我实现MFCC算法与Java。有一个样品code三角过滤器和MFCC在爪哇。这里是链接： MFCC的Java 不过，我应该遵循code，MATLAB写的：的 MFCC Matlab的

I am implementing MFCC algorithm with Java. There is a sample code for triangular filters and MFCC at Java. Here is the link: MFCC Java However I should follow that code written in Matlab: MFCC Matlab

我的问题是，在Matlab的code它谈论性和对数过滤器但没有任何关于在Java的code 。我应该测量的对数的表现 和线性滤波器，但我实现了Java code并没有什么有关。此外，我不明白这些fbins和cbins在Java的code？

My question is that at Matlab code it talks about linear and logarithmic filters however there is nothing about that at Java code. I should measure the performance of logarithmic and linear filters but I implemented that Java code and there is nothing about that. Also I didn't understand what these fbins and cbins at Java code?

推荐答案

Java的code不包括非线性变换，基本映射你的线性频率刻度梅尔规模​​。 'cbins'指的每个滤波器组的中心（频率）。

The Java code does include a non linear transformation, basically mapping your linear frequency scale to the mel-scale. 'cbins' refer to the center (frequency) of each filter bank.

Java的code正是一个MFCC code应该是什么样子。因此，它应该很好地帮助你。

The Java code is exactly what an MFCC code should look like. So, it should work well for you.