Improvements in the speech quality for a DECT system

The 32 kbits/s ADPCM media DECT system is a very simple algorithm that produces a good subjective speech quality when the transmission channel conditions are ideal. When the errors produced in the radio channel appear in bursts the speech quality is degraded rapidly, and the algorithm can not recover the synchronisation in a reasonable time (250 ms of resynchronization time when the channel produces a 2 ms burst error). This is an undesirable situation that generates a very unpleasant noise. This paper presents speech quality studies that proves an important improvement when the speech codec is switched from the ADPCM 32 kbts/s to ADPCM 24 kbits/s including a channel coder for the remaining bits, together with the results obtained using the G723 24 kbit s/s ADPCM joint channel coding where the subjective speech quality has been improved from 3 to 4 MOS (Mean Opinion Score) when the signal to noise ratio is equal to 15 dB and the cochannel interference is equal to 10 dB in a Gaussian channel. The speech quality improves from 2 to 3 when the C/I is 10 dB and the SNR is equal to 45 dB in a Rayleigh channel with the incorporation of a shortened BCH block code. An additional Mute subsystem that handles the PCM samples and is helped with uncorrected bit information provided by the channel decoder subsystem has also been incorporated. Simulations with both subsystems shows important improvements in the performance of the global system.

[1]  Shuji Kubota,et al.  A very low power consumption ADPCM voice codec LSIC for personal communication systems , 1994, 5th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, Wireless Networks - Catching the Mobile Future..

[2]  Robert F. Kubichek,et al.  Output-based objective speech quality , 1994, Proceedings of IEEE Vehicular Technology Conference (VTC).