[X4U] OS X Speech capabilities question

[A. Scott McCulloch]

There are a couple (at least) of different "types" of speech  
recognition... what is built-in to OS X is capable of "recognizing"  
commands (this is similar technology to voice based automated  
telephone systems), but is not capable of "recognizing" the spoken  
word to turn it into text (which requires a very different  
technology). For transcription, you need something like ViaVoice or  
iListen.

You can find iListen here: http://www.macspeech.com -- I don't know  
the current status of ViaVoice, it was done by IBM, but I think they  
may have abandoned it.

However, as I already mentioned, these programs have to learn the  
individual's patterns of speech, intonation, etc. to get accurate  
results. If you've got a good quality audio recording, you *might* be  
able to get a reasonable first take at a transcription, but expect  
*lots* of errors since there will be voices on there that the program  
has not been trained with.

For my time and money, sending my dissertation research interviews to  
a transcriptionist was the only way to go.


Scott


On Oct 7, 2005, at 3:40 PM, Matt Gregory wrote:

> Something that would autogenerate at least a first cut at text  
> would have been like a godsend to her.  Not giving up on that  
> though.  It may not be built into OS X (I'm on 10.4.2 btw) out of  
> the box, but it's obviously able to recognize speech ... so is it  
> hackable, or AppleScriptable?