Study On The Evaluation Method Of Guqin Timbre Based On The Spectrum Quantization Model

"Qin" as "qin,chess,book,painting"the first four art.It is regarded as the art of philosophy.And guqin because of its clear,and,light,elegant music character,is beloved by the literati.Nowadays,with the death of literati and the change of people's aesthetic conceptions,guqin,with three thousand years of cultural history,is facing a crisis of inheritance.It is necessary and urgent to save the national excellent culture and inherit the guqin culture.Because of the uncertainty of the measurement of timbre of human ear auditory perception,it is of low credibility to judge a guqin's timbre with large randomness and unstable result.How to use modern science and technology to judge the timbre of a musical instrument in a rigorous and scientific way is of great practical significance to improve the scientific and authoritative evaluation effect.Guqin tone evaluation is a comprehensive analysis of the guqin music signal to identify the timbre conditions of guqin and evaluate its value.Guqin timbre of objective evaluation can be achieved by scientific and objective way based on the analysis of the sound of the voice quality features and value of guqin,can be widely used in the production of guqin guidance,business evaluation,etc.This paper proposes a guqin spectrum quantitative scheme for voice quality evaluation,the main idea is based on standard guqin music tone or a spectrum of acoustic signals are extracted,the spectral feature space based on clustering technology to quantify the spectrum feature modeling,testing will extract the guqin acoustic spectrum signal and standard of quantitative model matching tests is given according to the results of the matching of guqin voice quality score.Same tone or the same piece of voice signal extracting meyer to spectrum parameters(MFCC)as acoustic spectrum characteristic parameter,the K-Means algorithm for characteristic vector clustering,the spectrum quantitative formed by LBG algorithm spectrum quantitative yards in this model,each mono or music eventually is expressed as a set of quantitative code book.The matching quantization of the model is based on the minimum average code distance.The experimental results show that the proposed method is effective.