« Echoes of the Doomed Franklin Expedition | Main | »
20 juni 2006
Voice Recognition Software
YPSIDIXIT suggested the possibility of using VRS to her boss today, and it turns out this suggestion has raised all sorts of interesting issues about the process of writing and the difference between writing and speech.
VRS transforms spoken words into computer text on a screen. It also responds to verbal commands such as "delete." It is widely used for medical and legal transcriptions and for real-time TV news banners.
Y.'s job consists of writing, digging around in computer archives, and research, in roughly equal measure. Though I'd thought VRS would streamline and quicken the writing process, once the technology is mastered, it was pointed out to me that it would only externalize the mental process of translating thought into the sort of writing that would satisfy a reader's expectations.
However, speech is much more verbose, repetitive, and redundant than writing. VRS would serve only to throw the whole glob of mentally undigested word-goo onto the screen, where it might be even more cumbersome to handle. On the other hand, Y. is a highly visual learner, so it might be easier to actually see the word-goo instead of trying to mold it in my mind.
A parallel between speech and typical blog-style writing was drawn--and it was pointed out that most blog-style writing would never be good enough to publish, however enjoyable it might be.
Y. needs to do a whole lot more research before committing someone else's money to this project. She'd also like to ask kind readers if this game is worth the candle, in their valued opinions. Thanks.
Posted by ypsidixit at 20 juni 2006 12:03
Comments
A friend of mine did her masters thesis on voice recognition software. She once told me that some people talk like they write. She said that those people often seem boring in real life but write interesting things and the VRS works *great* for them. She also mentioned that a lot of people who seem interesting in real life pepper their speech with so much extra stuff that the VRS is nearly useless for them.
Posted by: lynne at 20 juni 2006 17:03
Back around 1999, Dragon Naturally Speaking was state of the art for natural speed recognition, and once it was trained up on my voice, I used it with great success for transcription of recorded conversations of others (ear bud in one ear, regurgitating verbatim what I heard into the microphone).
That's a bit different from using VRS in (i) the process of creatively forming new text or (ii) using voice to interact with the machine for everything (e.g., "Launch Firefox. Ask dot com search. Search-text quote Keep Ypsi Rolling quote text.")
For (ii) it's not efficiency that's gained, but rather the avoidance of keystrokes or mousing in general in light of repetitive, overuse injuries. If you're looking to build a case justifying the experiment at least, that's a fair point to make.
For (i) I think that the limiting reagent on my 'writing' speed is not my typing, but rather my brain. Even if I could make the alternatives I'm mulling over appear instantly on screen just by thinking them, I think I'd write just as slow. Instead of just never making it onto the screen at all, rejected bits of text would have to be deleted. As it is, there are many potential bits of text that get considered and rejected before the first word of that text-bit is even typed.
Do you work within easy earshot of colleagues? It would drive me nuts to hear someone using VRS all day long. Also, what's the ambient noise like where you work? State of the art 1999 showed somewhat degraded accuracy with the introduction of background chatter, fans, keyboard clicking from the next desk over, stomach growling, etc. Could be better now.
If you decide to try it out, please post about how it's going and give brand names. When I do a hardware upgrade (soon?), I'll be in the market for something to help with my transcription load.
Posted by: HD at 20 juni 2006 19:36
Lynne: I think that's a shrewd observation.
I think I fall into the first camp, not because I necessarily write anything interesting, but because I think I talk like I write and I'm not a garrulous speaker.
It could be I talk like I write because I write 8 hours a day (and then more in the evening on the blog) so that is my dominant way I communicate to the world. It's certainly my preference.
Posted by: laura at 20 juni 2006 19:39
If it's your preference, why would you want to use the speaking format? (Just a little confused here.)
Posted by: addiann at 20 juni 2006 19:51
HD: That is an informative and nuanced reply as usual. Thank you for the info.
Just out of curiosity, do you use VRS for your site? I'm not naming it on purpose, since you used only your initials and so may wish for some degre of anonymity.
In my case I'd be creating new text. Your point about the limiting factor being one's brain is a good one. Can I think faster than I type? Hmm. Doubtful. :)
Posted by: Laura at 20 juni 2006 21:03
Addiann: It is faster. I talk way faster than I can type, though I'm not a bad typist.
Posted by: Laura at 20 juni 2006 21:04
Laura wrote: "Just out of curiosity, do you use VRS for your site?"
Nope. When I upgrade to a machine with enough horsepower to run it, that will be one of the first pieces of software I install. Not that VSR requires so very much horsepower, just that my current machine is very underpowered.
Posted by: HD at 20 juni 2006 21:41
Mine too, an old iMac. Was just curious; thanks.
Posted by: Laura at 20 juni 2006 21:45
When Cara was in her phd program, a friend of ours, another phd student, had fairly debilitating repetitive stress injuries - a pretty bad handicap for a grad student. He had developed an almost completely hands-off computing style.
For mousing, he had Play-School or Fisher Price or some such brightly colored, over-sized trackballs, which he placed on the floor and used with his stocking feet. (He had bought several, in case they ever went out of production, so that he wouldn't have to worry so much about wearing them out.)
For writing, and for menu manipulation and so forth, he used...whatever Dragon Naturally Speaking was called circa 2002-2003. He was extremely well-versed in VRS technology and history, as one might expect, and I think I remember him discussing the corporate buy-out chain behind Dragon, but I don't remember the content.
Obviously, he didn't do much recreational typing (though he was very interested in my experience with blogging), but he used this setup for note-taking (I think he recorded and transcribed lectures and conversations, using something like HD's in-one-ear-and-out-the-other method), writing papers, and, of course, disserting. It worked pretty well for him.
I feel like I write like I speak like I write, and I think most people really do, as noted - so I don't think that's a huge problem. If you are accustomed to filtering or structuring your thoughts a little more before writing than before speaking, then it might take some practice, but you're for the most part transferring processes from your fingers to your mouth. (Me, I feel like I can queue up thoughts better through my fingers, but I think that's a practice issue.)
Posted by: Murph at 20 juni 2006 21:54
The sensual Taurus moon welcomes the summer solstice, a magical day to celebrate the light, heat and life brought to us courtesy of King Sun! So get out and soak up some rays. If you can't take the day off, then catch a few stolen moments -- maybe on your commute with your arm out the car window or during a solar break at lunchtime.
ARIES (March 21-April 19). There is a Japanese proverb that luck is like having a rice dumpling fly into your mouth. Today, you're caught off guard by the sort of surprising luck that is a lot to swallow!
TAURUS (April 20-May 20). There's enough to go around. So when you feel like fighting for your piece, it's time to slip into the mindset of endless abundance. Look into the depth of the sky, imagine the breadth of the ocean, try to count the stars.
GEMINI (May 21-June 21). Getting along with loved ones isn't hard, or easy, either. You can say almost anything if it's true -- the way you say it makes all the difference. You weren't born with all that charm for nothing. Use it!
CANCER (June 22-July 22). You believe those close to you can do better, and you're probably right. You don't mind being an agitator. After all, it's the grain of sand that makes the oyster form a pearl.
LEO (July 23-Aug. 22). Love is not about "what have you done for me lately?" This you may have to remind yourself over and over throughout the day as you rise above the circumstances surrounding you.
VIRGO (Aug. 23-Sept. 22). Groups of people who all seem to know one another can be intimidating. Even if you are "the outsider, " you're also exactly the fresh breath that a project requires to make it to the finish line.
LIBRA (Sept. 23-Oct. 23). It's not enough to simply assert your independence -- celebrate your independence! You're not controlled by anyone! When you change your role in life, loved ones will adjust to the new you.
SCORPIO (Oct. 24-Nov. 21). You've carried someone through the trials of life, and for that, you're rewarded, albeit in small, glad doses. When the success of those you teach is your greatest achievement, you're truly a leader.
SAGITTARIUS (Nov. 22-Dec. 21). Equal affection is rare. There is always someone who loves a little more and does a little more for the other. Usually, the romance pendulum swings back and forth. Today brings such a shift.
CAPRICORN (Dec. 22-Jan. 19). You can vividly imagine how you would like a loved one to behave, and yet, you won't tell this person what you want and expect. Well, maybe he or she will receive your mental message. Night reveals much.
AQUARIUS (Jan. 20-Feb. 18). Projects go well when you stop worrying about whether you're being "original" enough. You ARE an original. You can't help but be that -- there is no other you. Your only job is to be authentic to the YOU you are.
PISCES (Feb. 19-March 20). The ones you love make you laugh -- heartily and often. This is reason enough to keep them around. Be as thoughtful, helpful and supportive as you can be.
Posted by: Anonymous at 21 juni 2006 08:49
Thank you, dearest Oracle. I will try to rise above the circumstances surrounding me. Maybe on the elevator.
Posted by: Laura at 21 juni 2006 10:16
I tried this software a few years ago, and found it a mixed bag. It did take a while to train, which I was willing to do, but I also found it an odd way to write. I can think and type at the same time, but found it harder to think about how to speak things the way I wanted them *written* (as opposed to how I speak). So, I abandoned the process.
There are people out there using it successfully, but I guess it's not for me.
Posted by: Brian Filipiak at 21 juni 2006 10:33
Brian: It has been pointed out to me that in a way it adds an extra "layer" to the process of writing, the one you describe--figuring how to speak the way one writes, since those are different processes.
Posted by: Laura at 21 juni 2006 10:35
iapeb fhcg wcdbzuay ocihq kzhx jxpw tpslvnakc
Posted by: rnzqacts asmepk at 12 januari 2007 01:30
iapeb fhcg wcdbzuay ocihq kzhx jxpw tpslvnakc
Posted by: rnzqacts asmepk at 12 januari 2007 01:31