EXCLUSIVES

ViaVoice and iListen: Talk to Your Mac!

By Louise Masurat

ViaVoice for Mac OS X by IBM, and iListen 1.5 by MacSpeech are two Jaguar-compatible speech-recognition programs. Not only can you dictate text with them, you can also control your computer by using voice commands rather than a mouse or keyboard. In order to do so, you must first go through a training period so the programs can learn your voice and speech patterns.

Both programs allow multiple users to create their own voice profiles, which can be backed up ­ and even transferred to a different computer, so you donąt have to go through the entire (tedious!) training period again if you're switching between a laptop and a desktop computer. With both programs you dictate at a normal speed and in a normal tone of voice, and both actually analyze your speech patterns, not just how you pronounce words. Both allow you to add words to their built-in vocabulary, and they can even analyze your text documents for words to add. Both can read back text you have dictated and recognize voice commands that control your computer for opening applications, surfing the web, and more. And both let you easily create macros to enhance their native capabilities.

With ViaVoice, you dictate into IBM's SpeakPad, where you can say voice commands and dictation interchangeably. It allows natural dictation of numbers, dates, times, measures, and prices. You can dictate letters, reports, and e-mail and then transfer them to your favorite Mac program. Within SpeakPad, ViaVoice affords you multiple ways to correct and edit text.

If you have an earlier version of ViaVoice and are now using OS X, you should probably upgrade. According to IBM, this edition offers the following new and improved features: Aqua user interface/look and feel throughout the application; a faster training period; redesigned Voice Center; new Command and Preference Centers; optimization for G4 with Multiple Processors and for G4 with Velocity Engine; dictation, editing and formatting into ViaVoice SpeakPad; direct dictation into text fields of Mac applications; voice launching and surfing websites in Internet Explorer; voice navigation of the Finder; a vocabulary more than twice the size of any previous ViaVoice for Mac, providing greater accuracy, IBM ViaVoice Outloud Text-to-Speech (TTS) technology.

ViaVoice for Mac OS X requires a G4 or at least a 300 MHz G3 processor. Power Macs and PowerBook G3 systems introduced before August 1998 are not supported, nor are G3 systems upgraded from older machines. Also required are 192 MB RAM (256 RAM recommended), 600 MB of available hard drive space, a USB port, audio output jack or USB speakers, and a CD-ROM drive or DVD drive.

At $171.00, ViaVoice which includes a noise-canceling digital USB headset microphone. There are also available legal and medical vocabularies at $142.00 each.

The Medical Vocabulary is for general medical dictation, and is derived from analysis of thousands of actual patient records. It contains over 25,000 medical terms, allowing dictation of patient history, complaints and symptoms, general examination findings, laboratory, pathology, and radiology results, and course-of-action and follow-up therapy.

The Legal Vocabulary is for general litigation dictation, and contains over 25,000 common legal terms, carefully selected from recent legal literature, which can allow users to dictate words used in both civil and criminal legal documents; dictate legal abbreviations and citations as naturally as if speaking to a colleague or assistant; dictate Bluebook legal abbreviations; and dictate common Latin terms and phrases. Information for product purchase: Visit: www-3.ibm.com/software/speech/ Or call 1-888-ShopIBM.

iListen 1.5 is a Mac-only application. As with ViaVoice, you are able to dictate text, edit it, and format it with just the power of your voice. However, dictation, transcription, editing, formatting and speech navigation can be performed within any application, mostly freeing you from the keyboard and mouse. You can simply dictate anywhere you would type. TalkAnywhere technology allows you to dictate and correct directly into almost any application in Mac OS 9.2.2 or Mac OS X.

iListen also provides speech navigation. You can press buttons, control the mouse, navigate the Finder and file dialogs, open and close files, print, etc. . . . all by voice.

In addition, Transcription mode will generate text from your WAV and AIFF audio files, so for the first time on the Mac you can be productive with your digital voice recorder: record anywhere, transcribe later. And iListen's new Dock Menu provides you with one click access to iListen's most used functions.

According to MacSpeech, these are iListen's major features: Reliability to dictate in a normal tone of voice and at a normal pace using the Philips FreeSpeech 2000 speech engine, dictate, edit and format text directly in any application, correct and update your voice model in almost any application, hear what you said in your own voice; transcribe using a digital voice, recorder such as the Olympus DS-2000 with your Mac (let iListen read and transcribe your AIFF or WAV audio files into text).

It also has a vocabulary of more than 330,000 words developed in cooperation with Oxford University, speech control of the computer through built-in normal language and individually created commands, the iListen Floating Palette for constant feedback, multiple voice profiles, and text macro capabilities (insert up to 32,000-character standardized or predetermined phrases and text with a single spoken phrase). You can also open, close, and control applications and navigate the desktop by speech. It is AppleScript compatible, and Internet ready so you can surf the web with your voice.

iListen works with Mac OS 9.2.2, Mac OS 10.1.5, and Mac OS 10.2. It requires a Power Macintosh G3 or Power Macintosh G4 computer with at least 192MB of RAM (256MB RAM with Mac OS X). iListen product info can be found at: www.macspeech.com/products/iListen.html Purchasing info: http://store.macspeech.com/ Price: $99 without headset.

The major differences

With ViaVoice, you can dictate into any application, but can only correct if you dictate into the SpeakPad. iListen lets you correct in as well as dictate into any application. But the programs differ in correction options. ViaVoice offers you a choice of substitutions for words or phrases you identify as incorrect, and you can then speak the number of your choice, or you can repeat the misidentification until ViaVoice gets it right ‹ but of it does not you have to type in your correction. And it's very easy to add words to the vocabulary list. In iListen, you type the number of the correction you want, and adding new words is more difficult. Also, ViaVoice has an excellent manual, a nice card that lists voice commands, and excellent online help. The only help provided with iListen is a manual in PDF format that gets installed on your computer with the program. Of course, within the program you can get a list of voice commands available to you.

ViaVoice has separate editions for OS X and earlier operating systems; iListen will run on both 9.2.2 and OS X. With ViaVoice you can only use the (USB) headset provided with the application or a limited number of other headsets; iListen can be used with almost any microphone headset, although you can buy a bundle from MacSpeech that includes a headset and even, if you need it, a Griffin iMic so you can use iListen with an iBook or a Cube. And iListen will even let you connect a digital recorder to your Mac and use that as input, so you can dictate on the go and then let iListen transcribe your text into the word processor of your choice.

IBM distributes PC versions of ViaVoice in different languages in different countries. But for English-speaking Mac users who also need to write in other languages, MacSpeech has announced a series of LanguagePaks that will add support for a range of languages. Shipping as of this writing are UK English and Italian. Scheduled for 2003 shipping are Dutch, German, French, Spanish, Portuguese, Finnish and Swedish. Each LanguagePak will be available for $75.00. And MacSpeech does plan to release medical, legal and other special professional vocabularies in the future.

The bottom line

Both programs require a significant commitment of time to reach a level of output with few enough mistakes to justify dictating rather than typing for the ordinary user. If, however, you have a compelling reason to justify speech input, either program can be trained to perform with amazing accuracy. Each has its advantages. I happen to prefer the way ViaVoice handles corrections (even though it will only do it in the SpeakPad), and if you will be doing medical or legal dictation, ViaVoice is the clear choice. But iListen is an evolving project by a company committed to the Mac, and by the time you read this a free service release with a number of enhancements should have become available. iListen gives you more flexibility in your choice of microphone and I'm sure many people prefer being able to dictate and make corrections within numerous applications. Also, if you need to dictate in more than one language, iListen is the only choice.