[Date Prev][Date Next][Subject Prev][Subject Next][ Date Index][ Subject Index]

Re: Dragon NaturallySpeaking v10 W2K Xywrite



Flash,
1. You say it uses as much RAM as there is, then you talk about using it with different amounts of RAM--does that mean you tried it with different *physical* amounts of RAM? Or is there some setting?
2. I haven't found NatSpeak to work well with Xy, in that it doesn't
recognize commands. Ideally, you'd want it to have voice-activated versions
of BC ,XD ,DF etc. But lacking that, I can't go to the command line, do a
correct selection, set formats, etc. I'm betting there is some place in
NatSpeak that you can define the byte-level output of a spoken word, and
via that we could make NatSpeak Xy-friendly. As it is, I use NatSpeak in
Windows programs, such as Eudora and Word. But, I blush to admit, I'm not
using NatSpeak much at all, or not yet.
Y'all,
Here's an update on using speech-to-text with our favorite word processing program. Dragon uses up as much RAM as there is, and they recommend a minimum of one GB. I first tried using it with 750 MB, and found that it ran slowly. It is now running more happily with two GB. It will merrily dictate text into XyWrite, although some supplementary features are not available in Dragon to XyWrite; for example some voice commands. It works well with T'bird e-mail program, and can even transfer text to programs such as Adobe Illustrator and Photoshop by means of a separate transfer pad. The headset which delivered with the standard version is pretty uncomfortable but the program will work with almost any headset or halfway decent microphone. If you don't need to hear playback then of course a microphone without earphones will be adequate. The program is trainable, and you can manually edit the dictionary. Getting it to work is straightforward: you start the program, activate the microphone, and then start any other program into which you would like to dictate text: simply set the cursor where you want the text to begin, and start talking. The text just flows right into your destination program, be it XyWrite or e-mail or whatever. You can make corrections on the fly or later on, as you wish. This e-mail for example was entirely dictated using Dragon. It does however make a few mistakes. For example, when I say obscured it writes up skewered, or at any rate it did until I trained it assiduously. Yes, it recognized the final word in the previous sentence; I did not have to correct that. It confuses homonyms: profit and prophet, to, two and too, etc. (this sentence I had to correct manually) [Note to Auerbach: now here's a good example of a self-referential sentence]. I will find it very useful for translating from German to English on-the-fly, and for quoting long passages of text (my flatbed scanner does not have an OCR function). I have not yet figured out how to format spoken text bold or italic in XyWrite, but maybe Harry has figured this out for us already.

Cheers,


Harry Binswanger
hb@xxxxxxxx