Combined speech and channel coding for mobile radio communications

摘要

We present a combined speech and channel coding scheme for digitalnmobile communications. The speech coding algorithm is based onncode-excited linear prediction (CELP) and it provides good communicationnquality speech at a rate of 4 kbps. For the channel code, bothnrate-compatible punctured convolutional (RCPC) codes and puncturednReed-Solomon (PRS) codes are considered. In the case of RCPC codes, softndecision decoding is considered, in addition to the simpler hardndecision decoding. The modulation format chosen in our study isnΠ/4-DQPSK with differential detection. Unequal error protection isnused based on the bit error sensitivities of the different speechnparameters. The performance of the combined speech and channel codec isnstudied under different mobile channel conditions, such as fade rates,nsignal-to-noise ratios, and interleaving delays. The results indicatenthat, with no interleaving delay and a large channel signal-to-noisenratio, PRS codes and RCPC codes with soft decision decoding have similarnperformance, given in terms of the segmental signal-to-noise ration(SSNR) of the reconstructed speech. At low channel SNR, RCPC codes withnsoft decision decoding are significantly better than PRS codes. Informalnlistening tests were also conducted and the results indicate that thendifference between the quality of the reconstructed speech in annerror-free channel and that in a Rayleigh fading channel isnimperceptible at channel SNRs larger than 20 dB for both the RCPC andnthe PRS-based codecs