Please Note:
This Page is Maintained for Archival Purposes Only.
More Current Information Is Available
Here.
By incorporating speech, the most natural form of communications, voice-activated products make computers "people literate" and easily accessible to a wider audience of users.
Voice-activated products are being used in a wide range of areas, including:
An example of an industry where speech input is invaluable is food processing, where food handlers are required to record FDA information as they perform inspections. Keyboard data entry is impractical because using a keyboard while handling food product s would risk contamination. Speech input lets workers perform these tasks simultaneously without compromising health regulations.
Frees users’ hands and eyes for other tasks. In situations where users need their hands or eyes to perform other tasks, speech input allows them to complete tasks while simultaneously inputting data.
Simplifies computing for novice users. Voice commands are more natural and easier to remember than complex keystroke sequences. Speech input makes it easier to use computers, reducing training time and lowering training costs for new workers.
Improves data entry speed and accuracy. Speech input can result in faster data entry with fewer errors when used instead of manual typing.
Automates processes requiring instant data access. Speech recognition makes it easier to enter data on-line to improve the efficiency of automated processes. Because systems gain immediate access to statistical data, they can react more quickly to corre ct problems and speed throughput.
Protects workers against repetitive stress injuries. By reducing the amount of typing required to run your application, you also reduce workers’ risk of developing Repetitive Stress Injuries (RSI), the leading cause of work related injuries today. Speec h input gives people with RSI an alternative to taking disability leave, and helps them continue to be productive employees.
Developed by Dragon Systems, world leader in PC speech recognition, Dragon VoiceTools is a software developer’s kit that enables you to add a voice activated interface to your application.
Here’s how it works: A user speaks into a microphone attached to a sound card inside the computer. The card converts the sound into digital form. The Dragon Speech Driver, which handles voice processing and conversion, interprets the digital signal and compares it to all the words you have defined in your application. The Driver then determines which word was spoken and returns it to your application.
Dragon VoiceTools lets you incorporate speech recognition as the primary input source or as one source in a combination of input modes. It provides you with everything you need to build applications that recognize human speech as input.
Speech input can result in faster data entry with fewer errors when used instead of manual typing. you can integrate speech into your MS-DOS and Microsoft Windows applications.
Dragon VoiceTools’ speaker-independent acoustic voice models for American English ensure that your application will recognize speech from practically anyone.
You can add the capability to learn and adapt to each user’s voice for improved recognition accuracy.
Speaker-dependent speech recognition
Once the Speech Driver has adapted the voice models to a specific user, they become speaker-dependent. To accelerate this process, Dragon VoiceTools allows you to enroll the user to create speaker-dependent voice models.
You can program continuous speech recognition for digits. This feature is invaluable for speech input of zip codes, account numbers, social security numbers and telephone numbers.
You can integrate speech recognition into any source code that can call C functions.
Dragon VoiceTools contains all of the functions you need to control the Speech Driver’s powerful capabilities.
For best performance, speech recognition functions are performed in a separate Speech Driver that resides in extended memory. The Driver controls:
Word Recognition. The Driver compares utterances against the application-specific vocabularies to determine which word was spoken.
Vocabularies. The Driver handles the total collection of words defined in an application, along with the word pronunciations and frequency information. States. The Driver manages developer-defined application states so you can specify which words are available at any given time.
Microphone input. The Driver sets up communication channels between the speech driver and the microphone, and turns the microphone on and off.
Word adaptation. The Driver can build and adapt word models for the speech patterns of specific users.
User files. The Driver stores and updates acoustic information for words and phrases for each user.
Speech data collection. The Driver collects, saves and loads speech samples for training or recognition.
Developer-specified application vocabularies
Dragon VoiceTools lets you design the vocabularies for your application. The total vocabulary is limited to 10,000 words or phrases. Up to 1,000 words or phrases can be active at any time. You specify the words and phrases to be used within you applica tion in an easy-to-read text format and compile it with Dragon VoiceTools Finite State Grammar Compiler.
In a large-vocabulary application, such as inventory control for an automobile parts store, you could easily divide the inventory into categories. Then you would define a different vocabulary state for each category, such as chassis parts, engine parts, and transmission parts. Each state could have up to 1,000 entries. For example, the vocabulary for the state engine parts might include fuel injector, distributor cap, alternator, timing belt and spark plug.
Flexible Speech Model Options Dragon VoiceTools words by creating models of the sounds of human speech. You decide how to handle the speech models for your application, choosing one or any combination of these approaches:
Use the speaker-independent acoustic voice models included with Dragon VoiceTools. Build and adapt you own acoustic voice models. Include no models, and have users train the speech recognition.
(DOS and Windows) Dragon VoiceTools is a software developer’s kit that lets you integrate speech into your PC applications.
DOS and Windows Speech Drivers The Speech Driver is a program that handles the processing and conversion of speech into data that you program can understand. Dragon VoiceTools includes both DOS and Windows versions of the Speech Driver. Speech Driver Application Program Interface (SDAPI) Library A collection of C language functions that you can link to you program code to control the Speech Driver. These functions work with Borland and Microsoft C and C++ compilers. Finite Site (FSG) Compiler
A tool that creates a vocabulary file specific to your application. The FSG Compiler converts an easy-to-read text file that you create to determine the words and phrases your application will recognize at any particular time.
Derived from a broad range of speakers, speaker-independent acoustic voice models are included for over 110,000 of the most commonly spoken American English words and phrases. These models allow you to develop applications requiring little or no training .
DOS and Windows examples are included to guide you through the process of integrating speech recognition.
Programmer’s Guide A step-by-step explanation of the development process and conventions used in speech recognition.
A list of function calls in the SDAPI library, with explanations and examples of each function.
Microphone The microphone shipped with Dragon VoiceTools Developer’s System is certified by Dragon Systems to provide outstanding performance and noise-canceling capabilities.
Additional Required Component Sound Card (sold separately). For DOS, Dragon VoiceTools requires a Dragon Systems certified DSP audio board with a 1/8 inch microphone jack.
For Windows, you can use a 16-bit industry standard soundcard or a Dragon certified DSP audio board.
To sum it up, Dragon Voicetools is a software developer's kit that lets developers write speech-aware PC software applications for either DOS or Windows operating environments.
Processor: Windows 3.1 or above (not Windows '95) & DOS v5.0 Desktop & Laptop.
Telephone: 800-245-2133
Copyright © 1996 21st Century Eloquence, Inc.
Microsoft is a registered trademark and Windows is a
trademark of Microsoft Corporation® (tm)
Other brand and product names are trademarks or registered trademarks of
their respective holders.