We've seen a lot of graphics techniques that we can use to create virtual reality environments. But what is different about the techniques that we need for VR as opposed to doing games and films? Well, not actually that much. Virtual reality is a sense to using the same graphics techniques that you're using in games. Film is a little bit different because it's not real time you can use more computation intensive and less interactive techniques. But games in VR shared most of their basic technology, and that's why you very often see game engines like Unity or on Wheel being used to create virtual reality. Having said that in this lecture, I'm going to talk about some of the differences that you need to think about if you come from a games background. The first, is performance. As we've seen, VR can be very computational expensive and has to run into very high frame rates so VR graphics needs to be fast. Secondly, there are differences relating to the way we are viewing VR immersively rather than on 2D screen. There are differences about how we create presence which means plausibility implies illusion. And finally, we need to think about comfort so not creating nausea or other uncomfortable VR experiences. Performance is really quite straightforward. We need to end the two screens in under 20 milliseconds. That's a really strong performance constraint. So even on the best PC hardware today, you're still quite limited about what you can do. You have to be quite careful about the speed of your experience and on mobile, you are even more constrained in terms of your graphics. There are lots of techniques you can use to speed up your graphics but the big picture is that the more polygons you have the slower your rendering will be. So you need to be careful about the number of polygons in your models. You should also be careful with very high resolution textures and some of the more advanced shader and material techniques. We'll put a reading below with some up-to-date recommendations about how to optimize and speed up your graphics. Another issue is simply that you're viewing this in a head tract, head mounted display instead of on a TV screen. And that means certain techniques work better than others. One example, is a very popular graphics technique called bump mapping. It is like texture mapping but instead of color, the texture affects the direction in which the light bounces off an object. It can create the effect of surfaces that seem bumpy because light is bouncing off in different directions, in different places, creating a nice 3D textured feel. And this is used a lot in games to create objects that have simple geometry but appear to be more detailed because of this bumpy surface. Unfortunately in VR, this doesn't work so well. If you're viewing on screen, this bumpy texture is very effective but if you're viewing it in stereo from both eyes and moving your head around, you aren't getting the power lights movement that you get from real bumps. So the bumpiness doesn't look as 3Ds as it does on screen. In fact, we expect a more 3D effect because we're wearing this fantastic 3D hardware. I'm not saying that you can't use bump mapping but you shouldn't use it at large scale to replace polygons. You should use other techniques. There is actually a technique called parallax mapping which works a lot better. This is just one example of a technique that doesn't work so well in VR, and there are others. Sometimes techniques work better with VR. Virtual reality is all about creating place illusion, plausibility illusion, and embodiment illusion, the illusion of being somewhere, interacting with things, and inhabiting another body. Together, these illusions are called presence. Graphics can support or fail to support these illusions particularly the first two. Interestingly, you could think the better graphics are simply better. In some degree that's true. Having better graphics isn't going to hurt you but in an interesting set of experiments in the early 2000s, Mel's latest team investigating this and actually lots of graphics techniques did not improve presence that much. You got a lot of presence with actually quite simple graphics and adding some better rendering techniques didn't do much right up to sophisticated global illumination. But some techniques did make a big difference. For example, real time shadows create much more presence probably because they create a greater sense of depth. Having said that, I'm not saying don't use advanced techniques in global elimination. People still like good graphics but it might be one of the things you trade off to get better performance. You don't always have to have an immense amount of realism to feel. In fact, there were lots of VR experiences that is actually quite simple graphics and they can look lovely and be really really effective. The last thing, we need to take account of in VR is comfort. We don't want to cause nausea or other forms of discomfort. Nausea is primarily caused by movement but there are a number of things that can make this worse. One of these is high frequency movement. So movement of very detailed textures, something that can cause this is aliasing. We normally think of aliasing as the kind of jaggies we see on low resolution computer displays. So if you try to draw a straight line but your pixel resolution is low, you see these jaggies. This is worse in VR than on screen or print because as you move your head the jaggies move around. This can be quite disturbing particularly in the periphery of your vision if you've got very bright aliasing patterns. So another really important thing is to use anti-aliasing to prevent this. And again, we'll put some readings later that will give you the state-of-the-art in these techniques. So in essence, VR graphics is not really that different to go in graphics but there are things that you need to keep in mind. There are a lot more than I can cover in this lecture, so I'm going to put some links in readings later about best practices for VR graphics.
