Building Pop Up

[Heliocon]

Dont pull the - graphics cant handle it bs. thats a lie.

If they had an iota of sense they would make building at a distance simple boxes with roofs, and have a very low res texture on them. As you get closer the LOD increases, and then they either have more LOD switches for less and less detail, or tesselate the buildings based on a second model from the first LOD switch. This means there would be no popping at all, and only 1 real model would need to be made (and houses are really easy to make, its what you do as a first project, however towns are a bit harder and I dont know what exact system they are using for that...).

In any case LOD popping completely eliminates immersion. I cant believe the rediculous opinions on this board when people obsess over the contrast or brightness of the terrain/land, yet the fact that buildings suddenly spawn into the word where before there were none is hardly argued over/mentioned. Its one of the MOST important graphical features, and can easily be handled by any modern computer (even low end ones). The amount of rescources needed to load buildings at a distance into memory is near nothing, they have very little geometry and at a distance low textures (which should increase when you get closer, but no point in highres textures if you cant see them!).

More and more there engine is coming off as inneficient, badly designed and planned it seems to me. Who knows it might be fine in release, but with the current pop it looks absurd.

[CharveL]

Originally Posted by Heliocon

Dont pull the - graphics cant handle it bs. thats a lie.

If they had an iota of sense they would make building at a distance simple boxes with roofs, and have a very low res texture on them. As you get closer the LOD increases, and then they either have more LOD switches for less and less detail, or tesselate the buildings based on a second model from the first LOD switch. This means there would be no popping at all, and only 1 real model would need to be made (and houses are really easy to make, its what you do as a first project, however towns are a bit harder and I dont know what exact system they are using for that...).

In any case LOD popping completely eliminates immersion. I cant believe the rediculous opinions on this board when people obsess over the contrast or brightness of the terrain/land, yet the fact that buildings suddenly spawn into the word where before there were none is hardly argued over/mentioned. Its one of the MOST important graphical features, and can easily be handled by any modern computer (even low end ones). The amount of rescources needed to load buildings at a distance into memory is near nothing, they have very little geometry and at a distance low textures (which should increase when you get closer, but no point in highres textures if you cant see them!).

More and more there engine is coming off as inneficient, badly designed and planned it seems to me. Who knows it might be fine in release, but with the current pop it looks absurd.

Dude, just...go read up on the stuff before trying to make comments about things you don't really get. I mean, you know some terms and what they're supposed to do but not how they work together collectively. You might as well give us your insight into nuclear physics since you read a wikipedia explanation about half-life (hint: not the game).

[TheGrunch]

He's still right, though. The solution to building popup is just to create even simpler LODs. How much are a few hundred 16 tri buildings with 16x16 textures going to tax a modern GPU? Not much.

[zauii]

I think Oleg and his team have a "slight" edge on the skills and knowledge required for the job than our Heliocon..
it might seem obvious in principle but truth to be told he has no clue how it actually works together and especially not in their environment.

I agree completely with Charvel.

[Heliocon]

Originally Posted by CharveL

Dude, just...go read up on the stuff before trying to make comments about things you don't really get. I mean, you know some terms and what they're supposed to do but not how they work together collectively. You might as well give us your insight into nuclear physics since you read a wikipedia explanation about half-life (hint: not the game).

Charvel, really stop before I start using caps : http://forum.1cpublishing.eu/showthr...795#post233795

I have little programming experience, but know a good amount about game design (from multiple sources, did a bit with Digipen) have a few years experience with Maya aswell. Now there are plenty of things I do not know about, and I always preface my post with that (for example compared to others here I know very little about flying and tactics, engines or plane mechanics, much of the wars history or even things like HOTAS setups). What I do know about is computer hardware and software, and enough about graphics design and engines that I feel confident in the comments I make. Like said I did a good amount of graphics design in Maya (7.0 then 7.5) which was bought out by Autodesk (the only other major competitor for CGI and game model/graphics) and while I would have to wiki to check now, I believe autodesk scrapped their autodesk app, and now Maya autodesk is really only used (maybe opengl is different, not sure on that one). Anyway I used the exact same software they are using for the game.

For the record the link is my "victory I told you so post", where I layed out many features and uses of DX11 and opinions on how it should be used/how they will use it along with comp hardware about 5 months ago-1month ago. Ironically during that time a dev himself said they were only thinking about tesselating "plane wheels" and forum members said dx11 decreased performance. I had been arguing for tesselation to reduce workload for modeling details and remove the need to LOD and multiple models to reduce popping. I also talked about using direct compute for physics modeling for flight and for water/particles, along with dx11 shader pipelines being much more efficient on mutli core machines and that they should aim for that. People like you dismissed it, just so you know- Then we had that nice new interview which nearly point for point repeated every single thing I had been saying (and getting responses like you made to me) for months and months (2010). Now I probably should not have dragged this out, but if you have a particular point I made in my post you would like to contest, thats fine with me. Until then, dont reply with that bs to my post, because I have a paper due and I am not in a good mood

Wall of text over, continue

[CharveL]

I won't pretend to be an expert either, hell only the developers really know where the bottlenecks are since there's often a big gap between supposed benefits of a feature and actual implementation. And that's kind of the point we're saying here. You were spouting off stuff that I'm sure makes a lot of sense to you and what you've gathered, but without taking into account that you don't have nearly enough knowledge in order to suggest the developers are incompetent, sloppy or lazy with such surety.

I have no problem with much of the negativity around here, Oleg's a big boy and can take it without crying himself to sleep at night, as long as it's accompanied by reasonable concerns. I know it seems only these types of posts tend to get a response so it's a valid tactic I guess when you want to stand out in a forum.

No offense, but some of the reasoning stuff you rattled off was just...silly...especially when you draw such conclusions from them. I do hope you get a response though because I'm pretty interested in the topic.

[Kikuchiyo]

Pop up is a legitimate issue, but I suspect that it will be something that will always exsist (or at least until computers are capable of rendering all things at all times simultaneously). the power of your PC will determine how noticeable the pop up will be.

[kestrel79]

Originally Posted by =XIII=Shea

Pop up is very annoying when bombing from high alt

This is what annoyed me most with pop up in IL2. I spend all this time in my B25 getting up to alt and then I cannot see the buildings I'm suppose to target until I'm right on top of it looking down the sight. By the time I move the rudder and let the sight settle it's too late.

I hope this gets improved in CoD.

[Blackdog_kt]

Heliocon there's no need to be aggressive. Most of us are intelligent enough to understand that DX11 will obviously be better than DX10.

That doesn't mean it's a fully mature technology yet, nor that it can be done in time (they will need to get used to working with it first to the detriment of other stuff that are equally important for a lot of potential customers, like for example spending time on the aircraft), neither that everyone has a DX11 GPU yet, so i don't really share your anguish.

Granted, it's a matter of taste and i tend to focus more on things like "how do the airplanes fly, is it accurate?" and prefer it when they focus on that kind of stuff, but i don't think delaying the game for 6 months to add DX11 support before release would mean much for you.

You still wouldn't have a DX11-capable CoD either way until 6 months later.

As for the processing load, Luthier said that it's not so much the graphics but the positional references that hog the system resources. I think he said that they could reduce each house to a pixel and we would still get almost the exact same FPS: the CPU would still have to track each of that pixels' positions relative to the player's aircraft and we all know that a PC is only as fast as its slowest component for the given task. Even if our GPU could render everything just fine, in all likelihood it would probably have to "synchronise" with the rest of the components and wait for the CPU before fusing the complete set of data to present on the screen.

[Heliocon]

Originally Posted by Blackdog_kt

Heliocon there's no need to be aggressive. Most of us are intelligent enough to understand that DX11 will obviously be better than DX10.

That doesn't mean it's a fully mature technology yet, nor that it can be done in time (they will need to get used to working with it first to the detriment of other stuff that are equally important for a lot of potential customers, like for example spending time on the aircraft), neither that everyone has a DX11 GPU yet, so i don't really share your anguish.

Granted, it's a matter of taste and i tend to focus more on things like "how do the airplanes fly, is it accurate?" and prefer it when they focus on that kind of stuff, but i don't think delaying the game for 6 months to add DX11 support before release would mean much for you.

You still wouldn't have a DX11-capable CoD either way until 6 months later.

As for the processing load, Luthier said that it's not so much the graphics but the positional references that hog the system resources. I think he said that they could reduce each house to a pixel and we would still get almost the exact same FPS: the CPU would still have to track each of that pixels' positions relative to the player's aircraft and we all know that a PC is only as fast as its slowest component for the given task. Even if our GPU could render everything just fine, in all likelihood it would probably have to "synchronise" with the rest of the components and wait for the CPU before fusing the complete set of data to present on the screen.

My problem with charvels response is that he never specified what he had a problem with, or disagreed with. The reason I said what I said is that if the devs statements of true it is poor planning. I dont think its "lazy" but they knew what they were getting into, knew what they wanted out of the engine and they should of planned it in a way that takes into account all these problems. What I see is a lack of foresight that can be crippling to a game engine.

Now as far as DX11 is concerned - I think they should of cut out dx10 completely and just gone dx11 and dx9, because dx11 builds on dx10, but is far far easier to work with, so the time they spent on 10 could of been spent on 11, and would of been completed far faster. Also it would mean that they would have alot less modeling work to do (although this depends on alot of factors). Now I think mechanics are as important if not more than graphics - but they keep saying they want fps fidelity, which is absurd because atm it has no where near the detail of an fps, and no where near the detail it should have as a flight sim either, because when buildings sprout from the ground it screams "cheap indie game". That should not happen at all in 2011, there is no excuse for it with modern hardware and tech. Now maybe on mid range systems it does not happen, but since they dont care to update their non russian community, we dont know.

The supposed CPU bottleneck on buildings is due to bad optimization and tbh makes no sense.
1. Its because the cpu keep track of every building, irregardless of LOD or distance and has to tell the gpu to render shadows and lighting interactions on each of them irrespective of if we can see the effects or not = Gpu bottleneck. If not and it is purely CPU then I have no idea what the hell they are doing because you dont need to track a houses position, the damn thing is stationary. Why do you need to track its position if it is not visible? Because it blocks another object? Thats crazy, who the hell makes maps / engines that way? Also adding to that there no reason why they would need to track stationary objects for anything BUT LOS rendering on the GPU. Computers can dome enormous amounts of calculations per second, but its like we are adding 1+1 and we get 33.33333x66.6666^infinity. It just doesnt add up and it should not be an issue in the first place...

-p.s DX11 has been around for two years now, its time to move on.