Features
Above the Fray
Follow the Money
Market Research
Big Ideas
Technoptions
Developer Insights
Q & A
Opinions

Archives
Calendar of Events
ASP Directory
Other Cahners Pubs

Submit a Release
Meet our Staff
Contact Us
Advertise
Magazine Subscription

Talkin’ about FAAST Windows CE 1.0

by clayton crooks
May/June 2001

Writing applications for Windows CE devices? Check out a product from Fonix called FAAST - Fonix Accelerated Application Solutions Technology - that lets you add speech recognition and text-to-speech, or TTS, capabilities to Windows CE 2.1 applications. The software developers kit is a complete solution for developers who work with personal digital assistants or cellular phones that use Microsoft’s Windows CE.

The main component of FAAST is the FAAST Builder, an integrated development environment, or IDE, for developing FAAST applications. It’s a Windows-based program that displays a visual interface and graphical representation of elements necessary to create speech-enabled applications. The interface is similar to most IDEs and includes a standard menu bar and toolbar. It’s easy to become comfortable with the application, which offers the necessary functions to develop a speech application. Within the FAAST Builder, you can prototype and optimize your settings before you generate a speech content file that is ultimately linked to the Digital Signal Processor, or DSP, engines with standard C tools.

The speech-content file that is generated by Builder contains four main parts: TTS content, automatic speech recognition, or ASR, content, engine runtime parameters and functional flow information. Its main purpose is to instruct the TTS and ASR engines in both input and output functionality. Fonix provides engines as C libraries precompiled for many different DSPs, micro controllers and reduced instruction set computer central processing units. These libraries, which consist of a simple API, include support for TTS, ASR, voice decompression and audio support. You can create a complete application by combining the libraries, your custom code and the speech-content file.

The Fonix ASR technology used within FAAST Embedded works extremely well for small systems like PDAs that offer limited space and mediocre microphones.

It is speaker-independent and does not require training to a particular voice. During development, you have access to the entire ASR dictionary. When building the application, FAAST Builder analyzes the application and reduces the ASR dictionary to specified content.

The TTS technology offers two approaches: unlimited vocabulary or customized vocabulary. Unlimited vocabulary TTS provides an infinite vocabulary of understandable speech output. This system uses speech clusters and provides the means for concatenating them into desired words and phrases. The software allows modifications to emphasis, pitch and pause that provide correct meaning and context to spoken words and phrases. Like ASR, you can access the entire unlimited vocabulary when you are within Builder. Once you build the application, Builder will analyze the application and reduce the TTS size to the selected speech content.

Customized Vocabulary TTS uses a specific vocabulary of predefined words generated from cluster palettes, which contain the pieces of sound for spoken words. Thus you can design applications that produce words that sound as if spoken by a real person. You can use unlimited TTS to develop and prototype your application. Once you’ve finished your application and refined the vocabulary being used, however, you will need Fonix to create a custom soundbank to fit your specifications.

Overall, the FAAST SDK is an excellent tool for Windows CE developers looking to add voice interactivity to the applications they develop. The FAAST Builder offers a simple interface, it’s comfortable to work with and the TTS and ASR engines offer a range of functions that answer many needs.

Clayton Crooks is an independent software developer. Contact him at crooks@planetc.com.