decim = np.fft.ifft(base_spectrum)
dphase = np.zeros(decim.size, dtype='complex')
#
A = decim[1:decim.size]
B = decim[0:decim.size-1]
#
dphase[1:] = np.angle( A * np.conj(B) )
h = signal.firwin(128,7500,nyq=24000)
output = signal.fftconvolve(dphase,h,mode='same')
#return np.real(output)
output -= np.mean(output)
stereo = np.zeros(output.size*2, dtype='complex')
if (isStereo):
#diff = np.fft.ifft(stereo_spectrum)
w = self.stereoWidth # adjust to change stereo wideness
#print w
left = output + w/10 * diff
right = output - w/10 * diff
if(self.useAudioFilter):
left = self.lowpass(left,self.audioFilterSize)
right = self.lowpass(right,self.audioFilterSize)
stereo[0:stereo.size:2] = left
stereo[1:stereo.size:2] = right
else:
if self.useAudioFilter:
output = self.lowpass(output,self.audioFilterSize) # just the tip (kills the 19k pilot)
stereo[0:stereo.size:2] = output
stereo[1:stereo.size:2] = output
spectrum = np.fft.fft(stereo[::2])
self.plspectrogram = np.roll(self.plspectrogram, 1,axis=1)
self.plspectrogram[:,0] = np.log(np.abs(spectrum[spectrum.size/2:spectrum.size:20]))
if(self.toDraw and self.plotPlaying): # and self.count % 2 == 0):
if self.toDrawWaterfalls:
self.drawPlspectrum()
else:
sm = np.abs(np.fft.fftshift(spectrum[::20]))
self.toPlot = (np.linspace(-2.4e4,2.4e4,sm.size),sm)
self.replot()
if(self.toDraw and self.plotWaveform):
sm = np.real(stereo[::20])
self.toPlot = (np.linspace(0,output.size/48000,sm.size),sm)
self.replot()
return np.real(stereo)
# def updateimg(self,base_spectrum):
# spectralimg = np.zeros((base_spectrum.size/20,base_spectrum.size/10))
# for i in np.r_[0:base_spectrum.size/10]:
# spectralimg[np.abs(np.fft.fftshift(base_spectrum)[i*10]/10),i] = 1
# cv2.imshow('spectrum',spectralimg)
def demodulate2(self,samples):
# DEMODULATION CODE
# LIMITER goes here
# low pass & down sampling
lp_samples = signal.decimate(self.lowpass_filter(samples,16),int(self.decim_r1))
# polar discriminator
A = lp_samples[1:lp_samples.size]
B = lp_samples[0:lp_samples.size-1]
dphase = ( A * np.conj(B) )
dphase.resize(dphase.size+1)
dphase[dphase.size-1] = dphase[dphase.size-2]
rebuilt = signal.medfilt(np.angle(dphase)/np.pi,15) # np.cos(dphase)
output = signal.decimate(rebuilt,int(self.decim_r2))
return np.real(.5*output)
# utility functions #
def lowpass(self,x,width):
#wndw = np.sinc(np.r_[-15:16]/np.pi)/np.pi
#wndw = np.kaiser(width,6)
wndw = signal.firwin(16,width*990,nyq=24000)
#wndw /= np.sum(wndw)
new_array = signal.fftconvolve(x, wndw, mode='same')
return new_array
# calculate mean average deviation #
def mad(self,samples):
ave = np.mean(samples)
return np.mean(np.abs(samples-ave))
# calculate rms for power #
def rms(self,samples):
meansq = np.mean(np.square(samples))
return np.sqrt(meansq)
def play(self,samples):
self.stream.write( samples.astype(np.float32).tostring() )
def start(self):
self.streamer = MakeDaemon(self.demodulate_th) # run demodulation in the 'background'
self.streamer.start()
self.count = 0
self.sampler_t = threading.Thread(target=self.getSamplesAsync) # sampler loop
self.sampler_t.start()
