Memo to 3DVW Devs: “It’s the AVATARS Stupid!”
I’m following up my 3 part problems with 3D Virtual Worlds (3DVWs) with a series of at least three observations on what REALLY makes a good 3DVW. There is a surprising amount of myths in this regard. It seems most builders of 3DVWs do not understand their own potential audience.
Lets start with something that most all players understand, but as far as I can tell most game developers do not: “It’s the AVATARS Stupid!”
I have listed on my website something like 25 active 3D virtual worlds. Of these 25, how many have really great avatars?
The answer is zero. I counted … twice.
I’ll tell you what a really great avatar is. Its an avatar that looks like PS3 Home’s avatars, behave like There.com’s avatars, and is as versatile and adaptable as Second Life’s avatars. Not one 3D Virtual world fits all 3 bills, especially Home, There, and SL.
People define who they are in virtual worlds by their avatar. It is important for most players to have the means to create an avatar that is unique and at the same time beautiful.
Most 3DVW devs don’t get this. They are more focused on building a world to explore, or building the tools to let players build a world to explore. This is also important, no doubt, but no matter how nice your world is or how great the tools you develop, people will not stay very long if the avatars suck. I’ve seen it happen, heck I have done it myself.
Now I could spend the rest of this post trashing all the bad avatars, but that would not be helpful. Instead I am going to go over the almost good 3DVW avatars that are out there and tell you whats good about them and what needs to be in any future 3DVWs, if you want any chance to be successful.
Playstation 3 Home Avatars
I do not own a Playstation 3, so I am not an expert, but there are enough videos online to get the gist of the program. Basically the avatars in Home are realistic looking humans like any of the modern video games out there. They use morphing to make realistic looking joints and motion (failure of SL and There), high resolution textures and bump maps, clothing that actually has a thickness and folds and does not look painted on (another failure of others), lighting shading and highlights that looks more like real skin, real hair, real cloth, etc.
The primary negatives of the PS3 Home avatars is the quality and variety skin textures. Not just skin textures, everything. The lack of user customization means that everything about the avatars seems generic. They are also not very expressive, seems like a zombie world in PS3 Home.
Now I know what some of you are going to say. Why is realism so important? Many popular virtual worlds. both 2D and 3D have succeeded with unrealistic, almost cartoon looking avatars. I have even pointed out that there is such a thing as avatars that look too real. But, that’s not the case with any 3DVW I have seen. The truth is players prefer realism. In games that have custom texture markets (a subject for another memo post) the best sellers are almost always the most realistic ones. This is true in Second Life where realism is standard, and There and IMVU where avatars usually start with a cartoon look. Realism is an aesthetic people like in an avatar regardless of the environment.
There.com avatars
There has been around for six years now and is started to become dated. Despite this, there are aspects of There that are way ahead of their time. One is the behavior of the There avatars. First, the built in animations are first rate, better than Home even. They walk, run, sit, and stand very naturally. Their stances are constantly changing the way a real person standing for a long time would change.
Secondly, There avatars are great in the art of conversation. If you are using voice, the avatars look like they are lip syncing, and even gesture as they talk. If you are using text, it takes clues from the text to express mood. Use words like “happy”, “joy”, “fun”, etc, the avatar starts smiling. “sad”, “bad”, “bored”, and they start frowning. LOL or ROFL or ROFLMAO get increasingly expressive laughter. Thirdly, the There avatars seem aware of their environment. If someone is talking, or doing some kind of action, they turn and look at that person, and nod or maybe even smile with approval. They act differently around dogs, and with most vehicles they act like they are really interacting with them. If they are standing alone, they start to look bored even annoyed after a while. Yes, I am aware that this type of reaction is typical in single player video games since the 80’s, but surprisingly still uncommon in multiplayer games.
No other game comes close to this type of world interaction. IMVU can mirror some of it, but that is because Will Harvey is behind both programs. He is one of the few that gets it. Another one would be the short lived Google Lively, also created by a former There programmer. Lively handled two person interactions (handshake, hi-five, hug, kiss, synchronized dance, etc) far better than even There or IMVU.
Scripting in Second Life can mirror some of this action and interaction, but controlling facial expression via scripts in SL is very difficult, and turning a head to look at a speaker is damn near impossible. These simple actions make an avatar feel much more alive.
Second Life Avatars
It has been a complaint of Second Life since it opened almost six years ago that the default avatars are, well ugly. They look funny, they walk funny, the hands and face are damn near impossible to control. Those of us who came from There to Second life had to adjust to the lack of AI in these avatars. Suffice it to say that the strengths of Home and There avatars are precisely the weakness of Second Life avatars.
The one area where SL avatars are strong, which almost makes up for their weaknesses, is their customization. A good skin texture and shape makes them look a lot better, a script triggering animations makes them behave better. The “Appearance” options in SL has over 100 slide bars for customization, clothing and hair options number in the millions. Top and bottom halves of the avatar have 3 levels of clothing textures each, not including the base skin texture. You can attach objects to every part of the body, meaning if you want to look like another creature or “furry” you can. Attached objects can be scripted as well, so you can use particle effects to turn yourself into a human torch if you want. Customization is great for “role play”, easily the most popular activity in any 3D virtual world or MMORPG.
If you can imagine it, it can be done with SL avatars. No other 3DVW can even come close to this level of avatar customization.
Which leads to…
The ideal avatar is a combination of all three of the above, with the addition of the two person interactions like Google Lively had. I know the average Virtual World developer would balk at all the work it would take to pull off this much detail in looks, behavior, and customization, but it would make your world unique.
We have enough environments and building tools already, and these things can be upgraded as you go anyways. Avatars cannot be easily upgraded, especially if you allow future customization. If for example, Second Life fixed their avatars the way they should be fixed, all current clothing and prim attachments would basically become useless. The ensuing “restart” might ultimately be a good thing, but the aggrivation it would cause would be rather intense. It ain’t gonna happen.
Getting all three is probably beyond the current computing state of the art. The problem is combining Home-level detail with Second Life-grade customization — the bandwidth requirements for all those high resolution downloads would be prohibitive. And the rendering hardware needed to make it work (even heavier duty than Home needs because it can’t be optimized as well than it can be for professionally pre-designed content) would exclude 95% of the potential user base. In a few years, when the average computer is at the level of a current gaming enthusiast system and we can count on people having network connections that can reliably provide multi-megabit download speeds (and the network backbone improvements needed to deliver that to hundreds of thousands of concurrent users), it will be possible to deliver that combination.
You’ve done a good job of identifying the goal, and we can get closer than we are now. Certainly it should be possible to add improved avatar interaction to Second Life or a similar world, which would at least get us two out of three. Or we could have Home-level detail and interaction; again two out of three.
Excellent article and I completely agree.
@Shirley
I think current mainstream computer systems are perfectly capable of accomplishing the goals outlined in this post. However I agree that a majority of users are still running on older hardware that might exclude them.
I personally would be extremely happy with Home level detail for very customizable Second Life avatars before incorporating interaction AI which can be scripted.
Hopefully there are some developers out there “who get it” are working on the next best 3DVW that we can look forward to in the not too distant future.