Voice recognition software-- Does it work?

[Justin_Bailey]

@voguevixen

Not at all. I saw a demonstration of some VR software at EPCOT Center one time and thought “This is the way of the future!” I mean, supposedly it could even tell if you meant to say “which” or “witch” based on context, among other amazing things. I went home and bought (I want to say Dragon’s “Generally Speaking”?) I spent the 5 HOURS or so prepping it during installation by repeating phrases so it would recognize the timbre and inflection of my voice, or whatever. I was literally hoarse by the time I was done.

If I didn’t know you were a woman with a Justin Bieber obsession, I’d swear this post was written by my Uncle Stan. Needless to say, Dragon Naturally Speaking went into his Closet of Misfit Computer Products less than 24 hours after it was purchased.

[gaffa]

I get the impression that I’ll have to make a YouTube video of exactly how accurate Dragon NaturallySpeaking 11 is for me, in contrast to any number of posters here. If you’re using Windows 7 and a headset with an analog microphone, you really need to visit the Sound control panel, select the Recording tab, select the Microphone input, select Properties, select the Levels tab and adjust the Microphone Boost control. Just plugging a microphone in without making sure the quality of sound is optimized is going to deliver mediocre results.

The OP’s experience was about as far from optimum as you can get. Biggirl said:

I ask because I once saw a demonstration of a voice recognition program (don’t remember if it was Dragon) and that thing could not understand a word I said. I’m from The Bronx and I talk fast.

…which I take to me that there was NO training at all. Sorry, but that really isn’t fair to a very impressive piece of software.

Get a headset. An analog one if you can get the sound control panel set up correctly, a USB one if you’re not one to fiddle. Go through the setup of Dragon and when you get to the point were it is testing your voice take a moment and actually play back what it recorded. If it sounds even slightly noisy, go to the sound control panel and do the adjustments I mentioned above. I’m quite confident you can dramatically improve the quality of the microphone input to the system and therefore the accuracy of the program.

[Napier]

The Siri on my iPhone works fine. I ask it for directions, have it set reminders, send text messages to people, and so forth. These things frequently require names and other proper nouns that wouldn’t be in a dictionary. There is also a dictation feature available when writing notes, emails and the like, which I understand uses the same voice recognition facility.

In fact, I have not yet had the thing misinterpret a single thing I’ve said. I bet I’ve put a thousand words into this. It’s pretty amazing. I believe I speak pretty clearly, though I have a slight lisp that emphasizes the hard “s” sound - maybe that’s helping?

[gaffa]

@Napier

The Siri on my iPhone works fine. I ask it for directions, have it set reminders, send text messages to people, and so forth. These things frequently require names and other proper nouns that wouldn’t be in a dictionary. There is also a dictation feature available when writing notes, emails and the like, which I understand uses the same voice recognition facility.

In fact, I have not yet had the thing misinterpret a single thing I’ve said. I bet I’ve put a thousand words into this. It’s pretty amazing. I believe I speak pretty clearly, though I have a slight lisp that emphasizes the hard “s” sound - maybe that’s helping?

As mentioned earlier, Siri appears to be Dragon’s technology. The difference is that the microphone and audio system is known - everything is running on the exact same hardware. I suspect you’ll get more variability of results if you use it with a Bluetooth headset.

[jdc]

Dragon has 30 day return policy, so no harm in trying.
Headset mike, properly adjusted, and quiet area do make a big difference.

Decent headset comes with the product. Fanatics tend to upgrade.

If you can’t stand the headset, a really good noise cancelling mike can do almost as well as headset mike.

Initial training for Dragon is pretty short, but optimizing accuracy requires some additional training as errors occur. You also run it against a bunch of your emails and documents and it improves its recognition based on your style and vocabulary.

[KAndre]

For me, even the basic Speech Recognition in Windows works fairly well (I have 7 Professional, not sure if it’s in other versions of Windows) - although best with a headset.

[elfkin477]

I’m using version 11.5 right now as I compose this post. The biggest problem I find with using it is that their idea of a normal speaking speed seems to be slower than people around here speak normally. So, I have to speak a little slower in order for it to be accurate. I’m also fairly well convinced that whoever programs it is in the south, considering some of the very strange interpretations of what I say that only would make sense in a particularly tortured southern accent (I say the name “Ren” it says “read.” I say “Todd” it says “taught” etc). Overall, though, I think it does a pretty decent job.

[Karen_Lingel]

I once asked my Android phone to “Call Frances” and it showed me a map of Iraq. It asked first. It said: Did you mean “map of Iraq”? And while I was sitting there stunned, it decided no response means Yes, and there it was: a Map of Iraq.

[thelabdude]

I haven’t tried to call SS or Medicare for a year or 2. No punishment is sufficient for the over paid idiot that approved their phone system. It was terrible.