Automatic Speech Recognition (ASR)
LumenVox ASR Features
We have spent over a decade advancing speech recognition capabilities through the development of our Automatic Speech Recognition engine. These ongoing efforts have yielded a number of features that will help you in building world-class speech applications while maximizing return on your investment.
Distributed Client / Server Architecture - Seamlessly Grow Your Speech Environment
You can’t afford service outages or hardware failures. The versatility of the LumenVox client/server architecture allows your administrators to seamlessly grow speech environments as the business functions they support grow. This distributed architecture provides stability through redundant installations and achieves higher levels of performance through load balancing, without requiring increased processor load.
Flexible Licensing - Choose the Right Pricing Model for Your Project
With a selection of licensing options to choose from, it is easier than ever to deploy LumenVox’ suite of products. Options include per–port, monthly subscription, use-based, bursting, and Software as a Service (SaaS). In most cases, the LumenVox software will be deployed locally to your application, right where you want it for optimal results.
Let industry standards simplify development. Our adoption of standardized protocols and development tools make it easy to replace your existing ASR with the LumenVox technology.
supports MRCP v1 and 2, we are VXML compliant
can be written in GrXML or ABNF, with SISR and SRGS capabilities
Server-Side Speech Grammars - Efficient Support For Large Speech Grammars
LumenVox’ unique technology offers even more efficient support for large speech recognition grammars through advanced server and client-side grammar caching. Server–side speech grammars allow large grammars to be compiled once and saved so that they can be loaded again very quickly and without using up valuable processor or memory resources.
Voice Activity Detection (VAD) Accurately Separates Speech From Background Noise
Whether the call is coming from a crowded restaurant or inside of a speeding car, the LumenVox ASR engine separates speech from background noise using Voice Activity Detection (VAD). Human speech has qualities that make it distinguishable from other sounds, including energy level (volume), frequency (pitch), changes in frequency and duration. VAD listens for these qualities within the incoming audio to accurately detect the actual speech.
NBest technology improves user interaction and, with the LumenVox ASR engine, is particularly effective when callers need to spell names, street addresses or email addresses. Using NBest reduces the need for a caller to continuously repeat a letter over and over and will ultimately eliminate the need.
Instead of returning only the top scoring result, the engine returns several of the highest scoring, most likely answers.
Using these NBest results, the system can prompt the caller with, for example, “Did you mean ‘M,’ as in ‘Mary’?” When the caller responds, “No,” the system can go to its next option, “Perhaps you meant ‘N,’ as in ‘Nancy?'”