Thread: isometric view

What I say here represents my point of view, and is not necessarily correct

Learn 3D graphics programming, either using OpenGL or DirectX Graphics. You'll only need the basics. Then adjust your camera's orientation ( rotate it 45 degrees around the y axis, 45 degrees around the x axis ) and you get your isometric view with no pain at all.

You'll define every tile as a square, and treat it as a square ( since it IS a square ) and yet, it'll be isometric ( Diablo/Fallout/Syndicate views can be very easily immitated in 3D )

I've tried to learn Iso game programming a couple of years ago, and in the middle of the fray a programming guru that used to check in this board ( called Sunlight ) totally changed my mind ( and I do 3D graphics now )

The quality of your sprites and characters is of the most importance for an isometric game. Best to render them with a 3D animation studio (use to be Blender *sigh*), maybe touch up with something like Gimp or Photoshop if you have it. Render at a slight angle and everything with that characteristic 45% angle. Not too much of a drop in elevation, as you don't want to have to change the size of a character depending upon where it is in the viewable screen. Usually the ground fills the screen, but I've seen some isometric engines get away with having such an extreme angle that the sky appears. Usually there is some sort of "fog of war" or such that keeps the perspectives from looking bizarre.

A good mix of iso and 3D can be ideal in some cases and look a lot better. Prerendered background tiles with 3D characters, for instance, will allow you to put a lot more detail into the backdrop and save the rendering power for really classy 3D models.

That sounds like a good idea but wouldnt it waste a lot of processing power?

This is what I THINK. Don't take it for granted

If you do it in 2D ( iso ), you'll need to render your tiles too. So the only difference is in the 3D transformations done to convert from 3D coordinates to screen coordinates. These are the object-world-view-projection transforms.
Each is a 4x4 matrix, and the transformations are basically multiplications. Any Pentium compatible processor can handle multiplication effectively, even a P90.
I believe there's no waste of processing power. I'd say there's more processor power demanded & used, but it's not wasted.

3D ISO looks better than ordinary ISO, Plus being able to rotate the camera as you like (with some minimal work).

Additionally, collision detection becomes absolutely easy ( straight sphere-plane , sphere-sphere , sphere-box, box-box, box-plane collisions)

I really prefer doing straight, clear checks for collision ( like the sphere - plane check for walls, for example ) instead of doing some checks in some tile map and see where the player is.
3D algorithms tend to be real, which makes me comfortable

Additionally, lighting becomes very simple. Using lamber/gouraud shading, even adding specular highlights is very easy. Reflectance becomes simple,...etc
Doing these in 2D iso requires writing ... strange algorithms. Because you're not following the original physical law (or one of its simplifications), you're just immitating ( simulating ) the results.

About the "where's sunlight? why'd he leave?" thingy
I don't know, he left and then came back, and then left, and then came back under the name of Esss, and then he left.
I'd say it's a combination of one or more of the following :
- Not enough time
- Bored
- No advanced questions
- Being flamed (happened a couple of times, at least)

The best thing to do is to ask him!