| ||||||||
Cool Mac Gear• iPod Video • iPod nano • iPod 1G-2G • iPod 3G • iPod 4G • iPod Mini • PowerBook-iBook • Garageband |
Charles Moore Reviews iListen 1.6 Dictation Software For OS X
I’ve been a cheerleader for MacSpeech’s iListen dictation software since I first heard about it back in 1998. Not that I have anything against IBM’s a rival product, ViaVoice, which has its comparative virtues. It’s just that iListen is a Mac - only application engineered by a team of developers who are committed to the Mac platform. MacSpeech was founded by Andrew Taylor and several other alumni of the Dragon Systems PowerSecretary team. PowerSecretary was the first generation of dictation software for the Mac, RAM-hungry, and poky, and very expensive ($395), but better than no dictation software at all, but its discrete speech recognition engine required users to....speak....with....pauses....between....each....word dictated, which slowed things down, interrupted one’s train of thought, and the stop-start often proved stressful and fatiguing to the vocal cords as well. Dragon abandoned the Mac platform in early 1998, leaving the field open to other contenders. One was a port of IBM’s PC continuous speech dictation program, ViaVoice, The MacSpeech folks originally started with a clean slate, but eventually adapted the Philips FreeSpeech 2000 speech engine developed by t Netherlands-based Philips technology. MacSpeech’s original mission statement back in 1998 was music to a Machead’s ears: “We are Macintosh people, committed to the Macintosh platform. We have “* We believe in Speech and we believe in Macintosh. “We hope that’s enough for you.” By comparison, IBM’s commitment to the Mac seems a bit perfunctory. They did get a continuous speech Mac dictation product out the door first - - ViaVoice Millennium edition in the fall of 1999, but ViaVoice for OS X, which also beat iListen’s OS X product market by more than half a year, has not been significantly upgraded (other than a bug fix patch for OS 10.2.2) since it was first released in the fall of 2001. It’s a good product, with amazing speed and accuracy, but it can also be troublesome to install and keep working smoothly once you get it installed. The first public version of iListen was released as a preview at MacWorld Expo New York in 2000, and a final release, iListen 1.0.1 followed in the fall of that year. It was RAM hungry, and more than a little buggy, and had no spelling correction module, but it did have one big advantage over the ViaVoice Millennium Edition -- you could dictate text into virtually any text field that would accept keyboard input. ViaVoice ME required you to dictate into the program’s dedicated SpeakPad word processor, and then to transfer your transcribed text to its final destination via the Clipboard. iListen 1.1 followed in November, 2001, and versions 1.2 and 1.2.1 coming along in January and April of 2002 respectively, finally adding a spell correction mode and the ability to “learn” from corrections. iListen with OS X support, version 1.5, arrived in August, 2002 as a Carbon application that would work in both OS 9.2.2 and OS X. The problem was that it worked a lot better in OS 9.2.2, at least on moderately slow Macs like my 500 MHz Pismo PowerBook, than it did in OS X. When I say “better,” as I mean faster. While transcription speed in OS 9 was satisfactory, iListen 1.5 lagged horribly in OSX, making you spend a lot of time waiting for it to catch up. I found it easier to put up with ViaVoice’s foibles in order to get its all but real time speed, although iListan1.5 was a pleasure to use in OS 9.2.2. iListen 1.5 also brought a significant productivity enhancement to the Mac platform for the first time - Transcription. “Using a Digital Voice Recorder to record some audio for later transcription is an incredible time saver”, says MacSpeech’s Donald MacCormick. “iListen’s ability to read audio from AIFF and WAV files and put the recognized text directly into almost any application is remarkable”. The latest version of iListen is 1.6, released last week, although I’ve been experimenting with a prerelease version for over a month. I’m happy to report that the first thing that grabbed me upon checking out the new iListen 1.6 final is that it’s a lot faster. ViaVoice is still the speed king, but the improvement in this version of iListen is impressive, and its transcription lag is no longer an inhibition to to using at an OS X. Which is a good thing, because iListen 1.6 is not supported by the Classic Mac OS. However, Classic users have not been abandoned. An upgraded version 1.5.5 of iListen, incorporating some but not all of the version 1.6 enhancements, is bundled with iListen 1.6. MacSpeech says that version 1.5.5 will probably be the last upgrade of the classic iListen application. iListen’s installation process is generally painless, however once the application is installed and started up you still have quite a bit of work to do (typical of all dictation applications). The first time you launch iListen, it will walk you through the steps necessary to set up your microphone and create your first voice profile. After this initial setup, iListen will automatically begin your first training session. You do this by reading to the program aloud some stories so it can learn your vocal characteristics. Multiple users can share iListen on a single machine, but each must create his or her own profile, and must train iListen separately. Initial training will take at least an hour, as you will need to read several of the stories and then let the computer process them. MacSpeech suggests that you read as many stories as you can manage to further improve your recognition accuracy. The more stories you read, the better your accuracy will be. Both iListen and ViaVoice are big, RAM and processor power hungry applications. I used iListen up to version 1.2 in OS 9.2 on a 233 MHz G3 PowerBook with 192 MB of RAM, and it was a lot better than no dictation software, but performance was pretty draggy. I would suggest that a 400 MHz Mac with at least 512 MB of RAM would be the minimum system you would want for reasonable performance. iListen runs well on my 500 MHz Pismo PowerBook and 700 MHz iBook, both with 640 MB of RAM. In my estimation, by iListen’s biggest advantages over its ViaVoice competitor are: 1. Its comprehensive, no hassle, “dictate anywhere” capability. While the OS X version ViaVoice has the ability to dictate directly into a certain other programs, it’s really only at its best when transcribing speech into its own dedicated SpeakPad word processing application, from which you transfer dictated text to its ultimate destination via the Clipboard or in certain instances with a voice commands to supported applications. With iListen, you can dictate directly into virtually any application or text field that can accept typed text, and this facility makes iListen significantly less cumbersome and more versatile to use than ViaVoice is. 2. IListen is easier to install, and to keep working once it is installed, than ViaVoice. On the other hand, I found that iListen demands more training -- reading stories to the application so it can learn your voice -- than ViaVoice does. I got pretty good accuracy after reading ViaVoice X just one training story, and never bothered with more than three. I’ve found that you need at least four all with iListen for starters, and it doesn’t hurt to read a few more than that. However, at least once that task has been completed, you don’t have to do it again so long as you’re sure to keep your voice profile files backed up. 3. iListen has been a happy camper with a fairly wide variety of mics in my experience, everything from cheap PC - style microphones to Apple PlainTalk microphones, to the Andrea USB headset that ships with ViaVoice, to the VXI Parrott headset mic that is one of the mics MacSpeech officially recommends (and sells). MacSpeech suggests that for the best dictation results, you use a quality, noise canceling microphone, and that’s good advice, but their program also works well with lesser equipment. ViaVoice X, on the other hand, demands its own dedicated microphone. 4. IListen is also cheaper to purchase them ViaVoice, although it’s an apples and oranges comparison, since the IBM product ships with a microphone and iListen doesn’t. However, even with of the recommended VXI Parrott headset added, iListen still comes in cheaper than ViaVoice. Listen has three modes: Dictation, Spelling and Command. Dictation is the function I use most. You use Command Mode to control your Mac with your voice. In Command Mode, you can click buttons, control windows, launch applications, and use speech instead of pressing keys on the keyboard. Commands in iListen are written in AppleScript, the scripting language built into Mac OS. Commands can be global, meaning they work everywhere, or application specific, which means they only work in a particular application. With iListen’s Command mode you can launch applications, visit web sites, press onscreen buttons, Aside from the trip to the speed shop and a round of bug fixing, MacSpeech says that iListen’s correction function is substantially improved from the 1.5 build (you can read the list of changes and enhancements in version 1.6 in the appendix at the end of this review). That’s good because because it needed improving. Frankly, I find the correction functions in both iListen and ViaVoice to be so clunky and annoying to use, that I just don’t bother with them much, which is too bad, because that’s how both programs “learn” to improve your speech recognition accuracy, but I tend to find it quicker and less frustrating to just correct manually. I would also like to be able to paste text into the correction window from anywhere at any time and be able to correct it. The current system is cumbersome. ,p.However, as MacSpeechÕs Donald MacCormick explains:
Partly a philosophical issue, I guess. I appreciate the reasoning, and I will grant that iListen 1.6’s correction mode really is a lot better than it was in earlier versions, but it’s still not very good, in my estimation. I tried it out while dictating this review, but gave up in frustration after a few paragraphs. No doubt with more practice and careful study and memorization of the user manual (PDF), (and perhaps a faster machine than a 700 MHz iBook) I could get iListen correction working better, but I prefer software to be intuitive and easy to use, and correction mode in both this program and its competitor are neither. Uncommon or misrecognized words can be easily entered using iListen’s Spelling mode. You can use Spelling Mode for mis-recognized words that are used infrequently, such as surnames. iListen has a built-in vocabulary of approximately 30,000 words, with a background lexicon of 300,000 words developed in association with Oxford University Press. Other features: iListen’s “Learn My Writing Style” feature allows you to drag a text document you have written into a dialog window and select words you want to add to iListen’s vocabulary. iListen will add the words you selected to your personal vocabulary file. iListen already knows many rules regarding phonemes, and knows how you pronounce those phonemes from the training you have done, so no additional effort is required to add new words! You can also crate text macros for often repeated text such as signatures and form letters. Your text macros can be up to 32,000 characters in size (about 8 pages). iListen’s Voice Launcher will allow you to open over 150 Macintosh applications simply by saying “Open” followed by the application’s name. Voice Launcher commands are can be edited too. For instance, if you would rather say “Open Word” instead of “Open Microsoft Word” simply change the command name in iListen’s item editor. iListen’s list of Web Favorites will take you to over 150 useful web sites. You can jump to a web site by saying “Jump to” followed by the name of the web site. You can also add your own web sites easily using iListen’s item editor. ilisten’s Transcription feature will generate text from WAV and AIFF audio files. Using a Digital Voice Recorder, you can record now, transcribe later. iListen is a powerful program. I would have to say that the IBM ViaVoice speech recognition engine is better than the Philips FreeSpeech engine, with more speed, better accuracy (at least after initial training respectively), and a larger vocabulary, but iListen is a more comprehensive and versatile speech software package, especially if you take the time to learn how to ring its bells and toot its whistles, so to speak. With version 1.6, iListen is a significantly improved program. It’s far from perfected yet, and still has rough edges, but so does the competition, and this release has done considerable smoothing. Dictation software, even at tis present, imperfect stage of development, is a great convenience, especially if you suffer from typing pain or other handicaps that make typing difficult. If that sounds like you, or you just ant to try hands-free typing, iListen is a wonderful tool to have in your software library. System requirements: iListen suggested retail price is $99 without headset, $149.00 with headset/microphone. iListen packages Microphones For more information or to purchase, visit: iListen is also available at the Applelinks Think Different Store: Appendix 1 - iListen 1.6 Feature Summary Appendix 2 - New in iListen 1.6
Page: 1 - 2 - 3 - 4 - 5
|
Cool Mac Gear• iPod 1G-2G • iPod 3G • iPod 4G • iPod Mini • PowerBook-iBook • Keyboard Skins • Garageband
| ||||||
|
| ||||||||
