Python vggish_input.waveform_to_examples() Examples

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  print('all_examples shape:', all_examples.shape)
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  print('all_labels shape:', all_labels.shape)
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = sine_examples + const_examples + noise_examples
  all_labels = sine_labels + const_labels + noise_labels
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def extract_audioset_embedding():
    """Extract log mel spectrogram features. 
    """
    
    # Arguments & parameters
    mel_bins = vggish_params.NUM_BANDS
    sample_rate = vggish_params.SAMPLE_RATE
    input_len = vggish_params.NUM_FRAMES
    embedding_size = vggish_params.EMBEDDING_SIZE
    
    '''You may modify the EXAMPLE_HOP_SECONDS in vggish_params.py to change the 
    hop size. '''

    # Paths
    audio_path = 'appendixes/01.wav'
    checkpoint_path = os.path.join('vggish_model.ckpt')
    pcm_params_path = os.path.join('vggish_pca_params.npz')
    
    if not os.path.isfile(checkpoint_path):
        raise Exception('Please download vggish_model.ckpt from '
            'https://storage.googleapis.com/audioset/vggish_model.ckpt '
            'and put it in the root of this codebase. ')
        
    if not os.path.isfile(pcm_params_path):
        raise Exception('Please download pcm_params_path from '
        'https://storage.googleapis.com/audioset/vggish_pca_params.npz '
        'and put it in the root of this codebase. ')
    
    # Load model
    sess = tf.Session()
    
    vggish_slim.define_vggish_slim(training=False)
    vggish_slim.load_vggish_slim_checkpoint(sess, checkpoint_path)
    features_tensor = sess.graph.get_tensor_by_name(vggish_params.INPUT_TENSOR_NAME)
    embedding_tensor = sess.graph.get_tensor_by_name(vggish_params.OUTPUT_TENSOR_NAME)
    
    pproc = vggish_postprocess.Postprocessor(pcm_params_path)

    # Read audio
    (audio, _) = read_audio(audio_path, target_fs=sample_rate)
    
    # Extract log mel feature
    logmel = vggish_input.waveform_to_examples(audio, sample_rate)

    # Extract embedding feature
    [embedding_batch] = sess.run([embedding_tensor], feed_dict={features_tensor: logmel})
    
    # PCA
    postprocessed_batch = pproc.postprocess(embedding_batch)
    
    print('Audio length: {}'.format(len(audio)))
    print('Log mel shape: {}'.format(logmel.shape))
    print('Embedding feature shape: {}'.format(postprocessed_batch.shape)) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels) 

def _get_examples_batch():
  """Returns a shuffled batch of examples of all audio classes.

  Note that this is just a toy function because this is a simple demo intended
  to illustrate how the training code might work.

  Returns:
    a tuple (features, labels) where features is a NumPy array of shape
    [batch_size, num_frames, num_bands] where the batch_size is variable and
    each row is a log mel spectrogram patch of shape [num_frames, num_bands]
    suitable for feeding VGGish, while labels is a NumPy array of shape
    [batch_size, num_classes] where each row is a multi-hot label vector that
    provides the labels for corresponding rows in features.
  """
  # Make a waveform for each class.
  num_seconds = 5
  sr = 44100  # Sampling rate.
  t = np.linspace(0, num_seconds, int(num_seconds * sr))  # Time axis.
  # Random sine wave.
  freq = np.random.uniform(100, 1000)
  sine = np.sin(2 * np.pi * freq * t)
  # Random constant signal.
  magnitude = np.random.uniform(-1, 1)
  const = magnitude * t
  # White noise.
  noise = np.random.normal(-1, 1, size=t.shape)

  # Make examples of each signal and corresponding labels.
  # Sine is class index 0, Const class index 1, Noise class index 2.
  sine_examples = vggish_input.waveform_to_examples(sine, sr)
  sine_labels = np.array([[1, 0, 0]] * sine_examples.shape[0])
  const_examples = vggish_input.waveform_to_examples(const, sr)
  const_labels = np.array([[0, 1, 0]] * const_examples.shape[0])
  noise_examples = vggish_input.waveform_to_examples(noise, sr)
  noise_labels = np.array([[0, 0, 1]] * noise_examples.shape[0])

  # Shuffle (example, label) pairs across all classes.
  all_examples = np.concatenate((sine_examples, const_examples, noise_examples))
  all_labels = np.concatenate((sine_labels, const_labels, noise_labels))
  labeled_examples = list(zip(all_examples, all_labels))
  shuffle(labeled_examples)

  # Separate and return the features and labels.
  features = [example for (example, _) in labeled_examples]
  labels = [label for (_, label) in labeled_examples]
  return (features, labels)