擅长:python、mysql、java
<blockquote>
<p>My question is: which MFCC features should I use for speaker identification?</p>
</blockquote>
<p>我要说的是把它们都用上。从技术上讲,MFCC特性是从不同的滤波器组输出的。很难说它们中的哪一个有用。在</p>
<blockquote>
<p>In addition to this I am unsure on how to implement these features. What I would do is to get the necessary features and make one long vector input for a neural network.</p>
</blockquote>
<p>实际上,当你提取N个样本的MFCC时,你会得到一个类似于<code>N x T x 20</code><code>T</code>的数组,它表示经过MFCC处理后音频信号中的帧数。我建议使用<a href="https://machinelearningmastery.com/sequence-classification-lstm-recurrent-neural-networks-python-keras/" rel="nofollow noreferrer">Sequence classification with LSTM</a>。这样会有更好的结果。在</p>