How feasible would it be to analyse sound or voice in DM? I think that voice recognition, even at a very basic level, would add a new depth to BYOND games. So, any thoughts?
The technology for voice recognition is only just beginning to climb out of the dark ages. Its history in gaming even fairly recently has been spotty. I don't see it adding depth to many games because it hasn't even been used that extensively in console and PC gaming--it's really suited for very limited roles.
BYOND's DLL support I think won't really allow for this because it'd need to be able to handle constant input and take it at the client end, not just the server end. It would have to be implemented natively. That falls somewhat under the banner of enabling voice chat, a subject that has come up before. It might fall out of something like that, but voice chat does raise concerns over practicality and bandwidth, particularly in large servers. Also I'm not sure there's mass appeal in hearing a bunch of 13-year-olds bandying foul language for the sheer novelty of it.
So voice recognition is doable if we built in some sort of native voice chat system that would carry the commands to the server, or did some special processing on the client end to convert to commands, but ultimately I don't think it'll do much good or work that well. The technology has not yet reached its prime and in gaming its practical uses appear to be fairly limited.
All right, I tried to mould the subject towards gaming as I thought that would capture more interest on these forums than anything else. My real interest in it would be for voice commanded actions on your own computer. The phrase "Open Firefox" surely sounds so very different to "Shutdown" that even a basic analysis could differentiate between the two? In this case, also, bandwidth would cease to be a problem as such a program could be client-side only.
I have some voice recognition software installed that types out what I say, but doesn't have the capability to take commands or perform actions, besides the four built-in ones; "Listen", "Stop listening", "Log off" and "Shutdown" - so would it be possible to integrate such a program with a home-built BYOND program? Would it be relatively easy?
All right, I tried to mould the subject towards gaming as I thought that would capture more interest on these forums than anything else. My real interest in it would be for voice commanded actions on your own computer. The phrase "Open Firefox" surely sounds so very different to "Shutdown" that even a basic analysis could differentiate between the two? In this case, also, bandwidth would cease to be a problem as such a program could be client-side only.
Yes, outside of gaming I think voice recognition has the biggest potential. Voice typing software is already common on the market of course, though it requires a lot of training to recognize your words right and even then it isn't perfect. For desktop navigation tasks it'd be relatively interesting, but I think it holds more promise as an accessibility feature than a time-saver.
Of course when I think of using voice recognition software to increase productivity, I think of talking to my computer in natural language. Natural language processing is related to contextual thinking, which makes me think that a more intuitive interface between man and computer probably starts at the level of trying to figure out what you want based on minimal input. "Play a happy song" is a simple command, but to parse it the computer needs to know where you keep your music library and it should have some idea what you think of each song in it, mood-wise.
For my money I think voice recognition will really reach its fruition when it can be combined with a relatively sophisticated ability to parse natural speech into a command. There seems to be a great deal of room for improvement on both fronts.
I have some voice recognition software installed that types out what I say, but doesn't have the capability to take commands or perform actions, besides the four built-in ones; "Listen", "Stop listening", "Log off" and "Shutdown" - so would it be possible to integrate such a program with a home-built BYOND program? Would it be relatively easy?
With your program, probably not easy--not unless you built a custom app to take typed-in text and fire off events like sending messages to windows and opening programs and such. In that case though you'd have a focus issue, because that app would probably need to be in the foreground.
A different speech recognition app, one designed to interact with windows in a more complex way than just taking dictation, would probably be much easier to integrate.
BYOND's DLL support I think won't really allow for this because it'd need to be able to handle constant input and take it at the client end, not just the server end. It would have to be implemented natively. That falls somewhat under the banner of enabling voice chat, a subject that has come up before. It might fall out of something like that, but voice chat does raise concerns over practicality and bandwidth, particularly in large servers. Also I'm not sure there's mass appeal in hearing a bunch of 13-year-olds bandying foul language for the sheer novelty of it.
So voice recognition is doable if we built in some sort of native voice chat system that would carry the commands to the server, or did some special processing on the client end to convert to commands, but ultimately I don't think it'll do much good or work that well. The technology has not yet reached its prime and in gaming its practical uses appear to be fairly limited.
Lummox JR