Acoustic echo cancellation in Java -

I am implementing a VoIP application that uses Pure Java, a resonance problem occurs when users ( Built-in microphone mostly on laptops). Do not use the headset when it is not

Currently

is crazy and the bolts of the VoIP application are just plain deaths of Java Media Framework Essentials By the way, I want to perform some digital signal processing on the audio data about the speaker before I output it for it.

  Public synchronize zero addAudioData (byte [] ayAudioData) {m_oBuffer.enqueue (ayAudioData); This.notify (); }  

As you can see that audio data comes in and buffer is pressed into it. This is to meet cyclic connections and allow different packet sizes. It also means that before I run the audio data in the speaker line, I have access to the audio data as much as I need any fancy DSP operation.

I have done an echo canceller job, but this works requires a lot of interactive user input and I want to cancel an automatic echo.

Manually cancel resonance

  remove public static byte [] (int iDelaySamples, float FDAC, byte [] ASEMMLAGE) {M_A DEL BABFER = New Small [iDelaySamples]; M_aySamples = new byte [aySamples.length]; M_fDecay = (float) FDK; System.out.println ("echo delete"); M_iDelayIndex = 0; System.out.println ("sample length: \ t" + aySamples.length); For (int i = 0; i & lt; aySamples.length; i + = 2) {// sample summarize wOldSample = getSample (aySamples, i); // Echo less wNewSample = Delete (small) (wOldSample - fDecay * m_awDelayBuffer [m_iDelayIndex]); Set sample (m_aySamples, i, wNewSample); // update late buffer m_awDelayBuffer [m_iDelayIndex] = wNewSample; M_iDelayIndex ++; If (m_iDelayIndex == m_awDelayBuffer.length) {m_iDelayIndex = 0; }} Return m_aySamples; }   
  Adaptive filter   
 I have read that this is the way to go, in particular, a minimal mean square filter, however, I get trapped Most of the sample code for the above mentioned above are in C and C ++ and they do not translate well into Java.  
 Is anyone advised to apply them to Java? Any other ideas will also be greatly appreciated Thanks in advance.   
 
 
 This is the age! Hope this is also the right square, but you can go there:  
  / ** * This filter usually shares the least preconceived mean square on a * Orange byte. This cancels the actual echo * Echo cancellation occurs with the following formula: * * E = D - x '* W * * E represents the echo-free signal * represents the actual microphone signal with the resonance d. X is the transition of the loudspeaker signal W is an array of adaptive weight * * / public class cNormalisedLeastMeansSquareFilter implements IFilter {private byte [] m_ayEchoFreeSignal; // e private byte [] m_ayEchoSignal; // D Private Byte [] m_ayTransposeOfSpeakerSignal; // x 'private double [] m_adWeights; // w / ** * An echo signal must re-filter transfer before applying updated and weight. * @ Ultimate EECHosignal * @ Param AT Transposhal Signal * @ Param Ed Wites * / Public Literature Last Mensquake Filter (Byte [] AEcocognal, Byte [] AT Transposh Ophspeaker Signals, Double [] Advites) {M_Eichociginal = EECHO Signals; M_ayTransposeOfSpeakerSignal = ayTransposeOfSpeakerSignal; M_adWeights = adWeights; } @Override [public byte] applyFilter (byte [] AyAudioBytes) // {e = e - x * W m_ayEchoFreeSignal = new byte [ayAudioBytes.length]; For (int i = 0; i & LT; m_ayEchoFreeSignal.length; ++ i) {m_ayEchoFreeSignal [i] = (byte) (m_ayEchoSignal [i] - m_ayTransposeOfSpeakerSignal [i] * m_adWeights [i]); } Return m_ayEchoFreeSignal; }