Speech to text Recognition Quake

[Team Xlink]

Has anyone ever attempted a speech to text recognition implementation in Quake?

It would make communicating with people easier, especially when your on ventrillo with part of your team, whatever you say would get sent into text as well.

[dreadlorde]

Seems redundant. On a sort of side note, are there any actual teams/clans still playing quake (not nexuiz, AA, OA, &c).

[Spirit]

Sure, this year's EQL has 38 teams of >4 players each. http://eql.quakeworld.nu/eql12/

[qbism]

Step 1 would be to get Ventrilo audio into Quake. Is there ventrilo API + DLL? If this hurdle was overcome, there'd be no need for SST because players could listen in?

Maybe it would be easier to build ventrilo-like function into Quake if there's a GPL library for this. Also guessing it requires a separate voice server either built-in or out on the net somewhere.

[Spike]

voice chat > speech-to-text.
if only because it means you don't have to look away from the middle of your screen - big monitors suck!

libspeex and libspeexdsp provide all that is required to de-noise and compress voice chat.
Then all you have to do is capture audio, send it somewhere, decide who should hear it, multicast it, receive it, stream it.
Easy...

The problem with ventrilo/teamspeak is that its not automagic, you have to have a server somewhere, you have to be willing to give out details publically. etc. Its not 'join server, oh hello there!'. And if you're playing with friends then you can just use a separate app instead.
I mean... teamspeak uses libspeex by default anyway.

[qbism]

Going the other way with it, TTS with a compact system like eSpeak would be interesting. It would speak chat messages of course, plus high-priority console messages.

[Ranger366]

Voice Chat in Quake would be awesome, Half-Life had it too in 1998.
Never thought of this.

[leileilol]

Half-Life had it too in 1998.

wrong
Half-Life added it in mid-2001, and depends on a miles sound system for it