hdw09 / filtfilt Goto Github PK

I use C++ and MATLAB to test result of filtfilt(). But unfortunately they are not the same. Please refer to the attached file for result of filtfilt().
C++_filtfilt_output_30.txt
MATLAB_filtfilt_30_result.txt
filtfilt_test_30.m.zip
td_30.txt
filtfilt.zip

I try to compile filtfilt-java on Ubunut 20.04. But it failed to compile with the following error message.
Could you let me know which environment do you use for compiling filtfilt-java?

javac ./Test.java
./Test.java:5: error: cannot find symbol
import com.iiphci.dsp.Filtfilt;
^
symbol: class Filtfilt
location: package com.iiphci.dsp
./Test.java:33: error: cannot find symbol
ArrayList y = Filtfilt.doFiltfilt(B,A,X);
^
symbol: variable Filtfilt
location: class Test
2 errors

I use Java and MATLAB to test result of filtfilt(). But unfortunately they are not the same. Please refer to the attached file for result of filtfilt()
Java_filtfilt_output_30.txt
MATLAB_filtfilt_30_result.txt
filtfilt_test_30.zip
td_30.txt
Test.zip
.

我用的是Python版本filtfilt是scipy.signal.filtfilt函数，然后我发现输出和你的对应不上，然后我就从网上搜索看到了Matlab和Python在滤波过程中的差异，会在函数计算中差0.3个相位差，所以在Python函数中 output = signal.filtfilt(b,a,x,padlen=3*(max(len(b),len(a))-1)
就可以得到一样的输出了。Matlab中不支持更改padlen（参考http://dong.sh/posts/differencefiltfiltfunction/）

Hello, first of all thank you for this java code. However, ı found a bug filter method in Filtfilt class. the part of the bug code is

changeArray(A, A.get(0));
changeArray(B, A.get(0));

when A.get(0) was 1 value then B values does not change.
So ı change it.
New code is

double a0 = A.get(0);
changeArray(A, a0);
changeArray(B, a0);

c++版本我用了一下，设置了5阶36输入，然后输出和MATLAB根本就不同，我还以为是我哪里弄错了，用不同的阶数又试了几遍，还是错的，怎么能这么坑。你发出来不应该检测一下程序的正确性吗。

加班ing。。。

做eeg的信号滤波，用了3s的数据，第一次filter之后就已经出现了0.02的偏差，最后的filtfilt和matlab的filtfilt相差甚远。
我花了点时间做了数据比较，输入的数据是一致的，但是这个偏差值会越来越大，目测是filter中的zi计算导致的。

你好，请问有java实现的butter函数吗，返回参数b和a