Reply to thread

Message

<blockquote data-quote="Umbran" data-source="post: 5175481" data-attributes="member: 177">Well, if typing is out, there are two basic technologies that can assist you:&nbsp; webcams and speech-to-text software.&nbsp; Dragon Naturally Speaking is reportedly very good speech to text - kind of the industry standard.&nbsp; I may be incorrect, but I don't think you can shift the burden to her end - all of you talking, and her hooking up an instance of DNS to turn it to text on her end.&nbsp; I don't believe the software accepts arbitrary audio streams as input (though I could be wrong there), and I even more doubt that it would be able to differentiate speakers - you'd all be one &quot;person&quot; to it, so she couldn't tell who was saying what, in general, and the software's ability to &quot;train&quot; to recognize your personal speech patterns will fail if you have several people talking to it at once, limiting its accuracy.The other option would be for each player to have DNS, and feed its results into a chat program for her.&nbsp; But that'll cost you about $100 per person for the software.I don't know the basis of your personal decision to not use webcams, but they are pretty darned cheap - probably cheaper than the speech-to-text option.</blockquote>

[QUOTE="Umbran, post: 5175481, member: 177"] Well, if typing is out, there are two basic technologies that can assist you: webcams and speech-to-text software. Dragon Naturally Speaking is reportedly very good speech to text - kind of the industry standard. I may be incorrect, but I don't think you can shift the burden to her end - all of you talking, and her hooking up an instance of DNS to turn it to text on her end. I don't believe the software accepts arbitrary audio streams as input (though I could be wrong there), and I even more doubt that it would be able to differentiate speakers - you'd all be one "person" to it, so she couldn't tell who was saying what, in general, and the software's ability to "train" to recognize your personal speech patterns will fail if you have several people talking to it at once, limiting its accuracy. The other option would be for each player to have DNS, and feed its results into a chat program for her. But that'll cost you about $100 per person for the software. I don't know the basis of your personal decision to not use webcams, but they are pretty darned cheap - probably cheaper than the speech-to-text option. [/QUOTE]

Verification