Speech Engine v8.5

Speech Engine v8.5 Release Boasts New
Pronunciation Generator and Grammar Caching

Our latest Speech Engine is here, and provides a number of benefits to our customers.

The new pronunciation generator uses statistical pronunciation rules instead of hard-coded language rules. When a word is in a grammar and its pronunciation is not in our internal dictionaries, we use the pronunciation generator to guess how it sounds. Previously, we had to hardcode pronunciation rules for a language. Now, we train the pronunciation generator in a way similar to how we train the Speech Engine, so that it can better handle words in a language that don't follow normal pronunciation. A good example is a foreign name: in English the "j" is pretty much always pronounced as in "Jacob." But if you have a name like "Juan" the "j" is pronounced completely differently. Because our training set includes these kinds of foreign names, the new pronunciation generator is better able to understand the alternative ways a "j" might be pronounced.

In addition, we added grammar caching that significantly improves speed when working with large grammars. When a large SRGS grammar is loaded, the Engine has to compile that grammar into a binary format. This process can take tens of seconds or even a few minutes for very large and complex grammars. We now automatically save that binary file and cache it for future use. If somebody tries to load that same grammar file, we check to see if it has changed. If it has not, we do not spend the time to compile it all over again — we just use our cached copy, meaning we save all the compilation time.

Lastly, we have updated licensing which allows for distributed licensing servers and moving licenses around to different servers.

Broadsoft Compliance

Broadsoft Logo

LumenVox, an innovator of speech recognition software, today announced that it has achieved interoperability validation of its Speech Engine on the BroadSoft BroadWorks® VoIP application platform. BroadSoft's application software enables service providers to deliver hosted PBX, mobile PBX and SIP and IP business trunking services.

BroadSoft's service provider customers can now use speech recognition technology in conjunction with BroadWorks for enhanced customer service and better business efficiency.

"LumenVox's speech software is an example of how an innovative new technology can enhance what operators have to offer to their customers with BroadWorks," said Sandro Cianci, director of Business Development for BroadSoft. "Partners like LumenVox boost the value of BroadWorks for operators and end-users alike." More

Employee Spotlight

employee photo Matt Warshal joined LumenVox in the beginning of this year, and has already helped many clients in his role in the Technical Sales and Support department.

Originally from balmy Syracuse, NY, Matt is integral in assisting customers through integration, development, testing and tuning of speech applications. Matt, and the rest of the Support team can be reached at support@lumenvox.com.

On-Demand Free Videos

These on-demand, free training videos provide you with an arsenal of knowledge.

Introduction to Speech Tuning

Phonetic Spellings

Speech Recognition In Under an Hour

Our on-demand, free training videos provide you with an arsenal of knowledge about speech recognition application development, and more.

Active Voice Chooses LumenVox

Active Voice logo

Active Voice, LLC, a global leader in Unified Messaging, Unified Communications and speech-activated solutions, and LumenVox, an innovator of speech recognition software, today announced the availability of Active Voice SPEAK, an advanced speech-enabled auto attendant designed specifically for the small and mid-sized business environment (SMBs) and the mobile workforce.

"Until now, only large enterprises could afford powerful and robust speech-driven solutions. SPEAK changes this paradigm," said Eyal Inbar, General Manager of Marketing and Product Development, Active Voice. "We are pleased to work with LumenVox and Digium to deliver a practical and viable speech-enabled auto attendant that is designed just for the specific needs of SMBs. Lower costs combined with plug-and-play convenience help make SPEAK the ideal solution, allowing SMBs to see beneficial results of their investment very quickly."

Active Voice chose LumenVox and Digium because of their proven technologies, added Inbar. "Active Voice SPEAK leverages LumenVox's cutting-edge Speech Engine and Digium's open standard hardware, enabling SPEAK to be offered at a price point that is unbeatable in the market and at a level of sophistication that is yet to be experienced by SMBs. The bottom line of SPEAK is simplicity, affordability and ease of use." More

Hearing is Believing

Solution Showcase

We've launched our new Solution Showcase that highlights actual recordings of customer deployed speech applications!

If you are interested in highlighting your speech application, here is your chance. Each showcased solution contains an audio file, a link to your site, and a brief description of your application.

Please provide a .wav or .mp3 file (approx. 1-5 minutes long) of someone calling the speech portion of one of your applications, along with your logo and a description of your application.

Email all materials to us.


Come to the Big Apple for SpeechTEK!

Aug. 18-20, in New York City, we are featuring five LumenVox Partners that will be co-exhibiting on our booth.

Aculab offers a wide range of hardware and software building blocks for integration into high performance, wired and wireless communications solutions - from contact centres and IVR to prepaid services.

Cepstral provides speech technologies (Text-to-Speech) and services for the spoken delivery of information.

Inventive Labs
Inventive Labs announces Voice Elements, a .NET toolkit for building speech-driven telephony solutions, such as IVRs.

Prosodie Interactive
Prosodie is a leading technology services provider with more than 20 years of hands-on experience in delivering cost-effective IVR solutions.

VClarity VXML Voice Platform is a high performance, carrier-grade telephony server that is responsible for call control and speech recognition capabilities.


The annual Asterisk trade show is right around the corner, Sept. 22-25, in Glendale, Arizona --and we have only one spot left for our partner kiosks in our booth.

For a rate much cheaper than buying your own booth, you can get a partner kiosk in our booth that offers great visibility at these shows.

Basically, all you will need to do is bring your collateral and a representative from your company -- no lengthy booth setup or teardown, no expensive shipping costs.

© 2018 LumenVox, LLC. All rights reserved.