Essentially the fact that you're sampling and not directly rasterizing. If a poly is too small or too skinny to rasterize, you have to sample it, and a ray that you cast is essentially a sample. Granted, rasterizing with supersampling can have that effect too, but that's no less costly and it really only raises the limits on how small is "too small to rasterize" as all you're really doing is raising the resolution. A raycast doesn't necessarily have to be directed by screenspace, so you can just cast it to tiny geometry and see which pixel(s) are intersected and affect them accordingly. Doing this is more of an example of a hybrid rasterizing/raytracing kind of thing since you can still do immediate fill stuff, but the only real difference is that you're accumulating samples which affect pixels (though this by itself only works if all geometry is tiny). Breaking down all source geometry into subpixel size and then sampling and shading those is essentially REYES.
Depends. Years ago (32-bit era) we didnt even mind having flat surfaces with crowd textures pasted on them in racing games.
Well I agreeA lot of people will expect huge leaps in every department of a game's graphics. While this is true for certain departments, others can just be improved on while keepign the same basic fundamentals. Today's sprites look a hell of a lot better than yesteryears'.
Some nextgen games will start to use more and more real 3d models for spectators, others less so. It will all depend on how comfortable the team is with throwing away resources on something that is rarely looked at with much detail. Some things come in small steps, others in leaps and bounds. We're seeing leaps and bounds in physics and shader power, but other things are slower to evolve with every gen. We might expect too much out of nextgen at first, but in the end it's still a video game, and devs still work very hard to cut corners and "fake" certain elements when they can so that they can concentrate the detail up front.
So yes, we are seeing things progress very rapidly, but it doesn't always make sense to just throw away resources when you can have an alternative that looks almost as good to 80% of players.
just my two cents
Agree 100%A lot of people will expect huge leaps in every department of a game's graphics. While this is true for certain departments, others can just be improved on while keepign the same basic fundamentals. Today's sprites look a hell of a lot better than yesteryears'.
Some nextgen games will start to use more and more real 3d models for spectators, others less so. It will all depend on how comfortable the team is with throwing away resources on something that is rarely looked at with much detail. Some things come in small steps, others in leaps and bounds. We're seeing leaps and bounds in physics and shader power, but other things are slower to evolve with every gen. We might expect too much out of nextgen at first, but in the end it's still a video game, and devs still work very hard to cut corners and "fake" certain elements when they can so that they can concentrate the detail up front.
So yes, we are seeing things progress very rapidly, but it doesn't always make sense to just throw away resources when you can have an alternative that looks almost as good to 80% of players.
just my two cents
X360 has 3d crouds btw.
I don't know, dude. I mean, how can you say that muscles visibly tensing and tendons moving isn't more important than having a nice crowd? Honestly, unless the crowd is somewhat interactive, as in wrestling games, I really don't care. I'd rather have more realistic fighters than a realistic crowd.