MetaVoice
MetaVoice is a patent pending system that creates an accurate mathematical model of human speech.
Unlike conventional voice compression methods which condense the audio waveform to reduce bandwidth, MetaVoice models every element of the human voice including resonance, pitch, timbre, timing, and character. The result is the most accurate reproduction of the human voice without distortion, while consuming the lowest possible bandwidth, storage and processing power.
- Incorporate high quality voice to bandwidth-constrained environments
- Minimize packet loss, network delays, out-of-sequence information
- Achieve compression ratios as high as 53:1
- Available on multiple platforms
- Fixed and floating point implementations available for both DSP chip ports and embedded devices
The low bandwidth requirement of this technology means that multiple voice channels can run in parallel, making multi-point conferencing, chat, and gaming a reality for any developer. You'll also be able to speed up or slow down voice playback by up to 200% without distorting speech or changing its pitch.
MetaVoice Codecs
RT24: A MetaVoice 2400 bits/second, real-time voice codec designed for personal computing applications requiring minimal processing power.
RT29HQ: A MetaVoice 2977 bits/second codec containing an enhanced human speech model.
MetaVoice is designed to ignore ambient noise (such as traffic in the background), delivers cleaner sound and reduces the amount of data analyzed by the compressor. MetaVoice also takes advantage of a variable-bit-rate coding system which processes only as much information as is needed to reproduce the speech signal. Fixed-rate systems process the same amount of data at all times, even during silent periods. (Since not all applications support variable-rate coding, MetaVoice can also operate at a fixed rate.)
All speech compression algorithms represent a technical tradeoff between quality, bit rate, processing power and delay. Voxware's MetaVoice coding technology dramatically alters the tradeoff between quality and compression, allowing excellent speech quality in a fraction of the bandwidth.
For an example of Voxware's MetaVoice compression, check out our Compression Analogy.
Voxware and VoiceFont are registered trademarks, and ToolVox, MetaVoice, MetaSound, and VoxChat are trademarks of Voxware, Inc.
All contents Copyright 1996 Voxware, Inc., All rights reserved.
Voxware, Inc.
305 College Road East Princeton, New Jersey 08540
voice: (609) 514-4100 fax: (609) 514-4101
