Research On Acoustic Approximation Of Piano Based On Microphone Array

With the improvement of material life,people's demand for spiritual life has increased,As a king of musical instruments,the piano has gradually become popular with people.The piano needs to be able to use left and right hands flexibly,and it also meets the needs of the middleaged and the elderly.However,the piano's large body and high price make it difficult for many families.Although the elecpiano is small and inexpensive,it is difficult to produce a sound field that sound comparable to a piano in performance.This topic focuses on the difference between the piano and elecpiano in the sound field,and proposes a piano timbre evaluation system.Based on this system,the method of letting the elecpiano approach the piano level in the sound field effect is studied.The sound field approximation of the piano includes two parts.The first is the extraction of the time domain and cepstrum domain features of the existing piano and elecpiano.The piano timbre evaluation system is generated through the neural network model.The second is the extension of the elecpiano sound field through the speaker,and the generated evaluation is used.The system evaluates the current sound field.This method can let the ordinary family can also play the sound field sensory effect of the piano only when there is an elecpiano.The main research work and innovations in this article are as follows:1.A timbre evaluation system based on feature extraction and machine learning is proposed.18 microphones are used to fully acquire the sound field information;a microphone platform is controlled by a hardware platform independently developed by the laboratory;Write software to reconstruct and denoise a collection of audio signals;a correlation coefficient matrix is generated for a group of collected audio signals;and matrix eigenvalues are further extracted as time domain features;The collected audio extracts the Mel cepstrum coefficients,and is average-compressed as the cepstrum domain feature;an evaluation model is generated in conjunction with the BP neural network.2.According to the Haas effect,a speaker pendulum method is proposed which can quickly find a good effect in the sound field within the space.Through this pendulum method,the time for determining the position of the speaker in the sound field approximation experiment can be effectively saved.At the same time,this method also It is suitable for ordinary home use speaker to connect multimedia to get a better audio-visual experience.3.Proposed a elecpiano sound field approximation system that based on timbre evaluation system and Haas effect.The Haas effect method is used to determine the best place of the loudspeakers.In this area,the loudspeakers are slowly adjusted,and the microphone array is used to collect the sound field information of each place at a time,and the collected audio is extracted in the time domain and the cepstrum domain and transmitted.In the evaluation system,the evaluation of the approximation effect of the sound field in the current position was obtained,and through constant adjustment,the sound field sense of the piano could be achieved for both models.