def test_shape_bias():
""" test shape bias """
a = np.array([-1.0, -0.5, 0.0, 0.5, 1.0])
o = dsp.shape(a, bias=0.5)
e = np.array([-0.625, -0.5, -0.375, 0.5, 2.875])
dsp.normalize(e)
assert np.allclose(o, e)
def test_downsample():
""" test downsample """
a = np.linspace(-1, 1, 10)
e = np.empty_like(a)
for i in range(10):
if i % 2 == 0:
e[i] = a[i]
else:
e[i] = a[i - 1]
dsp.normalize(e)
dsp.downsample(a, 2)
assert np.all(a == e)
def arb(self, data):
""" Generate an arbitrary wave cycle. The provided data will be
interpolated, if possible, to occupy the correct number of samples for
a single cycle at our reference frequency and then normalized and
scaled as appropriate.
@param data seq : A sequence of samples representing a single cycle
of a wave
"""
interp_y = data
num = interp_y.size
interp_x = np.linspace(0, num, num=num)
interp_xx = np.linspace(0, num, num=self.num_points)
interp_yy = np.interp(interp_xx, interp_x, interp_y)
dsp.normalize(interp_yy)
return interp_yy
def main():
""" main """
# create a signal generator
sig_gen = sig.SigGen()
# create a wave table to store the waves
zwt = wavetable.WaveTable(num_waves=16)
# example 1: generate and save a simple saw wave
# generate a saw using our signal generator and store it in a new Wave
# object.
saw_wave = sig_gen.saw()
# put the saw wave into our wave table.
zwt.waves = [saw_wave]
# as we're only adding one wave to the wave table, only the first slot of
# the resulting oscillator in zebra will contain the saw. the remaining
# slots will be empty, because we haven't added anything to those yet.
render(zwt, 'osc_gen_saw')
# you could fill all 16 slots with the same saw, by repeating it 16 times
zwt.waves = [saw_wave for _ in range(16)]
render(zwt, 'osc_gen_saw_16')
# example 2: morphing between two waveforms we can use up all 16 slots in
# the zebra oscillator, even with fewer than 16 starting waveforms, if we
# use morph() to morph from one waveform to the other, to fill in the
# in-between slots.
# morph from sine to triangle over 16 slots
zwt.waves = sig.morph((sig_gen.sin(), sig_gen.tri()), 16)
render(zwt, 'osc_gen_sin_tri')
# of course, we don't have to use all 16 slots. we could use only the first
# 5, for example.
# morph from sine to triangle over 5 slots
zwt.waves = sig.morph((sig_gen.sin(), sig_gen.tri()), 5)
render(zwt, 'osc_gen_sin_tri_5')
# example 3: morphing between many waveforms
# it is possible to morph between any number of waveforms, to produce
# interpolated waves between the given waves.
# morph between sine, triangle, saw and square over 16 slots
zwt.waves = sig.morph(
(sig_gen.sin(), sig_gen.tri(), sig_gen.saw(), sig_gen.sqr()), 16)
render(zwt, 'osc_gen_sin_tri_saw_sqr')
# example 4: generting arbitrary waves
# a custom signal can be used as an oscillator.
# in this example, one slot is filled with random data, but any data,
# generated or, say, read in from a wav file, can be used.
# the custom signal generator function automatically normaises and scales
# any data you throw at it to the right ranges, which is useful.
zwt.waves = [
sig_gen.arb(np.random.uniform(low=-1, high=1, size=128))
for _ in range(16)
]
render(zwt, 'osc_gen_random')
# example 5: pulse-width modulation
# SigGen has a pulse wave generator too.
# let's use that to make a pwm wavetable.
# pulse widths are between 0 and 1 (0 to 100%). 0 and 1 are silent as the
# pulse is a flat line. so, we want to have 16 different, equally spaced
# pulse widths, increasing in duration, but also avoid any silence:
pws = (i / 17. for i in range(1, 17))
# generate the 16 pulse waves
zwt.waves = [sig_gen.pls(p) for p in pws]
render(zwt, 'osc_gen_pwm')
# example 6: processing wave forms
# the dsp module can be used to process waves in various ways
# let's try downsampling a sine
downsampled = dsp.downsample(sig_gen.sin(), 16)
# that downsampled sine from probably sounds pretty edgy
# let's try that again with some slew this time, to smooth it out a bit
slewed = dsp.slew(dsp.downsample(sig_gen.sin(), 16), 0.8)
# generate a triangle wave and quantize (bit crush) it
quantized = dsp.quantize(sig_gen.tri(), 3)
# applying inverse slew, or overshoot, to a square wave
slewed_square = dsp.slew(sig_gen.sqr(), 0.8, inv=True)
# overshoot might make the wave quieter, so let's normalize it
dsp.normalize(slewed_square)
# morph between the waves over 16 slots
zwt.waves = sig.morph((downsampled, slewed, quantized, slewed_square), 16)
render(zwt, 'osc_gen_dsp')
# example 7: longer wavetables, more processing and writing a wav file
# wavetables can have any number of slots, this one has 120 slots
lwt = wavetable.WaveTable(num_waves=120)
# similarly, a signal generator can generate any number of samples
# a waveform coresponding to the frequency of C3 at 44.1 kHz would
# have approx. 337 samples.
mc_sig_gen = sig.SigGen()
mc_sig_gen.num_points = 337
# create ever-decreasing wave folding distortion over the wavetable
lwt.waves = [
dsp.fold(mc_sig_gen.sin(), (lwt.num_waves - i) / 50.)
for i in range(lwt.num_waves)
]
wavfile.write_wavetable(lwt, os.path.join(make_osc_path(), 'folding.wav'))
# create ever-increasing wave shaping distortion over the wavetable
lwt.waves = [
dsp.shape(mc_sig_gen.sin(), power=i + 1) for i in range(lwt.num_waves)
]
wavfile.write_wavetable(lwt, os.path.join(make_osc_path(), 'shaping.wav'))
def test_normalize():
""" test normalize """
a = np.array([0.0, 1.0, 2.0])
e = np.array([-1.0, 0.0, 1.0])
assert np.all(dsp.normalize(a) == e)
def test_normalize_dc():
""" test normalize with negative input """
a = np.array([0.123, 0.123])
e = np.array([0.0, 0.0])
assert np.all(dsp.normalize(a) == e)
def test_normalize_neg():
""" test normalize with negative input """
a = np.array([-1.0, 0.0])
assert np.amax(dsp.normalize(a)) == 1.0
assert np.amin(dsp.normalize(a)) == -1.0
def test_normalize_zero():
""" test normalize with amplitude zero """
a = np.array([0.0, 0.0])
assert np.amax(dsp.normalize(a)) == 0.0
