SUBBAND SPEECH CODING AND MATCHED CONVOLUTIONAL CHANNEL CODING FOR MOBILE RADIO CHANNELS

Due to increased radio spectral congestion, the trend in future cellular mobile radio systems is toward digital transmission. The recent advances in spectrally efficient modulation techniques and high quality low bit rate speech coding have further aided this move. However, mobile radio channels are subject to signal fading and interference which causes significant transmission errors. The design of speech and channel coding for this application is therefore challenging. In this paper, the effects of digital transmission errors on a family of variable-rate embedded subband speech coders (SBC) have been analyzed in detail. It is shown that there is a difference in error sensitivity of four orders of magnitude between the most and the least sensitive bits of the speech coder. As a result, a family of rate-compatible punctured convolutional (RCPC) codes with flexible unequal error protection capabilities have been matched to the speech coder. These codes are optimally decoded with the Viterbi algorithm. On a Rayleigh fading channel with differential four phase shift keyed modulation, more than 5 dB gain in channel signal-to-noise ratio can be obtained by using 4 levels of unequal error protection over conventional designs that utilize only 2 levels. This gain is achieved over a large range of channel signal-to-noise ratios, at no extra bandwidth requirement and only a small complexity increase. Among the results, analysis and informal listening tests show that with a 4-level unequal error protection scheme, transmission of 12 kb/s speech is possible with very little degradation in quality over a 16 kb/s channel with an average bit error rate of 2 . 10(-2) at a vehicle speed of 60 mph and with interleaving over two 16 ms speech frames. The SBC speech encoder/decoder and the RCPC channel coder/decoder have been implemented on a single AT&T DSP-32 floating point signal processor. The overall end-to-end delay is about 88 ms.