Volker Jantzen, CEO, SVOX
Q Please give us some background on SVOX.
A SVOX AG was founded in April 2000 as a spin-off company of the Swiss Federal Institute of Technology (ETH) in Zurich, Switzerland. Its core product is text-to-speech software. The multiple award-winning SVOX technology is used in fields such as mobile devices, automotive, call centers and telecommunications. SVOX differentiates itself by offering customized text-to-speech. With SVOX's software architecture customers are offered a text-to-speech engine adaptable to their technical and market needs. Alongside clients such as Swisscom, Harman/Becker, D-Link and ONCE, SVOX also collaborates closely with partners such as VoiceGenie, Telisma, Siemens, Ericsson, Fonix and A.R.T.
Q What is your target market?
A SVOX is an enabling technology company that focuses on its core strengths: Embedded text-to-speech solutions.
Consumer Devices. Mobile devices, consumer electronics and appliances represent significant market opportunities for text-to-speech. In order to bring text-to-speech capability into various devices we have partnerships with mobile phone manufacturers, consumer electronics companies, microprocessor chip manufacturers and embedded operation system companies. Automotive. New governmental regulations are forcing telematics manufacturers to provide secure and hand-free content to drivers in navigation systems, in-car phones and on-board controls. TTS is the underlying technology that allows translating digital text into speech. Consumer Applications. Mobile phone manufacturers, mobile carriers with branded ODM phones and multimedia content (e.g. ringtone) providers are all looking for new revenue streams from innovative products such as SVOX voice download or MMS. SVOX is the only TTS vendor that has solutions for the Symbian OS. MMS. Currently, the first mobile handsets with multimedia message service (MMS) capability are reaching the mass market. We are working with talking face companies to make MMS with unique cartoon character and human faces that speak with TTS. Visually Impaired. With our partner Code Factory, we developed Mobile Accessibility, a unique application for the Symbian OS that allows people with visual problems to utilize a mobile phone to its fullest capabilities.Q What differentiates your text-to-speech (TTS) offering from those of your competitors?
A SVOX offers the most modular and scalable architecture: The SVOX kernel software is language-independent. All language-specific data are kept in separate lingware packages. Each lingware package consist of a language module and one or more voice modules. Thus, the SVOX kernel only needs to be integrated once while additional languages and voices can be added later without additional software integration efforts. Even the size of the lingware packages can be modified any time during the development process without any changes to the SVOX software kernel. The SVOX architecture guarantees for highest possible software reuse and enormous cost savings over the whole software integration and product life cycle.
SVOX offers mixed-lingual TTS: The modular architecture enables the unique feature that various languages spoken with the same TTS voice even within one sentence. Especially in Europe, many text like e-mails and news contain English inclusions German or French texts which are handled correctly by SVOX's mixed-lingual TTS technology.
SVOX allows corporate branding: SVOX makes it easy and cost-efficient to generate corporate voices that enable our clients to brand their applications with their own custom voice.
SVOX has a very small footprint. Due to optimized concatenative techniques used by SVOX TTS, the software can be compressed small enough (under 1 MB) for embedded uses without noticeable loss of synthesis quality.
Q What languages do you currently offer and what languages are planned for the future?
A Currently with have various voices in the following languages: German, U.S. English, Italian, French, Spanish and U.K. English. Portugese and Japanese are under development. In 2004, among other languages, Dutch, Swedish, Danish and Chinese Mandarin will follow.
Q How can TTS voices help to build a company's "brand?"
A SVOX allows corporate branding. SVOX makes it easy and cost-efficient to generate individual voices that enable our clients to brand their applications with their own custom voice. The corporate SVOX voice becomes a design element of the product or application (e.g. within a car or on a mobile device). Corporate voices can also be used to target specific market segments such as the children, teenager or professional market.
Q Please describe an innovative deployment of SVOX's TTS.
A SVOX developed with its partner Code Factory Mobile Accessibility. This is a unique application for the visually impaired market which allows people with such a disability to utilize a mobile phone to its fullest extend. Visually impaired people are able to receive SMS/e-mails, write them back, check their calendar, ect. We provide this application for Symbian Series 60 phones as well as for the Smartphone SPV in Europe. SVOX has been inundated with orders for this application and we plan to expand the availability to other parts of the world and mobile operating systems in the near future. We are also working on an application for the "Elderly Market," a huge untapped market which will allow elderly people to easily operate a mobile phone and utilize some of the features which today are intimidating or impossible to them.
Q Compare the differences in embedded TTS compared to server-based TTS in terms of quality, footprint, robustness, etc.
A Server-based TTS sounds more natural than embedded TTS due to the footprint. Embedded TTS is compressed which - depending on the compression rate - has some effect the voice quality. Nevertheless, the TTS is very understandable and pleasant to listen to. Server-based TTS has footprint in the magnitude of 200 MB, whereas embedded TTS can be below 1MB, as our products SVOX Mobile and SVOX Genie have.
Q What do you see in the future for TTS technology?
A SVOX's research department is working on a revolutionary technology that allows the transformation of one voice into another. We extract certain parameters of a speaker's voice, e.g. by letting the person read one or two paragraphs. With a proprietary algorithm our basic diphones are converted to match the characteristics of the speaker's voice. With this technology, for example, celebrity TTS-voices could be produced or emails could be read with the sender's voice when a voiceprint is attached. Furthermore, we are always looking into making TTS emotional. We want our voices to be able to laugh, cry and smile. Lastly, offering the best possible voice at the lowest possible footprint is an ongoing effort here at SVOX.