Research and implementation on objective speech quality evaluation
Jianghao Lin · Journal of Guangzhou University · 2011
Objective speech quality evaluation plays a pivotal role in SALL(Self-Access Language Learning).This thesis firstly introduces the pre-treatment unit of voice signals in the process of objective speech quality evaluation.Then,it also presents diverse voice features extraction algorithms and makes comparisons between them.After that,the MFCC feature,which is closer to human auditory model,is chosen in the extraction process.Meanwhile,the whole process and its results are provided.Finally,we compare the advantages and disadvantages of current evaluation models(DTW and HMM),and propose an evaluation method which adopts the HMM model,and thus design a system to test and verify the objectivity of evaluation using this method.