Disguised voice chat

Love SL. The one thing which really gets to me is the time it can take to "talk" through a keyboard. Simply speaking is much faster; I could type for an hour or speak the same content in ten minutes. Skype-like functionality could revolutionise SL.

Obviously, anonymity has to be part of the SL experience, but that's possible. If you have a Mac, try Garageband -- it allows live voice changing effects to be applied, including gender changing and other modfications.

It probably wouldn't be feasible everywhere, at least at first, but it would make a big difference in the long run. It could also bring a great many new residents -- why just call someone when you can meet them in SL for a virtual drink instead?

Thoughts?

Hiya, here's an interesting article on the topic:

http://www.gamegirladvance.com/archives/2003/07/28/not_yet_you_fools.html

From what I've seen, the technology is not yet up to the task of making a changed voice sound natural, and I'm glad. And if you've got some people using voice changers and some not, it's kinda obvious.

On the suspention of disbelief:
http://en.wikipedia.org/wiki/Uncanny_Valley

Adding voice to a virtual world puts it into the uncanny valley.
(Also, the XBox 360 and the PS3's graphic qualities are dead center in the uncanny valley--take a look at their demo for the new basketball game and note the look of stupidity on the character's face)

Very interesting article that brings up several valid points. However, it's somewhat dated, and the many, many comments tend to politely disagree. I do too.

Agreed that voice changing could expose some identities; if that's a concern, make it mandatory. Even put everyone through a vocoder to make us all sound "robotic" and provide a text-to-speech engine for those who simply don't wish to talk. For an example of currently released technology, use Garageband's voice changing function to change Trent Reznor into a girl (he sound like a "punk rock chick" with the Garageband-formatted version of "The Hand That Feeds".

http://nin.com/downloads/index.html

There's a valid point in the article about accents being undisguisable, but to some degree that's true even of text. Different nations have different colloquialisms, and when asked, I'm yet to find someone who wasn't happy to say at least generally where they were from. Besides, if we're typing English, we're capable of speaking it, at least to some degree. Those who can't type English quickly are already somewhat excluded; voice would, if anything, improve this.

I'm not sure I agree with the uncanny valley problem in relation to SL -- yet. Final Fantasy tried too hard to be real; The Incredibles realised that would be suicide and used caricature instead. SL avatars have their own "look" which won't change for some time, and I don't think it's real enough to cause a problem. We're nowhere near the Final Fantasy level of realism yet, though admittedly we could be nearer with carefully built skins, animations and great big pipes for bandwidth. SL wants to be as close to hyperreality as it can get, ASAP, and the uncanny valley will be crossed sooner or later. Voice could take us a step closer, but that's inevitable.

Voice removes an important barrier between SL and reality. It would allow us to talk and emote at the same time or even talk while taking a stroll. I've experienced voice in SL, mostly through live DJs. At first it's odd -- "That's what he really sounds like??" -- but almost immediately you realise there's a whole extra sense we could be communicating with.

Disclosure: I don't game much, and haven't experienced any other games with voice chat. The comments on the article linked above were generally positive about voice chat. I have used videoconferencing and received a Skype call, and I know it's not perfect.

OK, so this is an old topic, but it's the most recent (i think) that mentions text-to-speech.

I really like SL, but was amazed that it didn't already have Text-To-Speech as standard

I'm used to playing UT2004, and with the right voice packs client side, the Text-to-speech voices are very realistic, i even have a pack with a british accent (i'm from the UK)

I would like to vote on having this feture implimented, it would all be client side so no extra lag, Windows XP has the ability already and that is what UT uses.

Even with 30 ppl on a server in UT2004 the text is queued and spoken quickly so i do not see a problem with having lots of ppl speaking at once

Anyway, i can't find a proposal for this so i may just make one

The voice changing software that I have heard about works, to my knowledge, by directly changing various acoustic properties of the voice sounds received as input.

If voice recognition software improves to where it works real accurately for almost everyone, one could do voice changing through a different process: convert the speakers words to text, at least internally, and convert the text to voice in a normal text to speech manner, like the program TextAloud does, using good quality voices like the NeoSpeech Kate voice.

The result wouldn't sound real, but who cares?

Streaming all that sound about would be murder on bandwidth unless it was done p2p, but then it would lag clients instead.

A text-to-speech I like, on OS X there are APIs precisely for doing this, can even go a step further and have dialogue boxes read out to improve accessibility.
Speech-to-text I believe is also an OS X API, would be a cool time saver as well, just press a key, speak, and the jist of it will be put into the chat box. Okay, so it'll be a bit erroneous but it'd save a lot of time.

proposal made

http://secondlife.com/vote/vote.php?get_id=2326

(I don't know if I'm allowed to cross post like this but <shrugs>

I've done it plenty of times. Audio streaming does not cause lag, even when done simultaneously with SL.

There are, however, many excellent reasons not to have audio as an integral part of SL, mostly to do with mismatch between the player and the character voices, and the fact that a trained ear can spot synthesized/modulated voices.