Quote:
Could you get some type of gameplay using that model?

I'm not entirely sure I understand what you mean.

I managed to spend a bit of time today fixing some of the bugs in the code. In particular the aspect ratio bug (thanks for the heads-up). Here's some revised pics:




The code is now usable (running/jumping), meaning that I can test and debug the camera control methods as I please. Since I already have 3 different methods implemented, I'll crack open Texniccenter tomorrow to start writing the article. I've re-arranged the topics in the article differently from my earlier post, now they are in a format that should be easier to digest.

I plan to switch between implementing a new camera method, and writing it up. This way the article and code are updated regularly, and together. The current list of autonomous camera methods I plan on implementing/have implemented are:
1. Projection [done].
2. Polar coordinates [done].
3. Spherical coordinates [done].
4. Quaternion interpolation.
5. Proportional controllers.
6. Steering behaviours.
7. The constraint-based method developed for my PhD (hopefully my thesis results will come back before I release the code to this one though).

I will write up some other methods in the article, but won't implement them. Mostly because they're unlikely to be of much use to game developers. Once these methods are done, I'll move on to the replay/cinematography camera methods, which is the topic of the second of my 2 camera control articles. I'm not sure if I'll use the same framework for these, or whether I'll strip it down to the bare essentials (ditching PhysX), or move to C# and MDX. Either way, there will be fewer camera methods available for replays/cinematography than for gameplay.

The current build doesn't have occlusion avoidance, despite it now being an option (see pics). I plan on adding this code once all of the camera methods are done. Since I plan on writing this up after I finish implementation of the gameplay cameras, if anyone is really hanging out for the occlusion stuff, I can put it at the top of the list of things to do.

For the eager people wanting to look at the code, the current source and documentation can be downloaded from here (zip) at 3.3Mb. This zip contains all of the "art" (and I use the term loosely) assets as well as the code. It's a Visual Studio 2005 project, using Feb 2006 DX and PhysX 2.3.3. I have my doubts that any of you will get it to compile, but good luck [smile].

EDIT: Just uploaded the executable at ~200Kb. You'll need to download the exe and the source, and put the exe in the same folder as the PhysX dll's. Lot's of messing about, I'll fix it for next time.

The controls are: up/down/left/right cursor keys control the movement of the model, space for jump. For the free-roaming camera, the WASD keys handle movement, while the mouse handles 'looking'.

Remember, the code is to demonstrate camera control not engine design (and is free), so please refrain from "this code sucks" comments. Other comments are welcome.