Design And Implementation Of Echo Cancellation Based On The Android Platform

With common promotion conference calls and hands-free phones, voice qualityof communication has always been the hot spot of the study, while the existence ofthe echo signal is the main reason affecting the quality of communication. Therefore,it is very urgent to echo cancellation of practical application, in the field of acousticecho cancellation, the core of the adaptive filter is based on the available data updateand iterate the weights, and track the actual path of the echo.This paper studies the adaptive echo cancellation algorithm, from the aspects ofconvergence speed, algorithmic complexity and disorder to compare the performanceof the algorithm, considering various factors under the fusion NLMS, SIPNLMSPNLMS algorithm advantages put forward the improved algorithm. Double sidedetection algorithm is an important part of the echo canceller, analysis theperformance of each detection algorithm and its practical application, theautocorrelation detection algorithm improved algorithm is proposed on the basis ofMECC. Pretreatment module of signal preprocessing, on both ends for adaptive filterto provide a good sample; Nonlinear processing module is the residual error forsubsequent processing. Combined with the four modules designed a complete echocancellation, and matlab simulation. Through the experimental results to theconvergence performance of the filter and signal-to-noise ratio on the double enddetection is analyzed, finally puts forward solutions to problems, for echo cancellationalgorithm implementation on android system.Then introduce echo cancellation of hardware and software resources. Firstintroduce the key technology of android talkback system, system analysis and designof the communication process and the software architecture (Java layer, JNI layer andc code). We used the eclipse development environment, JNI for the development oftechnology and the NDK intercom system. The software and hardware environmentecho cancellation for transplants.The last is the transplanting of the echo cancellation algorithm. First echocancellation algorithm written in c code, modify the software architecture, can makethe Java and c layer for data transmission. Resolved through repeated test and modifythe code, Java and c layer data does not match the system flash back, receives the dataout of sync with the recording thread problem. Finally add voice sample function, canbe saved by saving the microphone signal, remote signal and a comparative study of residual signal echo cancellation, not only in hearing is further verify the effect ofecho cancellation in the actual data. The realization of echo cancellation schemeunder the Android system to other systems of echo cancellation has guiding function.