Research On Distortion And Recovery Of Acoustic Signal In Enclosed Space

The phenomenon of acoustic multipath is inevitable in the sound propagation in bounded space,because the reflection,scattering and diffraction of sound will occur in the sound propagation process due to the existence of the boundary.There is an urgent need to solve the problem of acoustic signal distortion in room,car and underwater acoustic shallow sea communication.The existing processing methods think that the reverb and the direct sound can be distinguished in the multi-sound process,but in fact the reverb is homologous coherent interference and non-additive cancellation.Moreover,the channel is extremely complex and unstable,and it is difficult to obtain the channel information.Therefore,the signal recovery problem is considered by mathematical physics instead of processing from the perspective of constrained channel processing.The signal recovery algorithm proposed in this paper provides a new idea and method for correcting signal distortion caused by aliasing,and also provides a solution for inverse correlation of one-dimensional solutions,which is of great significance for realizing the recovery of sound source signals.In this paper,the mechanism of acoustic signal distortion in two typical bounded Spaces is first studied,that is,the acoustic multipath effect in rectangular room and shallow sea waveguide environment.The transient acoustic field is simulated by coherent virtual source method(CISM)and COMSOL simple normal wave theory respectively.It is found that the autocorrelation function has a strong correspondence with the waveform,that is,the autocorrelation function can also reflect the time-domain characteristics of the signal to a certain extent.According to the law of conservation of energy and the Wiener-Khinchin theorem,the sound energy is confined within the finite space volume,the energy radiated by the sound source is equal to the energy absorbed in the whole space,and the autocorrelation function and power spectrum are a Fourier transform pair.Two correction formulas of autocorrelation function for transient acoustic signals in closed space and shallow sea waveguide environment are derived.The non-normalized autocorrelation function can be obtained at 1m distance from the sound source equivalent sound center.In order to solve the problem of inverse autocorrelation function,the known constraints are iteratively applied in time domain and frequency domain,and finally converge to the original signal.The accuracy and applicability of the algorithm are verified by using the simulated free field signal as a reference.Then,through numerical simulation and experiment,the transient signal recovery in the reverberation chamber,the reverberation pool and the channel pool is realized under three bounded sound fields.The correlation coefficients after recovery are all above 0.9,and the convergence rate is fast,which proves that the method can effectively realize the recovery of the transient acoustic signal in bounded space.Finally,the restriction conditions of signal in known segments are discussed.In this paper,the physical causes of signal waveform distortion are analyzed from the physical mechanism of sound propagation,and the autocorrelation function,a deterministic observation quantity,is found by using the law of conservation of energy,and a new technical way of signal waveform recovery is found.