Engine Design (Programming)

[Kopavel]

See http://forum.keeperklan.com/wfto-kopavel-t718.html

[Hapuga]

Kopavel, are you familiar with DK2 engine? Do you know how to manage it's files extensions?

[Hapuga]

I am developing a DK engine, and encountered a couple of problems I feel would be advantageous to discuss.

Problem 1. The dungeon.
subproblem 1.1. generating dungeon geometry
each cell (most cells?) has unique geometry (noise applied over base geometry) - so we either should generate it on the fly, or employ some vertex shader tricks.

subproblem 1.2. rendering dungeon geometry
rendering each cell individually will result in huge amount of DIPs that would eat all the perfomance at once. Should we generate a solid vertex buffer, or use instancing in some way?

Problem 2. Lighting
given possible number of lights in a scene, we probably the only option is deferred lighting. However, torch lights are mainly static, and could be stored in lightmap or vertex color?

Solutions would be heavily based on target hardware.

As far as I understood, WFTO 0.2 doesnt adress these problems.

(to be continued...)

1. How about generation on encountering? this will save resources.

2.how about making a basic pattern?

[Kopavel]

Kopavel, are you familiar with DK2 engine? Do you know how to manage it's files extensions?

Nope and nope. But I dont think DK-2 engine is something we have to look at.
I remember DK-2 lagging when I zoomed out, this leads me to idea that they were rendering their dungeon blocks with individual DIPs (very unoptimized).
More importanly, DK-2 has nothing to do with shadows (even DK-1 had more advanced lighting).
Anyway, I've still experimenting with possible solutions I've described above. Here's a working screenshot, where I managed to combine vertex lightmaps with normal mapping:

This screenshots also presents you DOF, Bloom, Fog, and some temporary debugging information.

[Kopavel]

1. How about generation on encountering? this will save resources.

2.how about making a basic pattern?

Yep, thats it. As soon as finish experimenting with these, I'll post my detailed solutions. Basically, I expect the same from you

[Madkill]

Target hardware should be based to the highest performances the team can get this too, from there its just simple back-tracking to go from the set 'High' to Medium and then low.



Anywho, is it really safe to give away secrets to others who are wanting to make a DK engine?


By safe I do of course mean, are people around here that desperate for some shoddy piece of crap, than being patient for the current project to finalise?

[dotted]

Anywho, is it really safe to give away secrets to others who are wanting to make a DK engine?


By safe I do of course mean, are people around here that desperate for some shoddy piece of crap, than being patient for the current project to finalise?

If anything collaboration is what makes things move forward in open source projects.

[Hapuga]

Target hardware should be based to the highest performances the team can get this too, from there its just simple back-tracking to go from the set 'High' to Medium and then low.



Anywho, is it really safe to give away secrets to others who are wanting to make a DK engine?


By safe I do of course mean, are people around here that desperate for some shoddy piece of crap, than being patient for the current project to finalise?

Have you seen this "shoddy piece of crap"? So how can you judge? It is his engine and he can do anything he wants with it, so stop acting like a fool. Are you a fucking game developer, to say "for the current project to finalise?"? There can be 100.000 dk projects because they all will be fan-made. So who prevents?

[Blutonium]

Nope and nope. But I dont think DK-2 engine is something we have to look at.
I remember DK-2 lagging when I zoomed out, this leads me to idea that they were rendering their dungeon blocks with individual DIPs (very unoptimized).
More importanly, DK-2 has nothing to do with shadows (even DK-1 had more advanced lighting).
Anyway, I've still experimenting with possible solutions I've described above. Here's a working screenshot, where I managed to combine vertex lightmaps with normal mapping:

This screenshots also presents you DOF, Bloom, Fog, and some temporary debugging information.

your engine already supports normal mapping?!

/drools

you seem to be giving our NBK devs a run for their money!

1) do you have a team or are you working alone?
2) How far are you going to take this engine of yours, and how long will it take?
3) ever considered joining our team which has a pretty stable engine already and implementing normal mapping and so forth?

[Kopavel]

...

First of all, YES, I want to join your team, but most of my messages dont arrive at this forum! It says something about moderator who would check the message before it's posted (I also used your developer application feature). In fact, I started this thread to draw your attention.

I'm working alone on my engine, and of course it's capable of normal mapping along with dynamic / static lighting (and, as I already mentioned, several post-process effects). I've made a winning entry to Intel Level Up 2009 Contest with my engine (google it!)

I'm planning to implement visualisation system, and the game system (including digging the dungeon and building the rooms), after that clean up the code, write comments, and share it with you.