[Framework ? not Framewerk?]

I think we're over reacting a little bit here with the framewE/Ork change. LW is going to go through a lot of API change in the near future what with this merge and the integration of LUA and all that, so we might just have to prepare to spend some time performing code change in our project timelines. Personally, I'm fine with that - Its more work but I signed up for clean code and consistency when I bought LW. So change is expected, and I knew that from the start.

[LUA: How to animate models]

I would do your animation in the object:Render() function. That way animation only has to be performed on models that actually get drawn.

 

Hrm, are you able to point to an example of that function in use Josh?

[Important Update]

Finishing the framework docs today...

 

That's even better news!

After all this crazy new stuff in 2.3 I think we need some decent doco.

[Important Update]

Fantastic work Josh. Nothing better than consolidating codebases to make things more efficient and simpler.

[Very jerky movement at high framerates]

Okay. You guys might want to post full config details - resolution, SSAO/HDR/etc.

[Very jerky movement at high framerates]

This is just a complete hunch, but if you're using KeyHit() / KeyDown maybe the keyboard isn't being polled quick enough. What I mean to say is, KeyHit() captures the key that was hit and then resets afterward. So at a speed of 100FPS, you might be finding that some frames turn up an empty KeyHit() because it hasn't polled the keyboard yet.

[A few thoughts after playing Oblivion]

One thing I learned while watching an interview with the makers of FUEL, is the roads are procedural as well. They also said the game would have been 20GB bigger instead of 1 GB bigger, if procedural maps were not used. So it looks to me like the artists would have no control over map layouts, unless they were allowed to use 20X more space. B)

 

Ah but they do. Clearly the game must have some key places and areas in the game. So they tell the procedural engine to put this and that in specific places each time and then let it work out the bits in between on the fly.

[Nvidia 3D Vision]

Yeah I tend to agree with Niosop. PhysX is an Nvidia exclusive feature and now that quite a damn few games use it, I do make the concious decision to buy Nvidia over AMD.

And as for CUDA, thats a whole new concept Nvidia are hoping will really kick off in datacentres in the future.

I think Josh could be right in that thats what Nvidia aim for at the start, but both PhysX and CUDA will, in the future, make them money, all things going to plan.

[A few thoughts after playing Oblivion]

I wasnt aware of the fact theres id-tech6 already. No game has been relased with tech5 so it's a bit early to talk about tech6.

 

EDIT: Took a look at the .pdf from your wikipedia link and its refering to tech5. No tech6.

Ah you are correct Michael. A typo on my part. Still, you appeared to be talking about the MegaTexture implementation in idTech4 (Doom 3, Quake Wars), while idTech 5 (RAGE) takes it to a whole new level.

 

I would also bet a procedurally generated environment would not have as good a hand for balance as a human would.

 

I may not be getting what you mean by 'balance' but I disagree with you, for the most part - We have terrain generation working perfectly now. You'd be crazy to try to hand paint your own terrain (heightmap) when completely realistic ones can be gen'd at the click of ones fingers. Of course, I agree that one will need to come in a touch up and personalise the terrain afterwards.

But I think thats where games will continue to head in the future. More and more will be generated on the fly, which allows us developers to concentrate on more important assets or parts of the world we're building.

[A few thoughts after playing Oblivion]

Well, for terrain at least it makes sense. I bought Fuel. What an amazing game. Can you imagine if they had made an FPS with that?

 

haha. Gee you've really jumped on it eh. To be honest I don't think an FPS would have worked out at all. You'll see this the more you play - By the games design you are flying through the world at high speeds. So environment quality isn't a big issue. Thats not to say Fuel isn't beautiful, because it is. But when you're on foot in an FPS, its an entirely different matter. The world has to be much much more detailed. I think it could very well be done, but it'd take a hugely detailed procedural system to build up an interesting world to an FPS player level. It would require so many layers and passes.

I'm fairly certain the code is out there in pieces - Trees, grass, terrain, roads, caves, etc. It just all needs to be put together. I'd really love to see an open source project do sometihng like that. I wonder if there is already a project out there?

[A few thoughts after playing Oblivion]

The comments in the roads section mention procedural cities. The Pompeii city in this link is stunning. Its a pity about the price of the software though.

 

http://www.procedura...es/pompeii.html

 

Wow that app is very impressive!

I think procedural environments will become a bigger and bigger part of the gaming industry. We just can't keep on going the way we are - Takes too long to make decent large worlds at the moment. We need better tools.

[A few thoughts after playing Oblivion]

Wow, Fuel is awesome. I downloaded the demo. Can you travel across the whole world in the full game?

 

Aye. There is a lot to be said about procedural graphics. Fuel is a perfect example of how procedural graphics can really make your game unique. There is simply no way anyone could justify building such a huge world by hand, and the fact is, a game like Fual just would not exist without the huge procedural engine they have under it.

 

Josh, you should really check out some of Shamus's other articles on Fuel. I know you yourself have done quite a bit on roads, so I know you'll be interested in em:

1. Fuel: Roads - http://www.shamusyoung.com/twentysidedtale/?p=5237

2. Fuel: Terrian - http://www.shamusyoung.com/twentysidedtale/?p=5267

[A few thoughts after playing Oblivion]

Wait wait. You're talking about MegaTexture. That tech is the precursor to the iD Tech 6 technology.

 

Wikipedia Quote:

id Tech 6 will use a more advanced technique that builds upon the MegaTexture idea and virtualizes both the geometry and the textures to obtain unique geometry down to the equivalent of the texel: the Sparse Voxel Octree (SVO). It works by raycasting the geometry represented by voxels (instead of triangles) stored in an octree. The goal being to be able to stream parts of the octree into video memory, going further down along the tree for nearby objects to give them more details, and to use higher level, larger voxels for further objects, which give an automatic level of detail (LOD) system for both geometry and textures at the same time. The geometric detail that can be obtained using this method is nearly infinite, which removes the need for faking 3-dimensional details with techniques such as normal mapping. Despite that most Voxel rendering tests use very large amounts of memory (up to several Gb), Jon Olick of id Software claimed it's able to compress such SVO to 1.15 bits per voxel of position data.

 

The main drawback of the Sparse Voxel Octree is the need for fast updating of the octree in order to represent dynamic objects. However, Jon Olick gave examples of alternatives which would not require this, but cautioned that their use would probably be better suited for id Tech 7. For id Tech 6, SVO will therefore be used for representing static geometry such as terrains and buildings. Dynamic objects such as vehicles and characters will be represented by rasterized polygons as is the case in most 3D games today.

 

 

Check out this presentation: http://s09.idav.ucdavis.edu/talks/05-JP_id_Tech_5_Challenges.pdf

[logical entities and their "models"]

But won't then the linepick fail to grab whatever is behind the light? With your code Macklebee, it'll collide with and dismiss the entity, but it'll act as an invisible wall for objects behind it no?

[A few thoughts after playing Oblivion]

Why not go further than the Stalker people and create a map system that can wrap around onto a planet or asteroid shape? If this could be put into orbit with other planets, it would be tres cool. I am joking because it sounds ridiculously difficult, but if you came out with an engine with that capability, it would blow the competition away.

Already doable no problems. Prey did that - Jumping between asteroids of medium sizes. The gravity was great fun to play around with. Garys Mod has a few levels with the space plug where you can fly ships around a fairly large level full of worlds. Not as nice as Prey, but definitely cool.

 

I think some of theese unlimited and non repeatable texture discussion is overrated.

A lot of games use repeating textures. Lets take Crysis or Cod4. With some decals and good cover of the terrain it's not visible.

 

Using models for terrains entirely is a pain in the wonderful person if you need to do small changes. And you are limited in size. It depends on what you want to do. A fast paced shooter doesn't need mega texturing because all stuff pases by in a fast way. In a wide open level with only some vegetaion it would be nice to have such a tech. On the other hand if you do a open desert level you probably will have lots of stones to break it up. New Medal of Honor will take place in afgnahistan. Im curious how they will handle the "old texturing style" with their setting. First screenshots are very cool to look at.

 

So, yes, it would be nice to have more terrain layers especially in large scale levels with 4kx4k sizes so you can have several kinds of environment in one map like a transistion from desert to woodland.

For example Battlefield 2 is limited to 6 texture layers. And there are nice maps out there too.

 

First go and fill up a 4k level with all the stuff and then ask for (world)streaming. IT takes forever to walk through such a large level. Even if you do a RPG you can add different quests where the player has to cross parts of the level. Plus you can add vertical geometry and caves/dungeons. There are so many possibilites to keep the player busy in a non-world streaming world.

You make some good points. But my main point, and I think one of the main attractors to iD's solution is this:

By having this one mega-texture, iD have developed a real-time editor where artists can just go in and literially paint out the landscape. You can place down and merge in as many layers as you want because in the end it'll all become one big texture. I believe the editing tool is capable of working on the same 'world' in parallel with other artists as well. So oyu could have a team of peeps painting away in realtime in the same world. The advantage here is that it could cut production times by an astronomically huge amount.

Of course, its a completely new concept and there have got to be some disadvantages. But the bigger we make our worlds, the production costs in time and resources continue to exponentially increase. I mean its getting to a point even now where most large worlds look noticably crappy because they simply cannot spend the time they would like on all aspects of the surface.

And I think this'll first come to light when we see RAGE released. I think we'll all see the subtle but massive impact this mega-texture technology could have on a large landscape.

[A few thoughts after playing Oblivion]

Terrains in id tech are just a mesh made in a modeling program, so the geometry is very limited in size.

 

*shrugs* I was talking more about the texture itself, not the geometry. My guess is that geom/polys will be Carmacks next big project after this.

[A few thoughts after playing Oblivion]

From the technical specs I have read of RAGE, it doesn't even allow terrains as big as what we have.

 

Well I think your comparing two different techs here Josh. Your terrain is just that, a plane of terrain. RAGE uses one giant texture for EVERYTHING. Buildings, roads, the lot.

As far as I know you can technically make the size of the texture as large as you want. Your only constraint is space to store the thing.

[A few thoughts after playing Oblivion]

I don't think its too far off. Carmack's Rage is about to pretty much remove any problems to do with textures. IMO the biggest hurdle is us, the humans. The bigger our worlds get the longer it takes to build them or the more watered down they become. I think the answer lies in perfecting procedural environments. You guys should check out the game Fuel - Its a truely HUGE world and its pulled off completely procedurally. Check out this article: http://www.shamusyoung.com/twentysidedtale/?p=5134

[Linked In]

Aye, I've been on LinkedIn for a few months now. Its pretty good, and I could see how it could be a great tool to progress your career.

Only thing I don't like is that some people seem to think the aim is to rack up friends. You're only really meant to add people you've worked with seriously and you know you can count on to vouch for you. For that reason I won't be adding anyone here and won't expect anyone to add me. At least until I produce something worth-while and get to know anyone.

 

Lumooja - LinkedIn is not really the same thing as a blog. I agree with you for the most part, but think of LinkedIn as a dynamic and professional resume/CV. Its a priceless tool once you've got some decent contacts and made a few impressions.

[Intel's Larrabee Cancelled]

No. Thats 70% (or whatever the real value is) of the INTEGRATED market. Larrabee is targets at the actual graphics market. Intel wants that, wants it bad.

[Intel's Larrabee Cancelled]

I realise that. But I don't doubt for a second that they'll stop. They can't stop, not if they want to catch AMD/Nvidia any time soon.

[Intel's Larrabee Cancelled]

No, if I went posting e-mails around I wouldn't be extended this privileged anymore.

Fair enough. Well you'll have to excuse my skepticism - I very seriously doubt Intel is out of the game in terms of consumer space graphics. That notion just seems completely wrong in so many ways.

[Intel's Larrabee Cancelled]

I e-mailed the Intel marketing manager who was heading up product management on Larabee, and she said that they were no longer pursuing Larabee as a consumer product, period.

 

Well that's certainly interesting - The graphics market is simply huge and growing and I seriously doubt Intel are just going to let that be. Do you mind if you post the email here? I'd like to see exactly what was said.

[Intel's Larrabee Cancelled]

DJDD i think you will need to read all the other news on the nets not just bit-tech

as its delayed and part of it is canceled

 

so there is no enduser version of the card coming out only the pro side

which means a lot of us will not be able to get the card

 

Thats my point carlb. Josh, and that news poster at Bit-Tech make it sound like Larrabee is dead. It is simply delayed. No more.

[Intel's Larrabee Cancelled]

Way to take the quote completely out of context Josh. The real quote is:

"Larrabee silicon and software development are behind where we hoped to be at this point in the project. As a result, our first Larrabee product will not be launched as a standalone discrete graphics product."

 

Please note the highlighted keyword.