The topic of Second Life’s default camera offset is an old one. The first SL blogger / user who brought it to people’s attention was Penny Patton, with her seminal post “A Matter of Perspective” from as far back as 2011. Ever since then, various other bloggers have written about it, basing their work on her own, hoping to get people to understand how and why the way the way we see the virtual world through the placement and behaviour of our camera affects the way we experience and use it. As a matter of fact, it even has a major impact on the very usefulness of the platform itself. I had written about the issue four years ago, and I have been pestering Oz Linden himself about it in various Open Development User Group meetings, but to no avail; he acknowledges the issue, but he cites the risk of potential complaints about “content breakage” as the main deterrent.
The points made in my older post, as well as Penny’s original post, are still current. Second Life does not have a first-person view (think Doom, Quake, etc) as its default. Instead, the default perspective is third-person, which allows you to see your avatar and what it does in-world, provides a much wider field of view, and allows a more realistic portrayal of close-up 3D objects. Within the third-person perspective camp, there are several approaches, all of them depending on the application. For instance, for a third-person shooter or driving game whose action takes place in a wide open space, you need to see as much of your surroundings as possible to have the best awareness of the situation around you. So, it makes sense to place the camera somewhat high above your sprite (be it a fighter aeroplane, a combat spaceship, an automobile etc), behind it and not too close. However, for games taking place indoors, in dungeons and maze systems where the action happens in close quarters, other views were chosen – isometric, first-person, etc.
Second Life uses a third-person view with the camera placed quite far behind and above the avatar. It’s what I call a “radio-controlled car view”, as it’s the same view you have while operating a radio-controlled car, and it’s a development on earlier third-person perspectives, from Tempest to Tomb Raider. With the default camera offset, the camera floats at least two meters above your avatar, virtually making it a toy rather than part of “you” – but we’ll get to that later. Since 2005, when Resident Evil 4 was released, the third-person view, for which Penny is a vocal advocate with the camera looking “over the shoulder” of the player’s character prevailed as the industry standard. Penny had explained the reasons for this, and I’ll reiterate them. Also, please note that, in this post, I’m going to reiterate and update several points I’d made in my older post from four years ago, as they’re still valid.
- Environment creation. Second Life is a virtual world platform where the user might wander in an open plain, in a forest, in a bar, club, house, or what have you. The over-the-shoulder view allows a builder to create accurately-scaled and usable builds like an outhouse or a small shed in the woods or a cramped washroom stall in a sleazy bar, or claustrophobic corridors in a maze-like environment (catacombs, sewers, labyrinths). Also, with this view, larger builds are more impressive, without needing to be upscaled at all. On the contrary, the default camera settings in Second Life place the camera so high above and so far behind the avatar that, should you enter a realistically-sized build, the camera ends up either in the ceiling, or just above the upper floor’s – ahem – floor, or inside the walls behind you. This makes it impossible to get around in a build that has the same dimensions it would have in Real Life. Thus, the default camera offsets force us to upscale everything by at least 20%, with the vertical axis “enjoying” an extra “bonus”. This has dire consequences to our builds, avatars, commercially-available user-created content, and, eventually, the usefulness of Second Life itself for several of its intended uses.
- Usability. The over-the-shoulder view provides the user with an additional advantage: a sense of place, which allows you to have a more intuitive “feel” of where your avatar is in relation to its surroundings. It’s obvious that this allows us to navigate a region far more easily.
- Immersion. Placing the camera near the level of the avatar’s eyes puts you, the avatar’s operator, into the world where the avatar moves and exists rather than making you an outside observer who merely uses a radio control system to make the avatar do whatever it must do.
Scale and Usability
Ever since Second Life debuted in 2003, in spite of technical advances, and in spite of a growing tendency among content creators to build realistically-proportioned and sized objects, which is in an infinite loop with the demand for more realistic default camera offsets, the viewer’s default camera offsets have remained the same. Let’s have a look at how they impact our view of the virtual world. As a setting, I’m using my in-world workshop / sandbox, where I have placed a build using a (non-resized) copy of Apple Fall’s already realistically-sized and proportioned Portobello Corner Store, which I’d written about last year and you can purchase at his mainstore, and a downsized copy of Trompe Loeil’s Brooklyn Carriage House. All images are uncropped and unedited.
With the default camera floating so high above and so far behind the avatar, it looks indeed like you’re operating the avatar from above. Also, notice how the lower parts of each object in the view (such as the avatar’s shoes) appear to be much smaller than they really are. Let’s go inside now.
As you can see, this view makes the place, and everything in it, appear smaller than it is. Plus, it causes all sorts of navigation issues, as moving around causes the camera to get inside walls, ceilings and other obstacles, thus making exploration of realistically-sized indoors areas a pain. And now, let’s have a look at the view from the front with the default camera offsets.
Note that, being a neoclassical build, the Portobello Corner Store has a high ceiling. But what if you want to enter a contemporary home? Well, depending on when your home was built, RL ceilings are 8″ (2.438 m) or, if it was built after 1994, 9″ (2.743 m). You will also encounter 10″ high ceilings (3.048 m). The apartments and houses I’ve encountered in RL usually had 10″ high ceilings, with the exception of the ground floor of some really old neoclassical homes, some of which followed the old European standard (13″ – 3.962 m). The one where we live now has a 9″ ceiling. As you can understand, with the default camera offsets, you’d have a very hard time walking and looking around in an realistically-proportioned and sized build.
Losing the architects and civil engineers
“Second Life is the best tool for architectural visualisation!”
– No architect or civil engineer, EVER.
In fact, all architects and civil engineers I know have dabbled a bit with Second Life and OpenSim, and duly abandoned them, because they make no sense whatsoever to them. The reasons they cite are the default camera offsets, the lack of interoperability with files generated by CAD applications, and the cost. Please, don’t start with the “OpenSim is free and you can set it up on a USB stick on your computer” thing. OpenSim requires significant set-up effort and is actually hobbled by the same issues that SL has – such as model import. No architect I know can justify the cost of renting a homestead or a full region, and the effort and man-hours needed to prepare a 3D model for use in SL – or OpenSim. Furthermore, a lot of third-party content sold in-world and on the SL marketplace is built to look “good” with SL’s default camera offsets, and is both oversized and non-realistically proportioned. When was the last time you saw a passage door with a thickness of 20 cm (approximately 8″) in RL? When was the last time you saw stairways with stairsteps with a rise of 50 cm? As you can understand, not only do SL and OpenSim require significant fiddling with the 3D model to ensure its physics are serviceable, but much of the content sold on their marketplaces is useless to an architect, civil engineer, or decorator. And don’t even get me started on prefab buildings. By insisting on not fixing the camera offsets because its decision-makers fear some users might whine about “content breakage”, Linden Lab has missed the train of real-time, distributed architectural visualisation, and I’m not sure if any time has been left for them to catch it. What’s ironic is that the fix to this would have been extremely easy.
That’s LL’s explanation. I can understand why they fear people might get up in arms about “content breakage”; the default camera offsets have been around for so long, that users have been conditioned to thinking they’re the norm, and thus have built their entire world with, for, and around them. So, the Lab fears that, if they roll out a new release of the official viewer, where improved, updating camera offsets are the default ones, people will go up in arms. In reality, no real content breakage will occur from a move to better, more realistic, camera offsets. Your furniture will still work just fine. The same applies to your vehicles, homes, everything. What will happen is that you’ll realise how oversized everything is. You’ll realise that your avatar is way too tall, dwarving even the tallest NBA players, with unnaturally long legs, ridiculously short arms and torso, and a tiny head. You’ll realise what an idiot you were for calling those people with realistically-sized and proportioned avatars “ageplayers”, starting drama, and filing abuse reports against them.
And then you’ll try to make your avatar look good, so you’ll get with the programme and start using a more realistic shape. But then you’ll see that your existing animations and poses suck, because, being made for the default camera offsets and for the T-Rex avatar, the arms stick inside your body, your thighs, and other body parts every hint of a chance they get. You’ll try to make your house smaller and more proportionate – but then it’ll be at odds with your idiotically oversized car, which is (of course) non-modifiable. And your furniture will probably require new animations and a fresh installation of the scripts. Chances are, you’ll end up doing an awful lot of work. Perhaps, like me, you’ll say “well, it was worth it” and not look back.
But, for all its failings and inexplicably bone-headed decisions, the Lab has understood a few things about its users. The Lab, through fourteen years of interaction with SL’s userbase, fears that you are far more likely to go up in arms in the forums, yelling, screaming and perhaps blaming those “cunning ageplayers” who “goaded the Lab into making camera offsets to accommodate kid and lolita avatars” and caused you to break everything in your inventory. A quick look in the official forums, on various Plurk accounts, and other such outlets for opinion on all things SL justifies this fear. Do you think I’m exaggerating?
What does this mean? Well, for the Lab, the situation is perceived as a Catch-22. So, rather than change the default camera offsets and ruffle people’s feathers, they opt to simply not do a thing and continue kicking the can down the road. Perhaps they’re waiting until more people have adopted better, more realistic camera offsets and then making a gradual transition, when the oversized content has become obsolete enough for people to not care about it anymore. In fact, in an Open Development User Group Meeting, Oz Linden had hinted to the idea of a UI that’ll enable users to easily change their camera offsets.