audio_2_z.py

from magenta.models.nsynth import utils
from magenta.models.nsynth.wavenet import fastgen

def wavenet_encode(file_path):
    
    # Load the model weights.
    checkpoint_path = './wavenet-ckpt/model.ckpt-200000'
    
    # Load and downsample the audio.
    neural_sample_rate = 16000
    audio = utils.load_audio(file_path, 
                             sample_length=400000, 
                             sr=neural_sample_rate)
    
    # Pass the audio through the first half of the autoencoder,
    # to get a list of latent variables that describe the sound.
    # Note that it would be quicker to pass a batch of audio
    # to fastgen. 
    encoding = fastgen.encode(audio, checkpoint_path, len(audio))
    
    # Reshape to a single sound.
    return encoding.reshape((-1, 16))
  
# An array of n * 16 frames. 
wavenet_z_data = wavenet_encode(file_path)