os.path.join {}
批量处理音频
import os
for i in range(80):
wavpath=os.path.join('D:/speech/test_wav/wav ({}).wav'.format(i+1))
pcmpath=os.path.join('D:/speech/test_pcm/pcm ({}).pcm'.format(i+1))
wav2pcm(wavpath,pcmpath)
print("finish!")
def wav2pcm(wavfile, pcmfile, data_type=np.int16):
f = open(wavfile, "rb")
f.seek(0)
f.read(44)
data = np.fromfile(f, dtype= data_type)
data.tofile(pcmfile)
批量生成特征谱图
for j in range(4):
for i in range(80):
inmat_path=os.path.join('D:/speech/new_feature_train/{}_{}.mat'.format(j+1,i+1))
ucm_seg = loadmat(inmat_path)
feature = sklearn.preprocessing.scale(ucm_seg["mfcc_imfcc"],axis=1)
feature = feature[1:60,1:60]
plt.figure(figsize=(5, 5))
librosa.display.specshow(feature, sr=16000,y_axis='log')
plt.axis('off')
plt.gca().xaxis.set_major_locator(plt.NullLocator())
plt.gca().yaxis.set_major_locator(plt.NullLocator())
out_png_path=os.path.join('D:/speech/sample_set/imfccgraph','{}_{}.png'.format(j+1,i+1))
plt.savefig(out_png_path,bbox_inches='tight',dpi=300,pad_inches=0.0)