BP神经网络在评价歌唱艺术嗓音中的应用

录制湖南科技学院30名无喉病、无上呼吸道感染的声乐专业青年大学生专业训练歌声信号，利用语音分析技术提取歌声声学参数第一共振峰、第三共振峰、基频、音域、基频微扰、第一共振峰微扰、第三共振峰微扰、平均能量，使用BP神经网络方法客观评价歌声质量，并与资深声乐专业教师的主观评价进行比较，误差在3．4％之内。结果表明BP神经网络方法利用评价参数能正确客观评价歌声质量，有助于科学地指导选拔和训练艺术嗓音人才。

Application of BP neural network in evaluation of artistic voice

The singing voices were recorded from 30 young music students who come from Hunan University of Science and Engineering.Their acoustic parameters,such as F1,F3,F0,vocal range,jitter,disturbance of F1,disturbance of F3 and average energy were extracted by the way of voice analysis.BP Neural network analysis was used to evaluate the singing voices objectively.The results were then compared with those of the subjective evaluation performed by the experienced professionals.The error between the two evaluation approachs was within 3.4%.The results show that the neural network analysis can be used as an objective instrument to evaluate the singing quality of artistic voices.This is helpful to instruct,select and train professional singers.