Digital Speech Coder

 

High-Compression Speech Coder

HCSC is useful for long-duration or small-storage digital speech recording and playback such as for handheld and portable audio/visual devices. It is also useful for bandwidth-efficient wireless voice communication.

HCSC achieves high compression ratios by employing very compact models for human speech production, judicious glottal excitation representation, and optimal trellis vector quantization.

HCSC is available for 4- and 5-kHz signal bandwidth, with corresponding sample rate at 8 and 11.025 kHz, respectively. HCSC-4 achieves 64:1 compression ratio (from the original speech data rate at 128 kbit/s to the compressed rate at 2.0 kbit/s), while HCSC-5 achieves 80:1 compression ratio (from the original speech data rate at 176.4 kbit/s to the compressed rate at 2.21 kbit/s).

HCSC-5 offers wider-bandwidth voice for applications which require significantly enhanced speech intelligibility such as electronic dictionary.

HCSC/VX: Combined Voice Compression and Voice Changing - HCSC/VX integrates two useful functions into HCSC: Voice changing (among male, female, children and robot voices) and speaking speed control. HCSC/VX is useful for voice chat for on-line games, language learning devices, e-dictionary, toys, children books, text-to-speech, and many other applications. Importantly, the two added functions require negligible increase in DSP resources (MIPS, RAM and ROM).

FEATURES

  • User-selectable multiple data rates:

    - HCSC-4: 0.8/0.9/1.0/1.2/1.4/1.6/1.8/2.0/2.4/2.8 kbit/s; with 2.0 kbit/s as normal operating data rate.

    - HCSC-5: 1.66/2.21/3.3 kbit/s; with 2.21 kbit/s as normal operating data rate.

  • Robust speech quality even with acoustic background noise.
  • I/O format: 16 bit/sample linear PCM. 8 kHz (HCSC-4) and 11.025 kHz (HCSC-5) sample rate.

 

High-Quality Speech Coder

HQSC, a medium-data-rate coder, is useful for high-quality digital recording and playback of speech and for bandwidth-efficient wireless voice communication.

HQSC achieves efficient speech compression by employing advanced speech production modeling (including formant, pitch and glottal excitation), optimal trellis vector quantization, and psychoacoutic perceptual weighting for speech analysis.

HQSC is available in three signal bandwidths: 4, 5,and 7-kHz, with corresponding sample rate at 8, 11.025, and 16 kHz, respectively. HQSC offers multiple user-selectable data rates. By "embedded coding", the lower-rate coding is easily realized by dropping bits from the higher-rate coding.

FEATURES

  • User-selectable multiple data rates:

    - HQSC-4: 3.6/4.8/6.0/7.2/8.4/9.6/10.8/12/13.2/14.4 kbit/s.

    - HQSC-5: 6.5/9.1/11.8/14.5/17.1/19.8 kbit/s.

    - HQSC-7: 7.2/9.6/12.0/14.4/16.8/19.2 kbit/s.

  • Robust speech/audio quality even with acoustic background noise.
  • I/O format: 16 bit/sample linear PCM.