摘要
在低比特率语音编码方法中,改进的多带激励编码方法可得到低噪声和稳健性强的重建语音,但运算量很大.作者在改进的多带激励编码研究中,将矢量量化方法引入频谱幅度的量化,不仅降低了运算量而且降低了编码比特率,使得在1.35~1.8kbit/s的编码比特率下的重建语音,与4.8kbit/s编码比特率下(按自激励声码器编码方法)重建的语音质量相比毫不逊色.
Among the low bit rate speecd coding methods, IMBE coding is a type characterized by low noise and high robustness, yet it is also the most complicated in computation. Based on vector quantization introduced into IMBE coding for spectrum amplitudes, the authors developed a coding strategy more efficient not only in bits spent but also in computation. With the VQ-IMBE coding, a bit rate as low as 1 350~1 850bit/s can be achieved without losing reconstructed speech quality in comparison with that obtainable with the self exited vocoders.
出处
《北京理工大学学报》
CSCD
1995年第2期205-211,共7页
Transactions of Beijing Institute of Technology
基金
国家自然科学基金
关键词
语音编码
低比特率
多带激励
矢量量化
声码器
voice coding/low bit rate
variable
multi-band excitation
vector quantization