Poll Results: Inworld Voice People Don't Use It

Top eleven reasons why voice isn't ubiquitous.
(... and why that might not be a problem.)

• 1) Not everyone has a mic
  If "downloading a client" is considered a significant barrier of entry, going RL shopping for a headset has gotta be right up there. If you're not using internet telephony chances are you haven't got around to it yet, or if you have not all your friends have. Which brings me to....
• 2) Only one person in chat without voice breaks the chat.
  How many times have you been having a conversation that's migrated from voice to text because someone doesn't have a dot above their head.Text is, most of the time, the best common denominator. Some folk may have trouble speaking or hearing (seeing lip reading is out of the question in SL) which is perhaps the ultimate reason to move a group to text. (Certainly this is often a challenge at Heron Island where the disability support groups need to run both voice and text at once for those who either can't hear or type reliably to communicate with each other.)
• 3) Immersion (but not as much as some imagine).
  Sure not everyone is going for a fantasy experience in SL, but for those who use the word "immersion" in describing their experience, fantasy is right up there. Nothing breaks immersion quicker than a battle dwarf with a Brooklyn accent or a lingerie model that talks like Sly Stalone. Personally I think the non-voice immersionists are not as large or influential a group as we imagined during the early voice vs. immersion blog battles. They're a vocal non-vocal minority :P Minorities however have some clout, in a world like SL where we are (by necessity) sensitive about issues of social inclusion.
• 4) Many SL activities are backchannel chat activities
  The three main times I see a crowd in SL are for parties, conferences/panel chats, and group discussions. It's rude to use voice over the music at parties and panel chats, and voice is non-optimal for groups above about five. The more likely you are to be in a crowd, the more likely you are to be somewhere that text is better suited.
• 5) Group voice chat is bound by legacy metaphors (In short, it's clunky)
  Here's a real criticism of the way buddies work in SL - finding calling cards in inventory and putting them into a folder to initiate a group chat?! Gimme a break. The voice UI was designed around a chat UI that worked when SL was small and groups more well defined. As SL has grown the nature of how we interact has outgrown the group and chat system. Why oh why didn't they buy up Jabber when they could before Cisco got it?! But I digress.
• 6) Vivox will never have the network penetration of skype.
  This isn't the fault of the SL voice system, which really is quite good for what it is. Skype however, is secretly a peer to peer system - even when you're not in a call, the skype client may route calls through your client (using up your bandwidth) in order to route around network problems. Skype is not the free ride it looks like - everyone on the network is part of the infrastructure. Vivox on the other hand relies on central servers for more of their system, so there's slightlty more boundary on growth. However, when it comes to proximity based voice like you get with the SL parcel channel, skypes network is completely inapropriate - SL voice would crush skypes loosely connected model.
• 7) We naturally employ multiple communications channels.
  I don't know anyone who only uses a single IM client these days, considering you get one built into gmail, facebook, and probably have some lint on your taskbar or several built into your operating system. Folk will naturally fall back on the comms channel they have established with contacts they seldom communicate with. Many of the folk you talk to, you do so with the channel you find them on, and only exchange other details as you continue to communicate - so SL buddies graduate to MSN or GChat, facebook buddies might get your email and phone number, twitter buddies may eventually get your skype id. SL voice will add to this mix but of course it won't replace them - nothing will.
• 8) Perception of unreliability.
  According to the metrics i've seen, SL voice is actually pretty reliable. We tend to use it though (if we do) the whole time we're in SL. Due to this close to 100% in world time useage, we become very sensitive to glitches and outages. Consider the time you spend per session SL to the time you spend talking on skype - for most users who have sl voice active there will be a lot more time logged with their ears open in SL than chatting away to skype buddies. That makes any interruption to sl voice services or glitches much more aparent. Skype doesn't glitch out much less than SL (really it's neithers fault, usually it's the internet itself) but because you only chat on skype when you are actually chatting, it seems like it does. Add this perception of unreliability to the teething glitches we all remember from rollout  last year, and SL voice looks a lot more unreliable (comparatively) than it is - which of course is what we complain about on blogs.
• 9) Second languages are prevalent in Second Life
  The SL community is truly international and so often in a group some members will be talking a second language. The difficulty of understanding spoken conversation in a non-native language with the many accents of folk in SL makes it hard for those with little of the spoken language to keep up. Text doesn't have an accent, and sticks around (an utterance doesn't disapear as soon as it's uttered) so many folk outside their native language community find text an easier medium for communication.
• 10) Voice is non-optimal for multitaskers.
  Folk in SL may be engaged in activities in another window - and in my experience they often are. Many folk in SL are there for enjoyment or communication while they engage in other online activities. Voice (as a temporal thing) demands your immediate attention, which is ok if you are communicating with people who you wish to respond to on a second by second basis. If however you are otherwise occupied, or trying to engage in multiple activities, voice becomes a nagging annoyance. In demanding your immediate attention it takes your focus off other things which may be of importance. In the case of a phone (or skype) call this isn't a problem, you just hang up when you're done, but 100% voice connectivity like you have in SL can cause frustration.
• 11) Voice doesn't work well with more than five people.
  Perhaps the most important element - voice chats become difficult with more than five people talking. This is plain old human nature, the more folk you have speaking concurrently the less conversational coherence you have. Text chats scale to much higher numbers before losing coherence. Of course talking can be moderated, but that's an unusual formality for impromptu chats, and a designated conversation moderator is thus a special case. With such a rich international community spontaneous moderation by conversational rules of politeness (which differ by culture) are less likely to emerge than in real life where body language is also available to cue speakers. Ironically, the much maligned typing animation serves this purpose well in text conversation.

Interestingly, all of these points were mentioned in discussions before the voice rollout way back when. We knew the sky wasn't falling on text (despite worry from several interest groups that it would.) The main reasons voice hasn't been universally adopted are because most people don't find it as useful, critical, or desirable as those who argue for ubiquitous voice use for their own reasons. At that time the voice dependant folk I knew were using the secondtalk skype bridge http://www.secondtalk.com/ or teamspeak to communicate. All of them (that I know) have migrated totally to the SL voice tool now for their in world activities - it may not be ubiquitous outside Second Life but is sure easier than sharing teamspeak server ips or skype account details with anyone you meet at random in world.

Having said that - 22 billion minutes of voice use in SL the second year after rollout isn't too bad at all. ( http://nwn.blogs.com/nwn/2008/09/look-whos-talki.html ) That's somewhere around 42,000 YEARS of streamed audio in a year (if they're using the same billions as I do). Quite how they got that figure given not much above 50k concurrency for most of it is beyond me, but surely they're not doing too badly for such a new entrant. Quite a lot better methinks than Gizmo (a skype competitor which nokia had an interest in at one point) and, we're told, about 20% of the volume of skype by the same metrics.