applications

lawful interception / ASR

ASR is the next-generation speech recognition technology for speech-enabled applications. It is speaker-independent and reliably recognizes large-scale vocabulary continuous speech, even in the noisiest environments such as wireless. ASR currently powers services that handle millions of calls every day, such as fully automated directory assistance services, voice portals, and automotive applications.

The Benefits to You…
ASR gives integrators the freedom to create services those are user-friendly and as complex as they want them to be in terms of vocabulary size, interaction flexibility and number of languages. ASR perfectly fits the requirements of each and every application scenario - however complex.
- Broad Vocabulary & Flexible Recognition – recognizes up to 1,000,000 words; supports isolated and continuous speech.
- Highly Accurate Speech Recognition – thanks to integration of neural networks and hidden Markov models, and detailed acoustic-phonetic units trained on large speech corpora.
- Extended Standards Support – optimized for VoiceXML applications; complete grammar standards support, both W3C SRGS 1.0 and SISR 1.0.
- Highly Accurate Phonetic Transcribers – specialized for each language (also used in acclaimed TTS).
- High Efficiency – low-computational power requirements enable a large number of recognition channels to run simultaneously, both with small and large vocabularies.
- Rapidly Extensible to new languages – the methodology that has been tuned for our wide range of languages is rapidly extended to any other.
- Powers Speaker Verification technology.

Simple Yet Powerful Technology…
A complete set of simple and powerful features guarantees truly robust speech technology, enabling:
- Improved barge-in capability to guarantee high reactivity and robustness to noise and background speech.
- A new patented speech enhancement method for improved recognition performances in noisy conditions.
- A flexible rejection mechanism which identifies any linguistic expressions that are not acceptable within a specific domain.
- Dialogue-flow management which is achieved through confidence values provided for all the Nbest hypotheses returned – on a sentence-by-sentence & word-by-word basis.
- Garbage rules definition to match arbitrary spoken sequences not modelled by the grammar. A sophisticated Speech Assistant Toolkit guarantees the rapid and efficient definition of Recognition Objects (ROs) and Recognition Packages, such as Grammar ROs and Language Modelling ROs. In “unpredictable” situations, ROs can be created, stored and deleted “on the fly”.
Significant memory requirement reduction: ROs can be both permanent (and therefore shared by all recognition channels) and dynamic (i.e. loaded run-time when required and discarded once they have been used).

ASR also provides:
- A re-usable built-in grammar library for each language (e.g. date, time, currency, phone numbers, etc.).
- Phonetic segmentation, which includes the phonetic representation and related time-stamps for each phoneme within a sentence. This is often a prerequisite, especially in avatar animation.

ASR Tuning Tools
ASR provides users with a tool package that automatically analyzes data collected in the field to improve service performance, including:
- Phonetic Learning – which automatically analyzes application data to identify frequent formulations that have not been covered and additional pronunciation variants, to improve a speech recognition grammar.
- Acoustic Model Adaptation – further increases recognition performance by using audio material recorded in the field (environment, speaker, channel adaptation), where a vocal application is used in a particular context.

ASR - Technical Specifications

Main Features
- Speaker Independent
- Open Vocabulary
- Noise robustness (e.g. in-car, wireless, etc.)
- Optimized for Telephonic Speech

Basic Technology
A combination of Neural Networks and Continuous Density Hidden Markov Models

Configurable Recognition Modalities
- Grammar based
- Continuous Speech Recognition with Statistical Language Modeling
- Free or Forced Phonetic Decoding

Key Features
- N-Best Decoding
- Confidence Scores at sentence and word level
- Tuneable Voice Detection sensitivity
- Improved Barge-In functionalities
- Speech Complete/Incomplete Timeout
- Garbage rules
- Grammar handling and fast grammar compilation on the fly
- Re-usable Built-in grammar library
- Multilingual grammars
- Voice enrolled grammars
- Natural Language Processing
- Optimized for VoiceXML applications
- Speaker Verification
- Word spotting plug-in

Tuning Tools
- Phonetic Learning
- Acoustic Model Adaptation

Supported Languages
American English, Canadian French, Brazilian Portuguese, Argentinian Spanish, Chilean Spanish, Mexican Spanish, British English, Castilian Spanish, Catalan, Valencian, Galician, Dutch, French, German, Greek, Italian, Polish, Portuguese, Swedish,Turkish, Russian

Grammar Formalisms
- JSGF (Java Speech Grammar Format)
- W3C SRGS 1.0 (XML and ABNF Form) + SISR 1.0

Supported Operating Systems
MS Windows (XP, Vista, Server 2003, Server 2008*), Red Hat Enterprise Linux (3, 4, 5*), SUSE Linux Enterprise 10.0
* also available for 64 bit version

Interfaces
- API (C/C++)
- Intel Dialogic Audio Source support
- DSR support
- Java

CPU Requirements
- Connected Digits Recognition: 80 channels on an Intel Pentium 3.2 GHz CPU
- Grammar with 10,000 words: 20 channels on an Intel Pentium IV 3.2 GHz CPU

Memory Requirements
- 15 MB per language shared among channels
- Few MB per channel depending on the recognition task (e.g. 5 MB for Connected Digits Recognition, 15 MB for a grammar with 10.000 words)

Above mentioned specifications and informations are subject to change without prior notice.

news

23-25 February 2009 BTT exhibited in ISS World MEA, Dubai Intelligence Support Systems for Lawful Interception, Criminal Investigations and Intelligence Gathering at Dubai, as Exhibition Sponsor On February 25th, between 8:30-9:30 am BTT will be demonstrating a Case study in Session A. www.telestrategies.com 7-11 October 2009 BTT exhibited in CEBIT Eurasia 2009 Istanbul Tüyap Fair, Convention and Congress Center - Hall 3 / D18 www.cebitbilisim.com Top 500 IT companies list of Interpromedya is published; BTT Ltd. is announced to be in the Top 10 list in category of Software Development of Security Applications. www.interpromedya.com.tr 17-20 November 2009 BTT exhibited in Milipol Paris 2009 Paris Expo Porte de Versailles in Hall 1. www.milipol.com

private area