NVMUG eNews 4/08/2006
Last updated 4/17/2006
iListen
Warren Walker demonstrated iListen, a program that enables your computer to listen when you talk to it.
With iListen you can give commands to your Finder, called navigation. With iListen you can give commands to programs, and can buy script packages for each program. And, with iListen you can dictate into almost any word processing program. Warren bought it to see if it would work with Adobe Photoshop CS. He also brought a Wacom graphics tablet.
Midge Lubot took pictures and notes of the meeting.
1. iListen Demonstrated by Warren Walker
Warren demonstrated the latest version, iListen 1.6.8 which takes advantage of the latest features in Mac OS X Tiger, like Spotlight. iListen will not work on a G3 Macintosh, and it needs 500 mg on a G4. It will not work on OS 9, it will work on older versions of OS X, but works best on OS X 10.3 or 10.4.
Warren Walker at his computer.
Warren said the iListen tutorial has you set up the mike by speaking to it, You then train it by reading known text to it. Speech recognition doesn't work on some phrases. MacSpeech, Inc, suggests one lesson with a voice coach if you are planning to dictate a novel into iListen.
iListen claims 95% accuracy.
Warren switched to command mode and then said, open text edit
to open TextEdit, and then told it to go to sleep.
There is a list of commands and a list or programs in the tutorial.
Each program needs its own script package. For example, BB Edit has over 500 voice commands. There are three different wake up commands. Warren was using wake up.
Warren was getting a different reaction from the program because the background noises are different than in his home. Warren said there are recommended mikes for the program.
This program can be purchased with or without a mike. Warren recommends using a headset, and one can be purchased with the program.
You can create a manual setting, and you can have more than one voice profile.
Some dictation programs only work with a special word processing program, but iListen works with almost any text program on the Mac. The computer can't tell the difference between typing on the keyboard and speaking.
There are 20,000 words in the iListen dictionary and it will pick out the closest word. It has contextual intelligence, it looks at a phrase.
You can put it into correction mode. This can correct misunderstandings. You can show it what it did, and that it should update the word.
Via Voice is another program like this, and is cross platform, but it is only rated around 60% accuracy, much lower than iListen's around 90%.
Only iListen can transcribe digital dictation. Warren told iListen to scratch the last phrase it heard.
If a dialogue box comes up, you can say press & the name of the button.
There is also a spelling mode, you can say each letter and you must also speak the punctuation. The commands are based on Applescript. Under 10.4 you can use Automator. Photshop CS was the first application that was scriptable.
You can't type a shortcut into iListen.
Warren read a paragraph to the program, one of the words was doppleganger and it typed bottle vendor. Mellanin became militant.
He then demonstrated using the Graphics Tablet. The graphics tablet is very useful with Photoshop when selecting parts of the picture or using the pen for retouching, but it has handwriting recognition.
Veer was looking for something to use to record thoughts, it was suggested that she use a small digital recorder.
I wrote Warren to ask if it would work the way he wanted it to with Adobe Photoshop CS, and he sent this information:
MacSpeech is working on a ScriptPak for PhotoshopCS. I will probably buy it when it becomes available. iListen is usable in Photoshop for my work. Other people might find it less useful. I could use iListen to select tools and I could create commands for some menu items. I was unable to create commands to run Actions. That would very useful. I need to study Applescript, or perhaps upgrade to Mac OX 10.4 and use Automator.
The following ScriptPaks are currently available for download from the MacSpeech website, at macspeech.com: iLife 4, iLife 5, AccountEdge 2004, BBEdit, FileMaker Pro 6, Final Draft 7, GarageBand 2, iCal, iChat, iDVD 5, iMovie 3, iMovie HD, iPhoto 5, iTunes 6, iWeb, iWork 06, Keynote 3, Mail 2.0, Mariner Calc, Mariner MacJournal, Mariner Write, Microsoft Entourage 2004, Excel 2004, Office 2004, PowerPoint 2004, Microsoft Word 2004, Microsoft Word X, Nisus Writer Express, Pages 2, Preview, QuickBooks 2006, QuickBooks Pro 6, Quicken 2004, Quicken 2006, Safari, TaxCut 2004, The Print Shop, TurboTax 2005. They range in price from $10 to $30. Other ScriptPaks are being developed, including Adobe products.
I just visited the website for the following information:
Retail Price $99 without headset, $149.00 and up with headset/microphone. (I found it cheaper at PowerMax) Vocabulary of more than 330,000 words developed in cooperation with Oxford University. Support for Automator (Mac OS X version 10.4 required).
iListen is currently available in the following languages:
- US English (North American)
- UK English
- Australian/New Zealand English
- Italian
- German
- North American Spanish
- South American (Castilian) Spanish
- European Spanish
Can I use an iPod with iListen?
Unfortunately, no.
Older iPods (pre Video, or "5G), record at a sampling frequency of 8KHz (no matter what microphone is attached). Unfortunately, this does not provide enough audio information for speech recognition, which means older iPods cannot be used successfully to transcribe recorded speech to text using iListen.
The newer Video iPod does support higher sampling rates (22-44Khz), however, no microphones are currently available that support the new hardware in the iPod. MacSpeech is excited about the possibility of using these new iPods for dictation, but until we have tested them for this purpose, we cannot recommend their use for speech recognition.
Warren
For more information goto: http://www.macspeech.com
