def test_split_to_blocks():
assert np.array([
[ 0, 1, 2, 3, 4, 5, 6, 7],
[ 6, 7, 8, 9, 10, 11, 12, 13],
[12, 13, 14, 15, 16, 17, 18, 19],
[18, 19, 20, 21, 22, 0, 0, 0],
]) == split_to_blocks(np.arange(23), 8, 6)
if __name__ == '__main__':
# Dissonance of two sines f_1 and f_2 is like
# amplitude modulation of abs(f_1 - f_2) onto mean(f_1, f_2)
# The amplitude envelope can be obtained as the
# absolute value of the analytical signal (x + i * h(x))
# where h is the Hilbert transform.
# abs(scipy.signal.hilbert(x))
# It corresponds to the AM demodulation.
# The beating frequency is 2 * (f_1 - f_2), since its absolute
# value has twice higher the period.
t = sample_time(0, 1, 200)
x = sine(t, 10) + sine(t, 12)
e = amplitude_envelope(x)
plot(t, x)
plot(t, e)
# generate_and_play(lambda t: np.sum(sine(t, f) for f in (440, 445)), duration=3)
# derivative of the amplitude envelope
# plot(t[:-1], abs(np.diff(abs(hilbert(x)))))
import matplotlib.pyplot as plt
from scipy.signal import hilbert
from synthesis import sample_time, sine
def amplitude_envelope(x):
return abs(hilbert(x))
if __name__ == '__main__':
# Dissonance of two sines f_1 and f_2 is like
# amplitude modulation of abs(f_1 - f_2) onto mean(f_1, f_2)
# The amplitude envelope can be obtained as the
# absolute value of the analytical signal (x + i * h(x))
# where h is the Hilbert transform.
# abs(scipy.signal.hilbert(x))
# It corresponds to the AM demodulation.
# The beating frequency is 2 * (f_1 - f_2), since its absolute
# value has twice higher the period.
t = sample_time(0, 1, 200)
x = sine(t, 10) + sine(t, 12)
e = amplitude_envelope(x)
plt.plot(t, x)
plt.plot(t, e)
# generate_and_play(lambda t: np.sum(sine(t, f) for f in (440, 445)), duration=3)
# derivative of the amplitude envelope
# plot(t[:-1], abs(np.diff(abs(hilbert(x)))))
def complex_tone(partials):
# partial = ((freq0, amp0), (freq1, amp1), ...)
return lambda t: np.sum(sine(t, freq=freq, amplitude=amp) for (freq, amp) in partials)
def complex_tone(partials):
# partial = ((freq0, amp0), (freq1, amp1), ...)
return lambda t: np.sum(
sine(t, freq=freq, amplitude=amp) for (freq, amp) in partials)