Multimedia Technology IV – Farag, Yang & Jiao (Eds)
© 2015 Taylor & Francis Group, London, ISBN: 978-1-138-02794-7
An adaptive multi-scale lattice vector quantization and
its application in low bit rate speech coding
Weijun He & Qianhua He
School of Economics and Information, South China University of Technology, Guangzhou, China
ABSTRACT: The codebook design is a critical component of low bit rate speech coding. This paper proposed
an improved Pyramid Lattice Vector Quantization (PLVQ) based on Adaptive Multi-Scale (AMS), abbreviated
as AMS-PLVQ. The AMS-PLVQ employs a novel lattice structure of Global Nonuniform and Local Uniform
(GNLU) aiming at improving the PLVQ performance. Based GNLU lattice structure, discriminating VQ scheme
is introduced