Bluetooth

Rubidium’s Hands-Free Voice User Interface (VUI) for CSR’s BlueCore5™ and CSR8670 Bluetooth® Chips

Rich Development Package includes Sample Source Code, Easy to Use API and Documentation and Provides for Seamless Integration and Fast Development Cycle

Rubidium Ltd, a long standing CSR “eXtension Partner” and a Bluetooth SIG member, has packaged its industry proven multi-lingual hands-free Voice User Interface (VUI)  into a self-contained software package for integration into Bluetooth applications based on CSR’s Bluecore 5™ and CSR8670® chips. Rubidium’s Bluetooth VUI package includes: Automatic Speech Recognition (ASR), Text-to-Speech (TTS), Voice Trigger (VT), Beam Former (BMFR) and Compressed Speech Playback (CSP) technology.

Rubidium’s hands-free VUI provides Bluetooth device vendors with a cost effective, easy to integrate and field proven solution. The package is integrated into CSR’s development environment, offering seamless integration into virtually any Bluetooth-enabled application without additional hardware or BOM items.

A simple-to-use and comprehensively documented API provides in-depth access into tuning parameters, but it can also be used out-of-the-box with default values. In addition, the package also includes reference source code of a “sample application” that demonstrates how it is used and integrated. This follows Rubidium’s policy of offering its customers dedicated, hand-in-hand support throughout the implementation process.

The technology is offered in two suites, depending whether the application should only support speech output or both speech recognition and playback

Some of the key functionality includes:

•  Out of the box recognition for any speaker and any accent- just talk and it works
•  Natural sounding compressed speech output supporting wideband 16kHz pre-recorded prompts
•  Concatenated TTS technology based on real human speech announces any name and number, acronyms and abbreviations
• Voice Trigger for continuously listening and low-power, fully hands-free activation
• Available in a variety of languages and accents (US English, UK English, European and Latin Spanish,  German, French, Italian, Mandarin Chinese, Japanese, Russian, Polish, Dutch, Brazilian Portuguese, and  others TBA)
• Fast delivery of any customer requested language or custom voice
• State of the art noise robustness maintains high recognition performance in harsh conditions
• Patent-pending Barge-in functionality allows the user to start speaking while an announcement is playing
• Advanced proprietary Beam-former algorithm (BMFR) provides dual microphone noise suppression and improved  ASR performance in noisy environment
• Very small footprint, low-resource usage
• Voice “personalities” (grumpy, happy, sexy, etc.) and celebrity voices (cartoon characters, movie stars, etc.)  available for unique sound and product differentiation.

Speech Recognition (ASR)
Reliably controls a large array of functions through voice commands, including call acceptance and rejection, pairing, voice dialing, battery status, music control and music selection.

• Can switch vocabularies in real time to follow the dialog’s state
• Adjustable recognition sensitivity and recognition timeout
• Noise robustness to provide best recognition performance in a variety of operating conditions such as – car, office and other noisy environments
• Integral beam-former supports dual microphone input for improved noise suppression
• Zero latency – results provided immediately when the user stops speaking

Text-to-Speech (TTS)
 Announces any name and number in any language with natural and human-sounding wideband voice:

• Announces names and numbers without programming or recording
• Adjustable volume, intonation depth and pitch
• Selection of multiple languages, accents and voices
• Large dictionary ensures correct pronunciation for optimal accuracy
• Support of acronyms, abbreviations, etc.

Voice Trigger (VT)
 “Magic Phrase” Provides Hands Free Triggering

• Integral Voice Trigger mode allows wake up of the product when the user speaks the “magic phrase”
• Completely hands-free user experience for ultimate user comfort and safety
• Optimized resource consumption allows continuous listening with minimal battery drain
• Accurate algorithm minimizes mis-alarms and false alarms

Compressed Speech Playback (CSP)
High-quality pre-recorded speech prompts can be used to play pairing instructions, connection status,  remaining battery time and other information.

• Wideband voice for highly intelligible and crystal-clear voice quality
• Any requirements of pre-recorded voice prompts can be supported to match any set of languages, announcements and voices (even several languages and voices in a single product!)
• Resource-effective due to highly-compressed low-bit-rate coding
• Number of prompts in a single product is virtually unlimited
• A large set of prompts is already available in many languages
• Rubidium provides recording services to support customer requests for unique prompts and voice talents, including tailor made voice personalities