I've tried using the Speech Recognition, and have had little success. Then again, I have a desk microphone, and it's not really the best for that sort of work.
Looked useful to me. It understood when he was telling it to delete long strings of words, so I can expect to write an entire document by talking into it. I don't need to code. =P
Speech recognition is not robust. It does not handle dirty signals without being trained with them.
Speech recognition is not for coding. It is not good at recognising symbols and precise formatting.
Speech recognition IS incredibly good at interpreting large amounts of continuous spoken word from a clean enough source once it has been trained to the speaker. People are idiots and speak one word at a time rather than following the fscking instructions and speaking normally. A major part of increasing the accuracy of speech recognition comes from context sensing. Something that's very difficult to do if it's reading in a single word at a time or if some idiot is trying to use it to write code.
People also have retarded accents and can't speak plain english. "Aunt" is not pronounced "Ant", for starters. It's closer to "aren't".
<!--quoteo(post=1639563:date=Jul 18 2007, 11:40 PM:name=Haze)--><div class='quotetop'>QUOTE(Haze @ Jul 18 2007, 11:40 PM) [snapback]1639563[/snapback]</div><div class='quotemain'><!--quotec--> Looked useful to me. It understood when he was telling it to delete long strings of words, so I can expect to write an entire document by talking into it. I don't need to code. =P <!--QuoteEnd--></div><!--QuoteEEnd--> figuring out what to delete is a lot easier than figuring what to type. you can safely assume that the user wants to delete the most recent entry, after that all you have to do is count syllables and (maybe) double check the starting point.
I wonder if anyone here will have heard of OS/2 Warp. It came out in 1994 and it had speech recognition; sure you had to train it over a period of 3 hours, but it did have speech recognition. It has only taken Microsoft 13 years to finally catch up. IBM has only recently discontinued support of an OS they no longer sell. Oh well, just another sad tale of a good operating system being tossed away in the wake of Windows.
PulseTo create, to create and escape.Join Date: 2002-08-29Member: 1248Members, Constellation
<!--quoteo(post=1639653:date=Jul 19 2007, 09:48 AM:name=Gree)--><div class='quotetop'>QUOTE(Gree @ Jul 19 2007, 09:48 AM) [snapback]1639653[/snapback]</div><div class='quotemain'><!--quotec--> I wonder if anyone here will have heard of OS/2 Warp. It came out in 1994 and it had speech recognition; sure you had to train it over a period of 3 hours, but it did have speech recognition. It has only taken Microsoft 13 years to finally catch up. IBM has only recently discontinued support of an OS they no longer sell. Oh well, just another sad tale of a good operating system being tossed away in the wake of Windows. <!--QuoteEnd--></div><!--QuoteEEnd--> What are you trying to say here? Speech recognition has been a feature of Windows and other operating systems for a <i>long</i> time. What's special about Vista's speech recognition is that it's better than what we're used to. Albeit not by much, as shown by the video.
Comments
[youtube]kJ861ehHwWQ[/youtube]
Speech recognition is not for coding. It is not good at recognising symbols and precise formatting.
Speech recognition IS incredibly good at interpreting large amounts of continuous spoken word from a clean enough source once it has been trained to the speaker. People are idiots and speak one word at a time rather than following the fscking instructions and speaking normally. A major part of increasing the accuracy of speech recognition comes from context sensing. Something that's very difficult to do if it's reading in a single word at a time or if some idiot is trying to use it to write code.
People also have retarded accents and can't speak plain english. "Aunt" is not pronounced "Ant", for starters. It's closer to "aren't".
--Scythe--
Looked useful to me. It understood when he was telling it to delete long strings of words, so I can expect to write an entire document by talking into it. I don't need to code. =P
<!--QuoteEnd--></div><!--QuoteEEnd-->
figuring out what to delete is a lot easier than figuring what to type. you can safely assume that the user wants to delete the most recent entry, after that all you have to do is count syllables and (maybe) double check the starting point.
I wonder if anyone here will have heard of OS/2 Warp. It came out in 1994 and it had speech recognition; sure you had to train it over a period of 3 hours, but it did have speech recognition. It has only taken Microsoft 13 years to finally catch up. IBM has only recently discontinued support of an OS they no longer sell. Oh well, just another sad tale of a good operating system being tossed away in the wake of Windows.
<!--QuoteEnd--></div><!--QuoteEEnd-->
What are you trying to say here? Speech recognition has been a feature of Windows and other operating systems for a <i>long</i> time. What's special about Vista's speech recognition is that it's better than what we're used to. Albeit not by much, as shown by the video.
can someone point me to the man or men responsible for the youtube seek-bar programming so that I may send them a "gift basket"?
oh man did you catch that zinger in the second video? "Let's not forget the first part of 'Recognition' is 'Rec' as in 'Train Wreck'"