Voci Technologies Inc, formerly dba Silicon Vox Corporation is building out on research done at Carnegie Mellon University, to deliver high-performance, low cost speech-to-text solutions for enterprise applications. With a focus n enabling users to extract actionable intellgence from voice, data, the company develops hardware-accelerated speech recognition technology, providing speech appliances based on the proprietary HyperVox technology. HyperVox is a speech recognition-specific architecture that dedicates essential hardware and software resources to computationally intensive aspects of speech recognition. HyperVox streamlines the processing of audio files by parallelizing and pipelining the conversion of speech to text, much like Graphics Processing Units (GPUs) do for graphics. The technology revolves around a custom high-speed memory interface that pulls in data quickly, balancing functional units with memory bandwidth. HyperVox is designed to be extremely flexible to support speech applications of any size, and is able to accelerate recognition across any domain and configuration. Voci has pioneered the implementation of speech-to-text algorithms in hardware to achieve fast, comprehensive speech recognition solutions.
