| Acoustic feedback is the most common problem that faced by the modern amplification system.Acoustic feedback is based on the self-oscillation of soundreinforcement system.Acoustic feedback can cause fuzzy sound in the amplification system.Low extent of acoustic feedback can reduce the system’s audio quality,while severe acoustic feedback can not only cover the normal sound release,but also damage the amplifiers.Some can even cause different levels of damage to people’s auditory system.Acoustic feedback inhibition is an aspect not to be neglected in the use of amplification system.Traditional acoustic inhibition methods will cause severe damages to the sound quality,or have difficulties to operate,so its actual effect is not ideal.Modern acoustic inhibition devices have now already overcome the problems faced by the traditional ones.They are faster and more effective.However,most of these devices are produced in foreign countries,and so are too expensive to gain a wide use.Therefore,the study of acoustic inhibition has high scientific research as well as markets values.This thesis,based on the mathematic theory of self-oscillation,analyzes the actual measured acoustic feedback audio signals,and concludes on the basic laws of acoustic feedback.Since the wavelet transform has the ability to reflect signal mutations better than the Fourier transform,so this thesis introduces a reversal superimposition measure based on wavelet packet transform to inhibit acoustic feedback.First is to divide the sampled audio signals into wavelet packet frequency sub-bands through wavelet packet decomposition,and then calculate the energy of each sub-band.Through the energy threshold,we can recognize those sub-bands having acoustic feedback,and then,carry out reconstruction processes in order to achieve acoustic feedback inhibition.In accordance with the band division of psychological acoustics,plus the masking effect of human ears,wavelet packet frequency sub-bands division will not bring influences to the sound quality after eliminating the acoustic feedback.Through the simulation analysis of the actual measured audio signals,this thesis proves that the reversal superimposition is accurate and feasible.According to the reversal superimposition measure,this thesis designs a DSP hardware platform with TI’s TMS320C6713 b chip as its core and the TLV320AIC23 b as its audio codec chip,and then build peripheral circuits based on the core chip.At the software system aspect,this thesis divide the system task’s priority,and gives the frame of system’s main function and interrupt function.It also thoroughly presents the implementation of wavelet transform algorithm on TMS320C6713 b,and gives the programming method of wavelet packet algorithm.Finally,this thesis achieves the goal of acoustic feedback inhibition based on DSP. |
