A Low-Complexity 2.4 kb/s Speech Coding Algorithm Based on Sinusoidal Excitation Linear Prediction
Cui Hui-juan · Natural Science Journal of Xiangtan University · 2010
This paper presents an improved low-complexity 2.4kb / s speech coding algorithm based on sinusoidal excitation linear prediction.LSF parameters reduce the computational complexity by the global optimal scalar quantization adding G.729's inter prediction of LSF quantitative technology,and ensure the voice quality by Direct extraction of Calculating energy parameters enhanced the capability of anti-error. Mode merger in banging saves the bit allocation.The results show that the synthetic speech quality significantly better than 2.4kb / s mixed excitation linear prediction algorithm,and its computational complexity is lower.