Not logged in, Join Here! or Log In Below:  
 
News Articles Search    
 


Submitted by David Hunt, posted on August 20, 2001




Image Description, by David Hunt



Here are a couple of shots from our game which is in mid development.

The game is called Ghostmaster - the idea is you've got to scare people from their homes - the more you scare them the more energy you get the better the hauntings you can command.

Our company is Empire and our team is called Sick Puppies, we're based in Oxford England.

The game features fully photon mapped pre-lighting. What this means is that we generate our light maps by simulating photons flying about the scene from the various light sources. Each time a photon hits a surface it splats a bit of it's energy onto the surface and some of it is reflected. Eventually if you leave the compiler running long enough you get wonderfully warm, diffusely lit scenes which look unlike anything you've probably seen in games before. A level has to be lit over night to get good quality.

For more shots see http://uk.geocities.com/ghostmaster_stuff/ - these are all in game shots, not pre-renders and the game is running at speed.

David Hunt


[prev]
Image of the Day Gallery
www.flipcode.com

[next]

 
Message Center / Reader Comments: ( To Participate in the Discussion, Join the Community )
 
Archive Notice: This thread is old and no longer active. It is here for reference purposes. This thread was created on an older version of the flipcode forums, before the site closed in 2005. Please keep that in mind as you view this thread, as many of the topics and opinions may be outdated.
 
Morlith

August 20, 2001, 05:32 PM

excellent! very, very beautiful lighting!

 
Jan Niestadt

August 20, 2001, 06:20 PM

Amazing pictures, but I agree that often 2d games are just as fun as their 3d counterparts.

The problem is not game developers, by the way - it's publishers. A lot of game developers would probably love to keep doing 2d, but it's nearly impossible to get a game like that published anymore.

 
nonamez

August 20, 2001, 06:24 PM

"A level has to be lit over night to get good quality"

wow ! But I guess you could take challenge and do it in real time ? lol

Its really nice images, but it probably would take 3 nights (and days ..) to calculate lighting in outdoor environment. (or even more time)
BTW what resolution are your lightmaps?

 
Matt Newport

August 20, 2001, 07:20 PM

Hey! Another game company in Oxford. Where abouts are you based? I work for Rebellion (down near the Westgate centre) - I know that UbiSoft have a team in Oxford but hadn't heard about Sick Puppies. We'll catch up with Guildford for highest density of games companies in the country at this rate :-)

 
Matt Newport

August 20, 2001, 07:21 PM

Great IOTD by the way!

 
Crocmort

August 20, 2001, 08:32 PM

Greatest game idea I've seen after Black and White... =) The shots are really great! But what about AI in the game? I hope you'll do people with personnalities, like kids who react diferently than adults ect. Do yo have an official site? I saw the one with all the screenshots, but their is only screenshots... =P I want game details! =) Myabe it's too early in development... =)

 
Elixir

August 20, 2001, 09:29 PM

I remember playing some Isometric game for the ol' Sega where you were supposed to scare people in some time limit. I didn't like the game because it was timed (never gives me time to explore =( ).

Anyway, the lighting looks incredible, if not unsurpassed currently. Keep up the great work! =]

 
vrempire

August 20, 2001, 09:48 PM

I love your lighting effect and ghostly surrounding..
Can't wait to buy it...
Can't wait to play games with excellent virtual reality environment.

 
337

August 20, 2001, 10:14 PM

Thanks a bunch for your detailed explanation of photon mapping, I appreciated it.

 
SirKnight

August 20, 2001, 10:16 PM

I went to your website there and your game looks very moody. I love your lighting, looks very cool. The smoke you create in some of those shots looks very good. That scene with the spiral staircase and the fountain in the middle of the room there is probably my favorite shot. It looks like a scene from a movie i saw once. But why isnt that fish on the top of that fountain not spewing water out of its mouth? To me it looks like it should be. But the water effects in the fountain looks cool anyway. Can't wait for a demo.

-SirKnight

 
Pseudo_Me

August 20, 2001, 11:33 PM

nah, that's just a 4d projection onto the 2d image. Then again, 3d is just 3d projected onto 2d.

 
James Matthews

August 21, 2001, 12:04 AM

I've gotta agree with SirKnight, that staircase shot kicks ass, although this shot is also incredible!

Is the smoke procedurally generated? Great idea and beautiful graphics - makings of an excellent game, keep it up!




 
RCMD

August 21, 2001, 02:01 AM

One of the most beautiful IOTD for a long time. The lighting is incredible, seems to me to generate a very good atmosphere for the game.

 
LimpnoodL

August 21, 2001, 02:04 AM

The game idea reminds me of the old playstation game by TECMO... called Tecmo's Deception...'twas one of my favourite PS games...

 
psykotic

August 21, 2001, 03:19 AM

From rereading your description of the lighting system, this sounds more like light-driven path tracing than photon mapping. On another note, why are you not using radiosity? Only view-independent (i.e. diffuse-diffuse) light interactions can be stored in light maps so using radiosity would produce better results faster especially if augmented with discontinuity meshing. Photon mapping and path tracing are more useful in solving the complete rendering equation and this includes diffuse-specular, specular-diffuse and specular-specular transport which cannot be stored in light maps. I am just curious =)

 
Matt

August 21, 2001, 05:27 AM

That's what I was thinking...
Though you would get caustics etc (specular-diffuse).

 
David Hunt

August 21, 2001, 05:39 AM

The idea is fairly simple. You control the ghosts - the idea being to clear the level of mortal creatures.

We've spent quite a while tweaking the interface and we've ended up with a sort of souped up sims view of each level (not isometric obviously). The buildings/levels are split into cross-sections - one for each floor - so you can scroll up and down with the roller ball and the floors magically appear/disappear. Each room can be highlighted and haunted or objects within it can have apparitions of various types attached to them. You start off with pretty basic tricks but as time goes your repertoire increases.

The more you scare people the more psychic energy you get - some characters are more sceptical than others and will refuse to see any of your psychic endeavours. Other things will also act to calm the characters down such as priests the FBI and exorcists.


David Hunt

 
Raspberry

August 21, 2001, 05:43 AM

how did you use the moon to light stuff?
or did you hard code an ambient :(

do you have light volumes? i.e. a bulb is a sphere that randomly projects photons?

i also realised that you can probably run X number of computers all with different lighting seeds at the same time, and add all the photon maps together before you do the second phase... (sorry, but i love parallel computing).

makes me wanna setup a raytracer/radiosity/photonmapping system...

:)

 
ShinMan

August 21, 2001, 05:50 AM

Hello people. I'm on the Ghostmaster team and I am responsible for the lighting code, so I'm probably well placed to reply to a few bits and pieces, although Psykotic has been handing out pretty good info on photon mapping...

Matt Newport...
We are based on Osney Mead industrial estate off Botley road. Hmmm..
Rebellion... you may know Neal Tringham then...

"yes, you get ALL lighting effects with photon maps"
Just to back this up... I sample texture colours for photon collisions and include alpha in the transmit/refelect/absorb probability distribution. There is a shot on the site that shows light coming through stained glass windows projecting onto the ceiling.


"The shots are really great! But what about AI in the game?"
Trust me, the whole AI thing is well in hand. You might even get our AI guy on here. I get to worry about making it look nice. I can trust other people here to make it play nice....


"Its really nice images, but it probably would take 3 nights (and days ..) to calculate lighting in outdoor environment. (or even more time)
BTW what resolution are your lightmaps?"
We do have outdoor environments. They don't seem to take any longer to get decent results. The default lightmap resolution is 0.3m, but we can reduce low detail lightmaps or even drop back to vertex lighting depending on the specific circumstance.

"this sounds more like light-driven path tracing than photon mapping. On another note, why are you not using radiosity?"
It is more like path tracing. Photon mapping sounds better for the PR people.... As for not using radiosity.....
1 - It is easier to code a path tracing system.
2 - It is easier to modify a path tracing system to include new lighting phenomena - eg 'shaped' lights, volume lights etc, coloured transmission.
3 - The same code could be used for generating 'hacky' GFX, like for offline processing of a caustic texture. Just because its diffuse-only now doesn't mean I won't want to do some view dependant stuff later.

I mean, come on... I could worry about form factor calculation, patch subdivision etc,etc, or I could just knock up a BSP tree a fling a few billion photons around it, and then tweak it to do all sorts of extra stuff...

EG---
Light sources are diffuse spheres of light rather than absolute point sources. Directional lights have a probability distribution of directions about the mean. Shadows cast by these lights are hard near the caster but soften with distance.

"who do you generate the artwork? with a package like 3ds or maya?"
Our artists, God rest their souls, are using Softimage for both character modelling and animation and the level design.

Andrew Jones
Sick Puppies.
Lighting and graphic effects programmer.




 
David Hunt

August 21, 2001, 06:00 AM

The fire is procedural. This will eventually be handled the way most people handle switchable lightmaps. Ie. we seperate each light and compress it's maps, when a light level changes we rebuild the affected lightmaps. In our case the math is slightly complicated as we use a slightly non-standard attenuation curve to give more response at low light levels. We're also hoping to link special light maps to specific events such as doors opening.


David Hunt

 
David Hunt

August 21, 2001, 06:06 AM

We use SoftImage and some plugins we've written to attach attributes to objects exporting to XSI format and our own custom compiler.



David Hunt

 
David Hunt

August 21, 2001, 06:18 AM

We simulate various light types.

Point lights are infact random vectors from a ball with a small radius. Ambient is a large globe surrounding the whole world (hence we don't get actual ambient inside the buildings.) We also do spotlights and shaped spotlights (uplighters and so on) and are currently working on a method of passing indirect illumination points through to the runtime lighting engine. eg. The spot on the ceiling or wall from an uplighter. However accuracy isn't our raison d'etre (sp?) - we're more interested in ambience, warmth, cold and hostile.


David Hunt

 
shrike

August 21, 2001, 06:21 AM

hard near the caster, softer with distance. you must mean the other way around, unless there is some weird problem with your light map resolution. maybe i just read it wrong.

 
David Hunt

August 21, 2001, 06:28 AM

Osney Mead - near the station - used to be Iffley Rd.

Also Mathengine (or whatever they rebranded to) are in Oxford, and Empire has another team in Kidlington called Razorworks (Apache Havoc). We've got Neal Tringham (one of your boys) working on the project!

We're trying to get a go-karting team tournament going after ECTS if you're interested (talk to me off list).


David Hunt

 
David Hunt

August 21, 2001, 06:35 AM

We're going to start gearing up the promo stuff soon - thought I'd get some peer-response first tho;)

We do have AI - affinity aversion and internal scripting - this is one of the areas we're trying to nail for ECTS ATM. The differences in the way the characters react is one of the keys to gameplay - the more they get freaked - the more resources you get to allocate - you've got to milk the freak.


David Hunt









 
David Hunt

August 21, 2001, 06:38 AM

"It looks like a scene from a movie i saw once" ;)

Strange that - most of our scenarios are supposed to be evocative or reminiscent of well known celluloid moments.


David Hunt

 
ShinMan

August 21, 2001, 07:06 AM

Well.... erm..

The shadow caster being the object which is casting a shadow.
For example, if you have a streetlamp casting a shadow,
then the shadow near the base of the streetlamp will be
hard, and the shadow will blur and loose definition further
from the streetlamp.

Andrew Jones

 
Ewan Higgs

August 21, 2001, 09:26 AM

Looks great! Your software group might want to look into "ibriQ"s

www.totalimpact.com

If you want to have a level be path traced overnight, you could use one of those to do it all day too but not forcing you to give up your workstation to let it process. If psykotic is correct (I trust he is..) and can be infinitely accuracte as time permits, you could let these things fly from now until the day when you need to send the CD off to the publisher and it would be pretty accurate! ;)

 
psykotic

August 21, 2001, 11:00 AM

1 - It is easier to code a path tracing system.

Yes, I agree. One of the problems with light-driven path tracing is that you need to completely finish the path tracing process or your lighting will look like piss. You will get irradiance gradients that will make the most hardened dopehead run away screaming. Solving the radiosity equation with iterative methods (the most common scenario) will progressively give you more and more indirect lighting. But if you choose to stop the solver after one iteration you will still have direct lighting which at least does not look completely unreal. It should be noted that photon mapping and irradiance caching helps with the above problem.

2 - It is easier to modify a path tracing system to include new lighting phenomena - eg 'shaped' lights, volume lights etc, coloured transmission.

Radiosity has no problems with 'shaped' lights (I assume you mean emitter meshes). What is usually termed 'volume lighting' is in fact light passing through participating media such as fog. Yes, this is much easier to do with path tracing. You simply do a line integral over the ray on a scalar field which is the superposition of all participating media fields. I cannot see how colored transmission is a problem for radiosity too, unless I am mistaking something here.

3 - The same code could be used for generating 'hacky' GFX, like for offline processing of a caustic texture. Just because its diffuse-only now doesn't mean I won't want to do some view dependant stuff later.

True, I made a small mistake in a previous post. Specular-diffuse transport (caustics) and any other light transport with a difuse destination can of course be stored in a light map for view-independent use as long as the (perfect) diffuse and specular do not move.

I mean, come on... I could worry about form factor calculation, patch subdivision etc,etc, or I could just knock up a BSP tree a fling a few billion photons around it, and then tweak it to do all sorts of extra stuff...

Come on, radiosity is not that hard. Just like getting a simple ray tracing working is much simpler than coding a software rasterizer, getting a path tracer to 'work' is much simpler than a similar radiosity system. A lot of people care about processing times in relation to correctness of the solution. While ray tracing methods are generally assumed to scale better to large scenes, I still think a radiosity processor is better and faster (assuming you only need to simulate diffuse-diffuse light transport) if you partition the scene into sectors and do visibility computations on a more macroscopic level before performing patch to patch visibility queries. Anyway, I am very impressed by your graphics engine and you obviously know what you are doing.

 
Revolver

August 21, 2001, 11:18 AM

Blah blah blah. :)

 
This thread contains 76 messages.
First Previous ( To view more messages, select a page: 0 1 2 ... out of 2) Next Last
 
 
Hosting by Solid Eight Studios, maker of PhotoTangler Collage Maker.