(NOV 2006) — The November release of the LumenVox Speech Engine has a new feature designed to help developers improve recognition when users are speaking digits.
We have built a new acoustic model — the set of audio that our Engine uses to understand speech — that is designed to only recognize digits. By training our models on audio data made up only of speakers saying digits, we have a new model just for digit recognition that works like an individual language model.
Using the digits–only model results in an accuracy gain of about three points when recognizing digits.
If you use the built–in digits grammar, you will automatically get the benefit of the new acoustic model.
You can also specify the model within your own SRGS grammars, by declaring the language as a digits language. You do this by appending –di to the end of the grammar name.
For the normal American English model, you would specify the language as en–US in your grammar. To use the English digits model, you would specify the language as en–US–di.
For example, the start of an ABNF digits–only grammar might look like this:
If you would like information on downloading the latest release of the LumenVox Speech Engine, please contact us. It is a free download for users with current software maintenance packages.
In addition to the digit models, LumenVox now has beta acoustic models for Spanish and Canadian French.
If you would like to participate in beta testing either of those languages, please contact us for more information.
We are also interested in acquiring more language data for those models. If you have high volumes of callers speaking those languages, contact us about potential partnerships in collecting and using that data.
We have begun adding Frequently Asked Questions pages to our Web site in an attempt to better address your technical needs.
If you have a question that's not answered in those pages, please email technical support.
© 2016 LumenVox, LLC. All rights reserved.