Masanobu Suzuki

Improved ADPCM voice signal transmission employing click-noise detection scheme for TDMA-TDD personal communication systems

In personal communication systems, better voice quality and lower terminal price are essential requirements, and adaptive differential pulse-code modulation (ADPCM) codecs are one of the voice codec candidates that have good voice quality and reasonable cost. However, in slow fading environments, burst errors induce click noise in ADPCM voice signal transmission. This paper proposes a novel error-concealment scheme to improve the ADPCM voice signal transmission quality of time division multiple access time division duplex (TDMA-TDD) personal communication systems in slow fading environments. The proposed scheme employs an improved error-detection technique with PCM differential detection and overflow detection coupled with a cyclic redundancy check (CRC) code. To compensate for residual click noise, an expanded CRC error-detection signal is used. Erroneous PCM signals are adaptively suppressed according to the click-noise detection information. Hardware simulation results show that these techniques significantly improve ADPCM voice signal transmission quality by 0.5 points and 1.0 compared to that of the conventional ADPCM muting scheme and no processing, respectively, with the mean opinion scoring (MOS) test in the slow fading environments typical of personal communication systems.

A very low power consumption ADPCM voice codec LSIC for personal communication systems

This paper proposes a very low power consumption ADPCM voice codec LSIC. The developed LSIC satisfies CCITT recommendation G.721 completely and is suitable for realizing the high-quality voice signal transmission expected from personal communication systems in fading environments. By employing a time sharing signal processing, a very low power consumption (only 500 /spl mu/W) for ADPCM codec core has been achieved. Moreover, the LSIC has a novel click noise detection function and erroneous received signals are suppressed digitally according to the click noise detection information. This makes it possible to improve the voice transmission quality significantly in fading environments, for instance, the MOS (mean opinion score) evaluation is improved by 0.5 points compared with that of the conventional scheme.

A voice transmission quality improvement scheme for personal communication systems-super mute scheme

To improve the ADPCM voice quality in slow fading channels, this paper proposes the super-mute (S-Mute) scheme by using the probability distribution of ADPCM code without error bits and the error patterns. With consideration of characteristics of voice and channel error patterns, it makes it possible to reduce the code error significantly compared with conventional muting, which has only two replacement code patterns. Furthermore, this paper shows the adaptive replacement of ADPCM code against the grade of burst degradation by applying the channel quality estimation scheme, the ADPCM code detection (ACD) scheme. As a result, when the proposed scheme is employed, the mean opinion score (MOS) improves by 0.6 points compared to the conventional muting scheme.

Double side information concatenated coding scheme for fading environments-experimental evaluation employing SNUFEC VLSI

In order to realize a high-coding-gain forward error correction scheme for fading environments, a novel concatenated coding scheme employing soft decision decoding with channel-status-information (CSI) for inner codes and also soft decision decoding for outer codes (double side information concatenated coding scheme) is proposed and its performance is clarified by a hardware experiment employing the developed ultra-high-speed (operation speed up to 60 MHz) and universal-coding-rate (from coding rates one-half to fifteen-sixteenths) Viterbi decoder VLSIC (SNUFEC VLSI). The proposed scheme improves the Pe performance to 1/3 of conventional one by employing the SNUFEC VLSI which can output the path memory circuit status 1,0 ratio as the soft-decision decoding data for likelihood information.<<ETX>>