words from nowhere??!!

[One]

when the microphone is on their are words like "and" "be" "what", being added to the beginning and end of a sentence,, or even on the entry field by themselves. I am not breathing into the mic . it is of good quality (i have tested a few with same results) I am not getting feedback from hal ,,, Even when it is dead silent it happens does anyone have any idea what is going on??
TIA![]

[onthecuttingedge2005]

quote:Originally posted by One

when the microphone is on their are words like "and" "be" "what", being added to the beginning and end of a sentence,, or even on the entry field by themselves. I am not breathing into the mic . it is of good quality (i have tested a few with same results) I am not getting feedback from hal ,,, Even when it is dead silent it happens does anyone have any idea what is going on??
TIA![]

I have an idea.

set HAL's microphone in front of a bugs bunny cartoon T.V program and watch HAL's responses.

every child needs bugs bunny.
Jerry[8D]

[One]

Gee, thanks.

There is enough damagedone by permitting the link between Hal and Answerpad. [B)]

[Bill819]

When the microphone is on even the littlist sounds can cause Hal to reapond. My airconditioner although very quite makes a real small hissing noise and Hal can hear it and will make so dumb statements because he did not understand what he/she/it heard.
Bill

[One]

yea,, I know. Thanks

[Carl2]

I am now using Dragon version 10 and in the past I had used Dragon version 7, Dragon version 9 with Hal.  I can remember Dragon adding words for what seemed no reason, ages ago I had built and audio filter for use with Dragon which improved the signal-to-noise ratio, just about all the microphones I had tried would give a signal-to-noise ratio of approximately 19 DB, using the audio filter I was able to increase this to 27 DB.  Dragon has also improved their speech recognition program to reject unwanted noise, coughing into the microphone would yield a series of question marks without putting words into the dialog box.
  One of the main reasons I purchased this version of Dragon is because it is available with the Bluetooth headset, at this time I am unable to use the audio filter but the signal-to-noise ratio is 23 DB, much more comfortable and convenient than the normal headsets.
 Carl2

[Carl2]

One
  I really meant to mention that earlier versions of Dragon were much more susceptible to adding noise to the text, later versions of Dragon had less of a tendency to add a series of words to the text being generated.  I can remember coughing into the microphone and seeing a series of question marks being generated in the results box but nothing was added to the text box.
   I had tried using the voice recognition program which comes with Vista and found it to be what I would consider almost useless especially compared to Dragon.  At this time I still have not been able to connect the audio filter to the Bluetooth  microphone and at this time I do not recall seeing any noise being added into the text.
  Before ending this post I would like to mention that microphones using the USB input are said to be less likely to add noise.   I've just tried coughing and tapping on the microphone without anything being added.
Carl2

[One]

[8]

[Carl2]

One,
  Glad to see you have fixed the problem, I think it is much easier to speak with Hal using speach reconition.  I'm curious as to what type of speach reconition you are using.
Carl2

[One]

[8]

[Art]

From what I've read, Vista has a built in speech recognition engine that works very well even with little training, right from the start.

Microsoft has been working on MANY platforms for speech including Vista desktop, mobile computing and their Speech SDK that offers developers of new applications a powerful and flexible method to incorporate speech in their applications.

There is even the One Voice Communicator for Vista that one can use their voice to control Windows Media Center to play music, slide shows, check email and even order pizza.

"I guess from what I can tell Micrisoft has'nt done much in the way of speech the last few years...."


I don't think this is the case.[8D]

[tedathome]

Just a comment, sorry, I had to, for me. The very best speech recognition I have found is a cheap mike with vista speech recognition. I overdrive it (vu meter in the red) and this works better for me than the mike I bought at Zabaware and following directions to the T. Why? I don't know, maybe only real words can get through this way, I don't know. I just like to experiment with electronics and sometimes it just works. Being a sound reinforcement enginier, I like to experiment in what they call "max headroom" which is the loudest you can get without feedback or distortion. I know I'm getting a little off topic but I'd like to say something about my stereo while talking about sound. I have a 250 watt per channel Marantz. It took me 6 mo's to wire the crossover network to my satisfaction. Four speaker colums, two per side. I can turn it up to 4 and the neighbors house can hear it loud and clear. The point I want to make is you can still talk to someone in the same room in a whisper and still be heard. In other words, experiment!
 I'm getting better than 98% acuracy and hal with learning level at 0 actualy learns this way and understands that not all that is heard is what was said and seems to compensate for it. OK, there I said my piece and I know I't doesn't match with what is ment to be the way to do it, but I just had to speak my piece on this one. If it works, then Murphy's law applys![]

[One]

[8]

[Carl2]

Vista is a bit of a memory hog, I use about 1 G just to run Vista. I don't think I'd try it in an older computer. Not impressed with MS speach rec especially when compared to dragon. Just saw Dragon Perferred ver 9 being given away for completing a program, works much better with the 9.5 update.
Carl2

[freddy888]

Just thought I would throw in that Dragon version 9 can be found a lot cheaper on eBay now too..