Speech Coding Software DualSpeech

Hitoshi Ohmuro, Kazunori Mano

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Low-cost highly-efficient speech coding is important for personal multimedia communications. In particular, the demand for real-time speech compression on a personal computer has been increasing. In this paper, we describe the algorithm of the DualSpeech codec developed by NTT and outline the software product. DualSpeech achieves real-time high-quality speech compression and decompression without any extra signal processors. The bit rate is 7.4 kbit/s (high-quality mode) or 4.6 kbit/s (low-bit-rate mode). The quality of the decoded speech is equivalent to that of ITU-T G.726 (ADPCM) or a mobile telephone (PSI-CELP). To achieve low complexity and high quality, we applied new speech modeling and high-speed code searching. DualSpeech software is now available not only for general computers but also for cheap RISC processors.

Original languageEnglish
Pages (from-to)549-554
Number of pages6
JournalNTT R and D
Volume47
Issue number5
Publication statusPublished - 1998 Dec 1

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Speech Coding Software DualSpeech'. Together they form a unique fingerprint.

  • Cite this

    Ohmuro, H., & Mano, K. (1998). Speech Coding Software DualSpeech. NTT R and D, 47(5), 549-554.