摘要
在背景噪声存在的情况下,电子耳蜗(cochlear implant,CI)的性能会受到严重影响,而传统谱减语音增强算法不能有效去除电子耳蜗中的各种非平稳噪声。为了进一步提高电子耳蜗使用者的听觉感知能力,本文提出了一种改进的谱减算法,其将带噪语音信号的功率谱进行Bark分带,在有声帧的每个子带内按照最小统计量控制递归平均方法进行噪声估计,并自适应地调节谱减参数。将该算法应用到电子耳蜗的前端预处理并仿真电子耳蜗合成声音。结果表明:相比传统的谱减法,该算法能够进一步提高合成语音的清晰度和可懂度。
The performance of the cochlear implant（CI） deteriorates rapidly in the presence of background noise,and the conventional spectral subtraction speech enhancement method used in the CI can not remove the non-stationary noise well. In order to improve the audio perception ability of the CI users,an improved spectral subtraction method is proposed,which divides the power spectrum of the noisy speech according to the Bark bands and estimates the noise by using the minima controlled recursive averaging method in every individual band when the speech frame is present.Then the adaptive spectral subtraction parameters are used in the algorithm.The proposed spectral subtraction method has been applied in CI,and the simulation result shows that the proposed algorithm may extract the speech features more correctly and the CI synthesized sound is of higher quality.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2010年第1期188-192,共5页
Journal of Biomedical Engineering
关键词
电子耳蜗
谱减法
噪声估计
Bark子带
Cochlear implant（CI）
Spectral subtraction
Noise-estimation
Bark bands
