New DirectX9 DM clone - entombed

[mikko]

Yep, normal mapping doesn't look anything like displacement or parallax mapping. Normal mapped surfaces still look flat even though the lighting is correct but displacement mapped ones look like they actually have dents and bumps etc. Especially at close ranges and when you don't see the flat polygon edges that tend to ruin the effect..

HL2 is still using light maps even with maximum settings. Maybe not everywhere but in most places. And every light source reflection is clearly cube mapped and not per-pixel. But it still looks great and I don't mind.

The only game where I've seen actual displacement mapping is F.E.A.R. One great game graphically and atmospherically in my books.. It just needs a monster machine for higher settings.

[DSE]

We all agree then that HL2 (probably the best game ever made) is really very good indeed.

Tom - Just to clarify - I am sure HL2 does not use per pixel shading (the lighting in the game isnt accurate enough, it must be light maps with per vertex lighting for dynamic lights) BUT I am not saying that their SOURCE engine cant do it (maybe lost coast implemented it, but I dont think it did). Per pixel lighting is very expensive BUT I think it looks much better than light maps or per vertex lighting and suits a dungeon based game where lighting is scattered. I dont really want to get into a debate about what is possible to code, because its a pointless argument. ANYTHING (within reason) is possible if the coder has the skill level and energy to implement it. If we go back to my project I have implementing per pixel lighting and the demo is going to be free for all to download and try, so I was just trying to make people aware of the possible hardware issues that they are going to have. Are you are coder, if so can we take a look at your projects ? Unless you have tried to code something you wont appreciate how technically difficult it is, especially if you want to please everyone with their 3+ year old gfx cards. If you look at it from my perspective, why would I want to write code for older cards when I wouldnt get any real pleasure from seeing the results. They would look out of date before I had even started, and the effort involved would be massive. This is one of my hobbies, not a commercial venture.

Gambit37 - can I ask why you posted "AHA" and what you possibly thought it would achieve ?

FYI - you can now get a radeon 9800 pro for around £70 on ebay. If you are into games its not that much money.

thanks,
Dave.

[Gambit37]

The 'aha' was meant as nothing more than 'thanks for the explanation' and not the 'there you go, it's possible, now go ahead and code it DSE' that you seem to be assuming I had implied.

I understand your view better now, so it's cool if you don't want to support older cards. I need to upgrade anyway!

[PadTheMad]

Even tho my 5200FX supports pixel shader 2.0, it kills my framerate whenever I use anything with it turned on

Looks like we all need to upgrade...

[Fozzedout]

I work with DSE, and he brought it in the other day.

I was gob smacked. Something that he hasn't mentioned is that he has also done the monster AI, so if a path is blocked it will walk around the dungeon taking the shortest path to get you.

For a hobby of his, this is a really impressive DM clone. The pictures on his site just don't do it justice.

BTW, just so I don't forget, my card is a GeForce 6200, with Vertex & Pixel Shader 3, so it should work on my card, and it only cost £60 from PC World...

Cya Tomorrow,
Fozzedout

[DSE]

Howdy chaps . I have now incorporated the physics into the game engine. And the really good news is it only requires a dedicated 500 node neural network to run, everyone ok with that ?

Anyway, take a look at the website to get an idea of whats been done.

Well, the physics work nice and as long as you dont pile more than 400 boxes on one level the game still runs at around 75 fps (consistant frame rate up to around 300 boxes).

One more step towards a playable release !! To be honest this did take a little longer than expected (doors in frames were a real pain) .... but it works really well now.

Cheers.

[PadTheMad]

Looking better every time DSE! And it's nice to see you've got physics working. Can I ask what physics engine you're using?

Oh and I know doors can be a bit of a bitch when you start bringing physics in

[DSE]

I am using novodex (now AGEIA PhysX). Its absolutely awesome and the API is very well thought out, also it has a nice plugin for 3ds max (3rd party) so you can export your convex meshes straight out of max into xml. Its then a fairly simple process to read that xml physics data (PML file) into the engine. I am working on high poly meshes for the renderer and I have created a set of very low poly models for the physics. Sorry if this is obvious, but its all fairly new to me and I am really pleased its working so well. Just putting the finishing touches to pits and pressure pads, so not long now !!!!

[Tom Hatfield]

Just to clarify, I'm afraid you are mistaken. HL2 does in fact use per-pixel lighting. That's what normal mapping is. You cannot do normal mapping without using fragment shaders. (Bump mapping through the fixed pipe uses height maps, not normal maps.) The reflections are cube maps; the shadows are light maps; the specularity uses fragment shaders.

Normal mapping is wholly possible on a PS1.x card. Every game to date that uses normal mapping (Half-life 2, Far Cry, Doom 3, Chronicles of Riddick, a gajillion others) runs and looks exactly the same on a PS1.x card as it does on a PS2.0 card — except Radeons, of course, because they have higher quality shaders.

What I'm trying to say is, there's no reason not to support PS1.x cards unless you simply don't feel like writing the extra shaders . . . I mean, you've already done all this work, why not make everyone happy?

On another note, the physics engine looks sweet. Will objects behave realistically when you throw them at monsters? What about that ettin, does he push the junk out of the way when he wades through it? I'm eager to see this in action.

[Patto]

DSE, that is looking very good. Never thought I would truly like the idea of playing a dungeon master clone in true 3d. But if your game plays aswell as it looks, I think I may well be converted

Would definately love to do a bit of playtesting.

Machine Specs:
Athlon XP3000+
1.5GB DDR400 RAM
Geforce 6600GT 128MB Graphics (PS3.0)

So it should run alright. Keep up the good work.

Patto.

[DSE]

Tom, firstly good to hear from you . I really dont care for getting into arguments on forums, its a waste of my time . So here is my final statement on the whole HL2 thing -

All I was going off (other than my gut reaction on seeing hl2) is the fact that if you search in google with something like 'hl2 per-pixel lighting' you get a few pages stating that it doesnt use it and none that say it does (I thought that was the real debate between hl2 source and ids doom3 engine). Anyway, lets leave the entire topic because until either valve post on this forum or someone can come up with a definitive answer this could go on forever. Also, it doesnt really have any bearing on what I am doing, it doesnt really matter if they achieved xyz because I am going to release the demo as shader model 2 only at the moment, then I will look into doing the 1.1 version (ps2 is surely a minimum now, shader model 4 is coming out soon). FYI - I have tried compiling my current shaders as 1.1 and the errors I get are not that serious, so you are right it is possible and I will look into it more when I have everything else finished. There are a lot of 3D projects which are started and never get anywhere close to finished, I would really like to finish mine (at least 10 playable levels with combat, spells, secrets, puzzles, etc).

Can I also say that its weird how wide the spectrum is between peoples computers these days (amongst gamers), after posting on this forum and getting the general feedback about hardware I thought the average machine would be of higher spec. There are people still running 5200FX cards and on the other hand some nutter has dual 7800GTX's, its hard to know where to pitch it.

Yep the boxes and stuff do bounce off the creature and as he walks the boxes move out of the way, its really wicked to see it in motion.

Patto, I was the same, I really wasnt sure if it would work. But trust me, its works well and the 3D just gives it more atmosphere and enables me to create a more interesting world. Keep checking back, the demo is coming. Just looking at your specs, I think you may need more ram

Thanks for posting.

[MitchB1990]

Man this will be so awesome when its realeased. I'd be up for beta testing too.

AMD Athlon 3500+
ATI Radeon X700+ Graphics card
only 512 DDR ram though :-/

[DSE]

Your system will be fine (I was joking about patto having 1.5gb ram, on my test machine I have 768mb). Is that washington usa or uk ?

[MitchB1990]

US on the westcoast

[Tom Hatfield]

We may be confused on our definitions of per-pixel lighting.

I would suggest shooting for lower specs if you plan to cater to the people here. A lot of us have pretty good systems, that's true, but a lot of us — particularly those fans from back in the day — don't. I'm beginning to think you're making this clone not to function as a great game in its own right, but to flaunt the technology used in your engine. Really, if that's the case, we can buy a copy of FEAR and enjoy it even better.

I think you should consider supporting cards that aren't PS-compliant. Of course it won't look as good, and if that's your whole objective, then you can safely ignore what I just said. The key to success is not only that everyone wants to play your game, it's that everyone can play your game.

It really annoyed me when companies started releasing games that could only be played on hardware TNL cards (like Tron 2.0), because I was still using a TNT2 Ultra at the time — immediate precursor to the original GeForce — which I knew damn well could run the game if only the programmers hadn't been short-sighted. Bad marketing move in my opinion.

That aside, are you planning to use the physics engine for anything other than flair? Because you have a very powerful tool at your fingertips, something that could make for some very flexible DM-style puzzles. I remember hearing complaints about Deus Ex 2, and how the game had this wicked awesome physics engine, but it wasn't actually used for anything in the game. Stuff just bounced around and fell over, and it looked good, but people were like, "why bother?" And then HL2 came out and blew everyone out of the water.

[Guest]

We may be confused on our definitions of per-pixel lighting.

DSE - ok, right, lets leave that discussion.

I would suggest shooting for lower specs if you plan to cater to the people here. A lot of us have pretty good systems, that's true, but a lot of us — particularly those fans from back in the day — don't.

DSE - I guess I have been a bit spoilt by my hardware, I will release it how it is and people can make their own minds up. My problem is that I love the latest gadgets/hardware AND really miss the game play of the older games that I used to play for a constant 28 hours straight (only stopped to go to the loo or because I was too weak to use the keyboard and mouse). My idea of an awesome DM style game would be lovely visuals with the original game ideas, stories, etc coupled with physics and more atmosphere. I wont achieve all of these, but I would like to give it ago and see how playable the end result is.

I'm beginning to think you're making this clone not to function as a great game in its own right, but to flaunt the technology used in your engine. Really, if that's the case, we can buy a copy of FEAR and enjoy it even better.

DSE - Have you ever tried getting up on the other side of the bed ? Can you not do both, it will be free ? IMO FEAR really wasnt that good, bored the piss out of me. Technically great of course, lovely per pixel lighting

I think you should consider supporting cards that aren't PS-compliant. Of course it won't look as good, and if that's your whole objective, then you can safely ignore what I just said. The key to success is not only that everyone wants to play your game, it's that everyone can play your game.

DSE - I totally 100% agree, but I still dont have the time/skill/energy/urge to do this.

It really annoyed me when companies started releasing games that could only be played on hardware TNL cards (like Tron 2.0), because I was still using a TNT2 Ultra at the time — immediate precursor to the original GeForce — which I knew damn well could run the game if only the programmers hadn't been short-sighted. Bad marketing move in my opinion.

DSE - Sounds as though they were a bit like me, more interested in if they could do it than if they should do it. At least my game is going to be free to download and play.

That aside, are you planning to use the physics engine for anything other than flair? Because you have a very powerful tool at your fingertips, something that could make for some very flexible DM-style puzzles. I remember hearing complaints about Deus Ex 2, and how the game had this wicked awesome physics engine, but it wasn't actually used for anything in the game. Stuff just bounced around and fell over, and it looked good, but people were like, "why bother?" And then HL2 came out and blew everyone out of the water

DSE - Yep, I am already thinking of some. I am also open to suggestions if anyone can think of any.

[Tom Hatfield]

Releasing it for free does give you a great deal of poetic and technical license. People can make suggestions, but we can't really demand anything. I appreciate the effort you're putting into this because it means other people will pick up the ball and build maps for your engine, provided you release some editing tools (or at least the file formats).

Just a quick question: can the physics API you're using do cloth simulation? Because it would be wicked awesome to have some interactive drapes, tapestries, tablecloths, etc. Creatures walking through would push them aside, and the cloth would drift and billow when you threw objects at it. I suspect you could incorporate wind, too.

As for puzzles . . . I'll get back to you on this, but if you did release some map-making tools, I'm sure the community would throw together some amazing things in due time.

[DSE]

This will be the last update to the render engine ... I hope. Anyway, you can all blame Tom for this one ... he mentioned ages ago that the walls looked a bit flat, well I have finally got parallax occulsion mapping working. See what you think.

I am now trying to finish the physics engine, but I have a number of small issues to fix before I can get on with combat.

Cheers,
Dave.

Tom - I think the physics engine can do cloth and fluids ... I need to get the combat done next then I will look into it. Probably after the alpha comes out.

[Tom Hatfield]

Hey, I just gave him the link.

[Gambit37]

Wow, those latest wall shots look muc, much better. This is shaping up to be impressive and could be the first 3D DM clone that actually feels a bit like the original while offering something new. Kepe it up!

One thing that is a bit weird in the new walls is shown in particular on the bottom left. The stones here do show highlights on their outer edge, but the deeper part of the bricks that is more hidden by other bircks is actually *lighter* -- it's quite a pronounced line that shouldn't be there. What causes this?

[DSE]

Yeah its not perfect, its the best I can do . There are even better techniques for doing self shadow which look absolutely awesome but I dont really understand them Out of interest I looked at the nvidia composer fx and it suffers from the same problem, I think it may be a by-product of making a 3D representation from a 2D texture (althought the irlicht engine has a parallax shader which is perfect, but again I dont understand it - just call me thicky ). The self shadowing would remove it though, maybe I will look into this after the other bits are done .

I think I just want to get the character sheets on the way, and then obviously I want to finish the combat and spell system. What are your thoughts (anyone) on the number of stats you should have in a game like this ? I think there two options

Basic - health, strength, mana, intelligence, ac, dexterity with character levels
Bonkers - something along the lines of advanced dungeons and dragons manuals

D

[Gambit37]

Re: Stats:

Keep it simple. The strength of DM was in it's simplified character sheet that made sense to mainstream gamers who weren't specifcially into RPGs with a lot of number crunching.

[DSE]

I agree !!!

Big shout now to any artists who fancy getting involved .... could someone come up with a character sheet like the DM one but in higher res and maybe a little more up to date (sacrilegious )

Any help much appreciated !!

[Suule]

0.02 $ from me:

Include be sure to include 'hidden skills' and luck. I liked that some skills required certain hidden skill level to become available...

[DSE]

will do. just to make sure we are on the same wave length, what sort of hidden skills are we talking about ?

[Suule]

The regular 'hidden skills' that DM had. Four for each 'class' + luck statistic. Concerning that they determined which commands could be used i think it would be great to include them.

Useful trivia: DM2 Hintbook actually gives them names.

[Trakl]

Since EFDM is still stuck I'm really looking forward to your project. Will there be a playable demo available this year?

[DSE]

Definately ... just keep checking back, I will be posting up some new screen shots soon to keep you all up-to-date .

thanks for your interest.
Dave

[Riviera]

Hi,

I am james, a student at queens in northern ireland doing computer science, currently i am working at microsoft in dublin for my years work placement. I am a great programmer and i know a fair bit about directx.
I implemented terrain with dynamic lod. I have done research in the area of shadows, bump mapping and data structures. I havent had chance to implement any of my theory but i understand most of it. I have also made a dm java clone in software, i havent released it, it never got finished. I also know c# and c++. I would love a chance to help you complete this game. I bet there would be alot of mundane programming for gameplay (example: item classes).

My emails are
h1312903@qub.ac.uk
t-jamesh@microsoft.com
rivierakid33@hotmail.com - i only use this for msn.

Peace out.

[DSE]

Thanks for your interest but the majority of the code is already written, my main problem is now coming up with interesting shaders for spells like fire, lighting, smoke, clouds, etc. If you fancy having a go at writing any of these, no problem ... give it a go

There are some new screen shots on the website, I am not totally convinced that the widescreen format works ... feedback welcome.

later.