convert waveform to matrix

convert_wav_to_matrix.py

from torch.autograd import Variable
import librosa
import numpy as np
import torch

N_FFT=2048
def read_audio_spectum(filename):
	x, fs = librosa.load(filename)
	S = librosa.stft(x, N_FFT)
	p = np.angle(S)
	S = np.log1p(np.abs(S))  
	return S, fs

style_audio, style_sr = read_audio_spectum(style_audio_name)
content_audio, content_sr = read_audio_spectum(content_audio_name)

if(content_sr != style_sr):
	raise 'Sampling rates are not same'
	
style_audio = style_audio.reshape([1,1025,style_audio.shape[1]])
content_audio = content_audio.reshape([1,1025,style_audio.shape[1]])

if torch.cuda.is_available():
	style_float = Variable((torch.from_numpy(style_audio)).cuda())
	content_float = Variable((torch.from_numpy(content_audio)).cuda())	
else:
	style_float = Variable(torch.from_numpy(style_audio))
	content_float = Variable(torch.from_numpy(content_audio))