Voice Database Set Capture and Manual Development

VFI possesses the world’s only facility specifically designed and operated to provide end-to-end voice capture and consulting services to the speech industry.

VFI has the capacity to produce over 200 voice database sets (voice sets) per year. The production of a voice set calls for the precise, phase-linear, reverberation-free recording of a pre-trained human voice articulating custom-scripted words and syllables to capture a full set of context-sensitive speech units (phonemes, diphones, and triphones) for a given language or dialect. The importance of a completely reverb-free environment, completely isolated from any extraneous sound, is critical for the successful capture of speech for TTS applications. VFI’s anechoic chamber provides such an environment.

Following capture of a voice set, a procedure called “tagging” is performed to permit the voice set to be converted to a voice font.  The voice font is then operated by a voice synthesizer in real-time. The process is called “concatenative synthesis,” distinguishable from the more elementary production of a “pre-recorded prompt” (PRP).  VFI also provides world-class recording for the production of PRP, IVR and Automated Speech Recognition (ASR) using the same talent under contract for TTS applications.  This assures seamless results and consistency in voicing for combined TTS/PRP/IVR/ASR applications.

By employing VFI’s all-inclusive voice capture and recording services, clients can realize substantial cost savings, while being assured that they are receiving the best the industry has to offer in this highly specialized and demanding field.

Voice Database Set Capture

VFI has unparalleled experience in capturing custom voice database sets (voice sets) for the development of custom voice fonts for TTS synthesizers. VFI has captured thousands of hours of voice samples specifically for TTS applications and has extensive samples in dozens of languages and dialects, adaptable to a myriad of applications. By partnering with VFI to create custom voice sets at off-the-shelf prices, clients can capitalize on VFI’s…

• Anechoic Chamber:  the only one of its kind available for large-scale, commercial applications

• Experience working with speech synthesis R & D teams to maximize platform-specific prosody and pronunciation during voice set capture

• Ability to carry out specialized voice capture that meets the most stringent technical specifications, attaining levels of resolution that surpass industry standards

Database Manual Development

Database manual development has proven to be both time consuming and expensive when done in-house.  VFI’s proprietary development methodologies free up valuable resources and significantly accelerate your delivery times while reducing costs.

Domain-Specific Manuals

Until now, “domain-specific,” or industry-specific development of manuals for TTS implementation has been an extremely costly, time-consuming proposition. Today, thanks to VFI’s unrivalled experience and proprietary methodology for the development of domain-specific TTS manuals, clients can realize unheard-of savings for even the most challenging applications, not only in hard dollars, but also in time-to-market. Among others, VFI has developed domain-specific manuals for the following industries:

·        GPS-based navigation

·        Medical referencing and prescription instructions

·        Financial reports and business briefings

·        Legal notices and insurance explanations

·        Shipping instructions and delivery guides

·        Meteorology reports and forecasts

·        Transit arrivals and onboard announcements

·        Inventory pull-lists and material control assessments

·        Name/place infrastructure for call centers

·        Time/date in voice mail applications

·        Reservation confirmations/flight progress reports

·        Entertainment applications

General Language Manuals

VFI also produces baseline manuals for any language or dialect. As with domain-specific manuals, VFI’s unparalleled experience developing foreign language manuals, combined with our proprietary methodology, means that clients receive the best service, at the best price, for custom-designed material.

Fast Service

VFI offers the fastest turnaround times in the industry for voice selection (3 weeks), manual development (3 weeks) and voice set capture (3 weeks). VFI can deliver a completely customized voice set to the client in an average of nine weeks in the following standard languages:

·       English, any American, British or other variant or dialect

·       Middle Eastern Languages, such as Arabic, Hebrew, and Urdu

·       Asian Languages, such as Japanese, Chinese (Mandarin or Cantonese), Korean and Hindi

·       European Languages, such as French, Spanish, Portuguese, Italian, German, Russian, and Polish

For less common languages and dialects, turnaround time is calculated on a case-by-case basis.

Voice Preservation

From time to time, TTS vendors receive requests for voice capture from the general public.  Often, such an inquiry is made in the context of an individual’s impending loss of voice function due to illness or death. These people anxiously desire to convert their voices into TTS voice fonts so they can continue to be “heard” as they sound now. This demand has rarely, if ever, been accommodated, in spite of the fact that it will likely grow over time as the cost of converting voice recordings into voice fonts falls dramatically.

In the future, production efficiencies and falling costs will enable anyone to synthesize his or her voice at an accessible price. VFI’s production expertise as TTS voice database suppliers allows us to answer this demand today. 

[back]