An 800 bps VQ‐based LPC voice coder

By Xianglin Wang, C.-C. Jay Kuo. 1998

In this paper, it is investigated how to efficiently apply VQ on LPC so as to realize a very low bitrate speech coder with reasonable speech coding quality. The coding bitrate has been pushed to 800 bps and the result is compared with that of a traditional LPC 2.4 kbps speech coder. In LPC10 standard(FS1015), scalar quantization is adopted to quantize the LPC parameters. However, when vector quantization is considered, the speech coding rate can be further reduced without obvious degradation of speech quality. In our coding scheme, instead of direct encoding of LPC parameters, LSF coefficients are encoded which are obtained from a transformation of LPC. In addition, in order to utilize the correlation among LPC parameters, three frames of speech are processed together so that some kind of prediction could be utilized. A weighted version of MSE is selected as the distortion measure in building the VQ codebooks and encoding. Also, due to their different characteristics, separate codebooks are built and used...

Subfields: Speech & Audio

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Speech & Audio

Citations in the corpus

Unsupervised Speech Representation Learning Using WaveNet Autoencoders (2019, background): “largest inventory of tokens, the overall encoder bitrate is low: 14 bits at 50 Hz = 700 bps, which is on par with the lowest bitrate of classical speech codecs [63].”