 |
|
|
| |
Telisma to energize the VoiceXML ecosystem with teliSpeech, its new Plug&Talk Networked Speech Recognition generation.2007 press releases | 2006 press releases | 2005 press releases | 2004 press releases
03.03.2004
| >Table of contents Print | Telisma to energize the VoiceXML ecosystem with teliSpeech, its new Plug&Talk Networked Speech Recognition generation.
| | | | 
| | Based on a new core architecture, teliSpeech features a long list of innovative technologies such as patented decoding on the fly, eXtra Large Vocabularies or noise catcher. TeliSpeech has been conceived to remove the barriers to mainstream speech services, to pave the way for the future and to become the preferred ASR technology for voice services providers.
Paris, May 5th 2004 – Telisma, European leader in speech recognition technologies for telecommunications announces the launch of its new software generation: teliSpeech. Driven by a user-centric approach, teliSpeech has been conceived to anticipate market requests for free expression, real-time interactivity, and customization. Developed in telisma’s labs, teliSpeech represents a milestone in telisma’s history. CEO Laurent Balaine says:
“TeliSpeech is a major “leap forward” for us, a winning ticket to the future. Speech is going mainstream and only software designed from the very start for standards will be able to support the innovations that will emerge in the next few years. A new generation of speech-enabled services is coming up, which will very simply be able to answer quite complex multi-parameters requests almost instantly. Services which will deal with much greater vocabularies than we are used to, and that will give access to updated information in real-time. The idea behind teliSpeech is to give an immediate answer to VoiceXML speech-enabled services expectations but also to anticipate the technical challenges this new generation of speech services will impose and to be ready to empower them.” Being 100% standard compliant, teliSpeech helps ASR go mainstream
TeliSpeech supports every standard, whether it’s a current or forthcoming standard (VoiceXML, SRGS, MRCP, ….). TeliSpeech thereby aims to become a speech recognition reference in an open, democratized marketplace. Typically, teliSpeech is the only networked speech recognition technology today which has been designed with native VoiceXML specifications in order to reinforce the essential bridge that voiceXML has built between speech and web technologies. TeliSpeech offers users a new speech experience
A lot has been done with teliSpeech in terms of both recognition accuracy and dialog fluency to enhance the user’s acceptance of speech-enabled services. No matter who you are, and where you call your voice service from, it understands you better with teliSpeech. Here are a few things teliSpeech does for you: | - Its unique Noise Catcher identifies any weird sound as a garbage-noise to be discarded and ignores it. So if you cough, the system will ignore it and go on… instead of asking you a weird question or interrupting itself.
- Calling from a mobile line, you benefit from the system’s Auto adaptative signal processing and special training to get better performance. And if it’s awfully noisy, do not worry, the ASR Client Voice Activity Detector and the Noise Filtering Systems detect the speech signals and clean them if necessary.
- A bit impatient? Telispeech offers the best real-time barge-in response time on the market.
- As for accents, teliSpeech has a real gift for understanding everyone. It can even take into account current foreign mispronunciations!
|
TeliSpeech adds value to existing services
TeliSpeech makes possible the creation of innovative services with high value added. It namely opens the way to: | - Direct access to hundreds of thousands of names
TeliSpeech can handle vocabularies as large as 2 million words. And it does it fast.
“Who would you like to call in London?”
“Dora Smith”
A directory assistance service using this technology will be able to understand this type of request and automatically dial the number associated with the name in no time. What if Dora Smith moved yesterday? Because teliSpeech reduces compilation time to a few seconds, it can update any directory service featuring thousands of names in real-time. - Sophisticated multi-criteria requests
TeliSpeech can manage complex grammars and perform semantic interpretation in the meantime. These innovative techniques open the way to much more sophisticated dialogs and less directed modes of expression.
Here’s the type of dialog teliSpeech is perfectly capable of dealing with:
“Hello, you have reached reservations, which train are you inquiring about?”
“I want to go from Berlin to Paris Monday night, any train leaving between 7 PM and 10 PM is OK.” - New dynamic voice interaction
Dynamism and real-time interaction are the major themes in speech today. And teliSpeech positions itself well ahead of its competitors for these features. For instance, teliSpeech enabled directory assistance services can compile a grammar of names depending on geographical criteria on the fly. - Multilingualism
TeliSpeech already supports 12 different languages including all major European languages, and new ones are going to be added. These languages are coupled with a unique real multilingual recognitionsystem that supports as many languages as desired within the same dialog call.
|
TeliSpeech's distributed platform guarantees top-quality services
TeliSpeech Voice Distributed Architecture meets strict mass market telecommunication constraints of quality of services and scalability. It is the only ASR solution on the market to present no point of failure. For carriers and enterprises, teliSpeech translates into optimized response time, and better accuracy for given resources. TeliSpeech reduces time and costs of deployment
TeliSpeech benefits from an extremely dense and very light architecture that translate into maximized efficiency. Intelligent compilation and decoding on the fly contribute to teliSpeech optimizing your resources.
Based on standards, you don’t need a Phd to install teliSpeech A few clicks are enough, especially in the MRCP version. In the same way, reliance on the XML language and friendly user interface tools enables teliSpeech to offer simple and friendly programming protocols that drastically reduce conception, development and test times. Finally, a new log management architecture for capturing detailed information permits easy and complete analysis and tuning of teliSpeech’s processes. |
|
|
| |
