Speech technologies provider Loquendo (News - Alert) has announced the latest version of its Speech Recognition offering, ASR 7.7 which includes a number of new capabilities that improve   the delivery and accuracy of speech.

 

The latest version offers support for 21 languages and includes specialized acoustic models for all types of telephony environments as well as for noisy in-car environments.

 

The offering is available in: U.S. English, U.K. English, Russian, Chilean, Mexican and Castilian Spanish, Catalan, Galician, French, Canadian French, Portuguese, Italian, Dutch, Polish, German, Greek, Swedish and more.

 

New features include a "Feature Normalization" algorithm that helps to improve recognition accuracy regardless of how the ASR was trained.

 

Also, a "Continuous Recognition Mode" automatically recognizes short, consecutive commands and an improved "Confidence Score" algorithm improves the calculation of confidence values to help make dialogues more robust.

 

Loquendo ASR multilingual Speech Recognition is available for both Windows and Linux operating systems and the latest offering includes an updated Acoustic Model Adaptation tool which "has been extended to support 16 kHz audio signals." Also, the offering includes Java support built on C/C++ APIs.

 

Intelligent Mechatronic Systems Inc. (IMS) recently partnered with Loquendo to add speech recognition and text to speech capabilities to their new iLane product offering in a bid to make it even easier for customers to continue communicating while driving,

IMS' new iLane product offers a hands-free, voice-based email solution so users can have their email and SMS messages read aloud to them as well as hear on demand content like weather and news updates or make and receive calls – all while continuing to drive safely and keep focus on the road.