0xfe

files = os.listdir("samples")
np.random.shuffle(files)

notes = ['A', 'As', 'B', 'C', 'Cs', 'D', 'Ds', 'E', 'F', 'Fs', 'G', 'Gs']
note_map = {}
for i, note_name in enumerate(notes):
  note_map[note_name] = i

count_by_key = np.zeros((8 * 12))
errors_by_key = np.zeros((8 * 12))

0xfe

import numpy as np
import matplotlib.pyplot as plt

# Generate Sine wave
hz = 20
x = np.arange(0, 1, 0.001) # 1 second
y = np.sin(2 * np.pi * hz * x)
plt.figure(figsize=(10, 3))
plt.plot(x, y)
plt.show()

0xfe

# Generate an integer-period sine wave to
# eliminate spectral leak.
hz = 20
x = np.linspace(0, 2 * np.pi * hz, 128 * hz)
y = np.sin(x)
plt.figure(figsize=(10, 3))
plt.plot(x, y)
plt.show()

# We're only interested in the real components

0xfe

points = 256
y = np.zeros(points, dtype="complex64")
y[20] = np.complex(1,0) # stick a 1 into the 20th position
plt.figure(figsize=(10, 3))
plt.bar(np.arange(100), np.absolute(y)[:100]) # Show the first 100 points
plt.show()

0xfe

points = 256
y = np.zeros(points, dtype="complex64")
y[20] = np.complex(1,0)
ifft_y = fft.irfft(y) * points
plt.figure(figsize=(10, 3))
plt.plot(ifft_y)
plt.show()

0xfe

x = np.linspace(0, 2 * np.pi * hz, points * hz)
y = np.sin(x)
plt.figure(figsize=(10, 3))
plt.bar(np.arange(75), y[:75 * 10:10]) # sample every 10th element for bar graph
plt.show()

fft_y = fft.ifft(y.astype("complex64")) * 2
plt.figure(figsize=(10, 3))
plt.bar(np.arange(100), np.absolute(fft_y[:100]))
plt.show()

0xfe

import numpy as np
import matplotlib.pyplot as plt
from scipy import signal

def spectrogram(xs):
  return signal.spectrogram(xs, 256, window=('hann'), nperseg=256, nfft=256, noverlap=200, mode='complex')

0xfe

def build_training_data(batch_size, sample_size):
  xs = np.random.randn(batch_size, sample_size)
  f, t, ys = spectrogram(xs)

  (num_rows, num_cols) = (ys.shape[1], ys.shape[2])
  print(ys.shape)

  ys = ys.reshape(batch_size, num_rows * num_cols)
  Ys = np.hstack([ys.real, ys.imag])
  return (xs, Ys, num_rows, num_cols)

0xfe

import tensorflow as tf

size = 2048
N = 15000

(xs, ys, rows, cols) = build_training_data(N, size)

model = tf.keras.models.Sequential([
    tf.keras.layers.Dense(rows * cols * 2, input_dim=size, use_bias=False)
])

0xfe

def make_wave():
  # Create four evenly-spaced tones
  waves = []
  hz = 0
  for i in range(0, size, 512):
    hz = hz + 50
    waves.append(np.sin(np.linspace(0, 2 * np.pi * hz, 512)))
  return np.hstack(waves)

wave = make_wave()