def main():
# Start interactive mode
plt.ion()
# Load sample
wavfile = sp.io.wavfile.read('sample.wav')
wavdata = np.float64(wavfile[1])
wavdata /= np.float64(np.iinfo(wavfile[1].dtype).max)
wavdata *= 10.0
stream = Stream(wavfile[0])
stream.write(wavdata)
# Search for training symbol
training_spectrum = np.array([complex() for x in range(64)])
training_spectrum[2 + 4*0] = complex(1, 0)
training_spectrum[2 + 4*1] = complex(1, 0)
training_spectrum[2 + 4*2] = complex(1, 0)
training_spectrum[64 - 2 - 4*2] = complex(1, 0)
training_spectrum[64 - 2 - 4*1] = complex(1, 0)
training_spectrum[64 - 2 - 4*0] = complex(1, 0)
training_signal = np.fft.ifft(training_spectrum)*40.0
training_spectrum /= np.linalg.norm(training_spectrum)
training_signal /= np.linalg.norm(training_signal)
for i in range(len(stream.buffer) - 64):
window_signal = stream.buffer[i: i + 64]
window_signal /= np.linalg.norm(window_signal)
#window_spectrum = np.fft.fft(window_signal)
#window_spectrum /= np.linalg.norm(window_spectrum)
corr = abs(np.correlate(training_signal, window_signal))
print(corr)
return
if(corr > 0.9):
print('Signal @ %s' % str(i/SAMPLE_RATE))
stream.pop(i)
break
# Display signal
spectrum = Spectrum(stream, COMPONENETS)
waveform = Waveform(stream, COMPONENETS)
spectrum.tick()
waveform.tick()
plt.show()
while(True):
plt.pause(0.10)
stream.pop(1)
spectrum.tick()
waveform.tick()
