Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Davion]

Look at how you drew the corners for one room by itself -- you did proper convergence, i.e. there's a vanishing point.  However, you have a different vanishing point for each room, which makes it look like they're angled toward each other.  Not gonna work out.

Yeah, I know, I was just throwing it out there as a way to show all walls.

I'd prefer something like the 3/4 top-down if it came to that, but there's still the problem of not seeing everything.

[Capntastic]

There's a ridiculous amount of image leeching going on in this thread, especially from Tormy's end.

[Align]

Hmm, 2D Zelda's were straight top-down like that. Could work... In fact been proven to work. Huh.

[Footkerchief]

Hmm, 2D Zelda's were straight top-down like that. Could work... In fact been proven to work. Huh.

Didn't those just use prebaked faux-perspective though?  Or do you mean the 3/4 topdown stuff in outdoor areas?

[texmith]

Speaking of vanishing points, I saw curiously little discussion of this:

I particulerly support better handling of slices lower in elevation. Wasn't one idea just to scale progressivly every z-level beneath the current one and use a parallax? That could work wonders with things like towers and pits, depending on if we scale up or down the size of lower levels.

It really seems like the most promising approach to me.  The main issue is what to do about the vertical surfaces connecting z-levels, which of course don't actually exist in the game.

You'd be able to see mud that should be covered by the top of a hill, but the top of the hill is off somewhere else. The vertical lines of walls that should line up would be even worse. It'd be quite easy to make a mock up of this with current mayday set (just take 2 screens on seperate levels and cut out the gaps on the top one).

Here's my idea:



A variation on the basic angled perspective. Make sure every wall is the same height, and by simply drawing the whole lower level first, and at a fixed offset, they match up nicely much like isometric. Each graphics set would need to set this offset value in an init file. This approach would fully support top down views and current character sets (just set the offset to zero). Would require ramps to have a specific direction but I can't think of any other complexities in programming. Not sure how easy it would be to draw it well.

Still has the problems of hiding items behind walls (although perhaps not too badly). Also, perhaps lower levels should be drawn darker or something like previous suggestions.

[texmith]

oops. quoted myself when i meant to edit

[Mikademus]

Look at how you drew the corners for one room by itself -- you did proper convergence, i.e. there's a vanishing point.  However, you have a different vanishing point for each room, which makes it look like they're angled toward each other.  Not gonna work out.

That's actually how many games looked in the 80's and still in the early 90's, absolutely nothing wrong with it, it only takes some getting used to. Which every player getting immersed in an ASCII world should know everything about already.

Yes, every room has its own perspective, and if you look at everything at the same time that's physically monstrous, but

 1) Isometric, cavalier and any other fixed-angle bitmap perspective is also unrealistic in that reagrdless of where on the plane or in the distance you look the falloff is the same, and Irecall the first time i viewed graphics like that my brain protested wildly

 2) From a cognitive perspective we tend to focus on one point at a time, and for that particular focus point the perspective will be quite pleasant.

 3) The dwarves COULD build highly reclining walls (which is actually how Myth: The Fallen Lords pulled off using a heightmap to model interior walls).

[texmith]

That zelda style pic would look much better if the opposite walls weren't the same shade i think. Looks like its just been mirrored which makes the lighting seem off.

I think the trouble with that idea is creating a scheme from it. Its all very well drawing rooms like that.. even creating a tileset and using that to build maps, but creating a scheme thats got to represent all sorts of screwed up designs with walls in any possible configuration seems too much. Already single hallways and double hallways have different angle walls there. Pillars would look very odd if you could see all the sides but pillars with less perpective would look wrong next to angled walls. Working out all the rules of where to use which angle wall would become very ambigous, complicated and very specific to one particular tileset.

[Random832]

I'd just like to note that there are 256 kinds of ramps - 144 you'll have to draw (maybe not draw, but even the ones that can be unified you'll have to make design decisions about, and some of them it's not obvious what they should look like) if your design allows horizontal flipping, and a cookie to the first person to explain why it's not 128. (and of course there's more if you want to consider adjacent ramps separately rather than treating them simply as "no wall")

Some basic (pre-rendered then cached) 3d rendering for terrain might be desirable - you don't have to hand-draw them, your design decisions _can_ be flipped and rotated in all directions, and maybe you'll also have/want to apply textures depending on natural vs rough constructed vs block, stone vs metal vs wood, nevermind that some kinds of stone have ridiculous colors and maybe should have unique textures.

If we're doing 3d rendering why not go all the way, you ask? Because 3D rendering in a parallel projection (like isometric or cavalier) rather than in perspective has a huge advantage over "true" 3D rendering - The same tile, in the same orientation, has the same pixels no matter what position it or the camera is in.

[texmith]

good point. ramps can get rather complicated. Thats an issue for isometric too.

[Calessa Lynn Orphiel]

As a strategy game enthusiast, I'm a huge fan of the Isometric view.  I'm a very visual person and I'm having a hard time seeing what the problem is that other people have with it.  Does anyone have a pic that explains what the problem is, exactly?

[Grek]

Why not  just use a horizonal vewing option? Toady would just need to allow the player to change from looking down parallel to the z axis to looking along the y ans x axises. / and * can be used for turning your "face". For graphics, just switch walls so that they have slightly different graphics when they are lone pillars vs. looking at a row of wall tiles from the side. Creatures, trees and other non-wall objects which cannot be walked through would use the same graphics as they do now. If you want to get really fancy, add a side box with a top-down veiw like you have in adventure mode while in horizontal veiwing mode. Later on, a mode with only 1 z level shown and larger tiles allowing for even better, but still tile based, graphics.

Walla! No need to muck around with isometric and cavalier. Just a slightly different pattern for the same tiles.

[Mike Mayday]

Texmith: that's exactly what I had in mind when describing 3/4 top-down.

Random: true enough, ramps will be a problem. I'll see how well I could render them automatically.

Calessa: the problem is with walls obstructing the view of tiles that are behind them. Take a look at the XCOM screenshot posted earlier: some important item could by lying by the southern wall and you wouldn't even know it's there. You'd have to either make the walls transparent (which is IMHO ugly) or lower the walls (which questions the purpose of implementing such a view in the first place).

Grek: A horizontal view would be completely unplayable, only useful for viewing engravings, I guess. What I propose is a "pseudo-3d" view that allows you to constantly play the game in it if you so choose.

[Tormy]

the problem is with walls obstructing the view of tiles that are behind them. Take a look at the XCOM screenshot posted earlier: some important item could by lying by the southern wall and you wouldn't even know it's there. You'd have to either make the walls transparent (which is IMHO ugly) or lower the walls (which questions the purpose of implementing such a view in the first place).

Yes...I was thinking about this yesterday...this is the biggest problem of the iso/cavalier display method. I think that this problem can't be solved. Like I've said, perhaps the rotatable method could work. What do you think about that Mike? [IIRC you have dismissed my suggested rotatable idea a couple of days ago, and I don't know why...but I don't remember correctly.]

[Mike Mayday]

Hm, I don't think I've dismissed it but here the deal: take a look at the mockup of cavalier and you'll notice the walls don't leave their tiles, so the problem is fixed.
On the other hand, imagine a single tile in iso with walls built all around it. No matter how you rotate the view, you won't be able to see the tile (this was brought up by someone else).
I'm not saying I don't want an isometric projection implemented- Chris' mockup was enough to sell me on it, BUT I think implementing a view that is fully playable without constant switching transparency or rotation is more important.