Today, for the first time, I tested the Voice Beta client of Second Life. In the picture above, you can see a number of avatars with a white dot above their heads. The dot means that they have a microphone on. When someone speaks, such as the avatar in the center foreground, greenish waves eminate from the dot - this is so others know who is speaking. As the area filled up with avatars, this became useful... though if one has preferences set to show names, it is difficult to see. At work, I run SL on a very high end machine, and even when there were 21 avatars in the area, with music playing and with other avatars chatting with voice, there was no packet loss (though others reported it on their machine). What was neat, too, was to hear folks from Canada, UK, Uruguay, Germany, Holland, etc. who were participating in the conversation. The drawback for many of those who do not speak English as a first language is that some are embarrassed by their accents, and some even have trouble keeping up with fast-talking Anglophones. I heard many conversation groups where people where speaking their native tongue, but in large groups, I'm sure English will predominate. Another concern for some is the privacy that text provides. However, from the standpoint of an educator, the voice capability will help overcome some hurdles.
The voice performance, overall, is really impressive. I had to move my mic RIGHT on top of my mouth to be heard well, but even using a headset/mic that is a couple of years old (it has a 3.5ml jack, whereas LL recommends a USB-mic), I experienced no problems. Well, there is an issue with the microphone dB boost, but that's not an issue at the SL end. I am told LL will be rolling out voice to the main grid on May 15th - at least to some areas. The folks I chatted with seemed to be skeptical it would work, noting that it will probably take a few more iterations of the main client for it be fluid. The concern I have is how much of the voice cost will be passed onto the end-user. For instance, for an educational sim, institutions spend $150 per month for tier (1/2 of what others pay). But will adding voice cost much more? I suppose we'll see.
Another issue is how to capture voice conversations. Currently, one can get a transcript of a chat very easily - either by copying/pasting from the chat history window or using a tool like Eloise Pasteur's NoteTaker. It turns out that Audacity works great to record in-world convo. Simply select recording to "What U Hear," and it captures voice coming through SL just as it would coming through any other source. Here is a MP3 (2.5 mb) file that captures some of the conversation I had while in the Voice Beta today. I made sure to get permission from the speaker before putting this on the web!! http://www.oero.soe.vt.edu/podcasts/sl_voice.mp3
The conversation talks about how voice will be received in the SL community. Obviously, some will be very uncomfortable with it, as they do not want to give away gender identity. The speaker mentions software that will change a man's voice to a woman's & vice-versa. He noted that one could detect the difference, but only to a slight degree. The software (US $39.99) he mentions is MorphVOX Pro - Voice Changer 3.2.2 by Screaming Bee. I'm sure it's one of many possibilities. The URL is: http://www.screamingbee.com/ (there is an evaluation version available).