PDA

View Full Version : Ilya answers regarding CoD scaleability


TheSwede
01-27-2011, 10:12 AM
I just logged in to Facebook and threw away a comment/question regarding how the new game engine will respond to future hardware configurations.

For those registered on Facebook
http://www.facebook.com/il2sturmovik/posts/149236541796325?notif_t=like


For those without:
" Simon, the game is indeed extremely scalable. For the past six months all we've been doing is scaling it back. Basically we had built a 2013 game and had to pull it back into the present. Growth in the next couple of years will be very easy. We'll just flick a bunch of switches back on."


I find this answer positive and I think the hard part is done for the team of 1C. Future expansions will support both more planes, new terrains and some exotic graphical features. ;)


//Simon

David198502
01-27-2011, 10:32 AM
sounds really good for all of us!those who have high end pcs and those who dont.

=PF=Coastie
01-27-2011, 12:08 PM
Well yeah they had to turn stuff off. They can't even take a single pic at high res with AA. But then again, they may only have a 2006 machine they are testing it on....LOL!

mr71mb0
01-27-2011, 12:21 PM
" Simon, the game is indeed extremely scalable. For the past six months all we've been doing is scaling it back. Basically we had built a 2013 game and had to pull it back into the present. Growth in the next couple of years will be very easy. We'll just flick a bunch of switches back on."

//Simon

Ahhhhhhh, my face is melting with excitement.

Sounds like the devs have been very sensible with targeting features and allowing enough scalability for the future. I also take this to mean that they will scale all features this way and not just eye candy.

Perhaps weather, AI, FM, DM, online environment updates as more cores / CPU power becomes available.

Baron
01-27-2011, 12:28 PM
Well yeah they had to turn stuff off. They can't even take a single pic at high res with AA. But then again, they may only have a 2006 machine they are testing it on....LOL!

They do have high end machines, reserved for programmers and "important, have to stuff". And before u say: "well, borrow one such machine for an hour and snap some screenshots", they allredy said that the "gameversions" on the high end machines are not fit for "human consumption" ...Iliyas words i belive. (correct me if im wrong).

Blackdog_kt
01-27-2011, 12:48 PM
Just like i was suspecting, they went about it the same way as with IL2.

We didn't have perfect landscape mode or detailed water in the early versions of IL2 either ;)

I like the way they are doing it to be honest.

Hecke
01-27-2011, 01:17 PM
hopefully they don't scale back too much. When new kepler is out, which is meant to be 3 times faster than fermi, there's again plenty of room for better graphics and physics.

LukeFF
01-27-2011, 01:26 PM
Just like i was suspecting, they went about it the same way as with IL2.

We didn't have perfect landscape mode or detailed water in the early versions of IL2 either ;)

Exactly

Tree_UK
01-27-2011, 02:14 PM
I cant believe they built a 2013 game built on DX9-10 and not DX11. That just doesn't add up.

TheSwede
01-27-2011, 02:19 PM
I cant believe they built a 2013 game built on DX9-10 and not DX11. That just doesn't add up.

With some common sense I think they prepared the engine so it is competitive in 2013 => DX11, DX12 etc AND prepared to handle the mainstream market of today => lots of computers out there with DX9 gpus.

Ilya has already mentioned that they have lots of features in store ready for activation when they are needed.


Bad comparison but BWM doesn't launch the M3 right away together with their newest iteration of the 3-series.

Cowboy10uk
01-27-2011, 02:22 PM
I cant believe they built a 2013 game built on DX9-10 and not DX11. That just doesn't add up.

Well Not all of us can afford to ramp up their Pc's every few months. I have a DX11 card, But I couldnt care less what Dx they us, So long as it runs well, Is a great Sim of the battle of Britain, and is moddable and upgradable, then isnt that all that counts. Im sure we can upgrade in the future, For now, Lets just get into those blue skys and enjoy ourselves. The graphics shown look great even if its not 11. Hell even the RAF dosnt have sims with perfect graphics. They have enough to do the job. Thats all we need. For a scalible graphics engine so it covers majority of our machines, it makes perfect sense.

BigPickle
01-27-2011, 02:29 PM
@ TheSwede, surely thats because they hadnt built it back then, but the sim has been built already.

Tree_UK
01-27-2011, 02:30 PM
With some common sense I think they prepared the engine so it is competitive in 2013 => DX11, DX12 etc AND prepared to handle the mainstream market of today => lots of computers out there with DX9 gpus.

Ilya has already mentioned that they have lots of features in store ready for activation when they are needed.


Bad comparison but BWM doesn't launch the M3 right away together with their newest iteration of the 3-series.

well i guess time will tell, but if DX11 isnt switchable, it only leaves us waiting for bigger CPU's and Graphics cards to cope with any features that are currently disabled, personally i think that is very short sighted.

addman
01-27-2011, 02:54 PM
Funny thing about all this is that a large part of the PC owners still uses WinXP and we all know WinXP won't go any higher than DirectX 9. Even though many gamers have bad-ass graphic cards that supports DirectX 10/11 they still use WinXP. Might be ONE of quite a few reasons why they choose to have the game scalable. If you make the game DirectX 10/11 only then maybe many potential customers might be lost.

TheSwede
01-27-2011, 03:06 PM
Funny thing about all this is that a large part of the PC owners still uses WinXP and we all know WinXP won't go any higher than DirectX 9. Even though many gamers have bad-ass graphic cards that supports DirectX 10/11 they still use WinXP. Might be ONE of quite a few reasons why they choose to have the game scalable. If you make the game DirectX 10/11 only then maybe many potential customers might be lost.

+1 on that one.

WinXP will still be used in 2012.

Blackdog_kt
01-27-2011, 05:32 PM
Funny thing about all this is that a large part of the PC owners still uses WinXP and we all know WinXP won't go any higher than DirectX 9. Even though many gamers have bad-ass graphic cards that supports DirectX 10/11 they still use WinXP. Might be ONE of quite a few reasons why they choose to have the game scalable. If you make the game DirectX 10/11 only then maybe many potential customers might be lost.

Exactly.

Also, correct me if i'm wrong but isn't DX11 mostly about tesselation?
I'm not an expert on graphics, but from reading that nVidia article a while back it seems that tesselation is mainly a technique used to add bumps and a relief structure to 2d items, in order to make them 3d without having to manually specify each and every curvature point/height data/etc. It seems like it's an algorithm that combines a couple of 2d sources to produce a 3d item on its own, on the fly. The advantage seems to be that it saves you some workload since it goes about it automatically, the disadvantage seems to be that you don't have as much control over how the surfaces are created, is that a good approximation?

Well, if that's the case then i really fail to see the use for it when the aim is to produce painstakingly accurate reproductions of military hardware when their silhouette will be closely scrutinized by hordes of rivet counters :-P

I wouldn't want to fly a 109 and have it create overdone bumps and whatnot on the metal fuselage skin that were different each time, plus that Spitfire screenshot we saw at one point (at high detail with AA/AF enabled) showed that "wrinkles" in the aircraft skin were already reproduced just fine with other techniques (like bump mapping).

In fact, i vaguely remember them stating this was their exact reasoning for not including it...it would destroy the accuracy of their 3d models and would need a complete rework of all models from the ground up to be of any real use.

I think it would be useful to add a true 3d feel to things like gravel surfaces on railroad embankments, masonry and stonework on buildings or maybe waves in the channel, but as for the actual units the needed workload is off the charts and would cause further delays, so i don't really mind them not including it at this point.

Heliocon
01-27-2011, 05:55 PM
Exactly.

Also, correct me if i'm wrong but isn't DX11 mostly about tesselation?
I'm not an expert on graphics, but from reading that nVidia article a while back it seems that tesselation is mainly a technique used to add bumps and a relief structure to 2d items, in order to make them 3d without having to manually specify each and every curvature point/height data/etc. It seems like it's an algorithm that combines a couple of 2d sources to produce a 3d item on its own, on the fly. The advantage seems to be that it saves you some workload since it goes about it automatically, the disadvantage seems to be that you don't have as much control over how the surfaces are created, is that a good approximation?

Well, if that's the case then i really fail to see the use for it when the aim is to produce painstakingly accurate reproductions of military hardware when their silhouette will be closely scrutinized by hordes of rivet counters :-P

I wouldn't want to fly a 109 and have it create overdone bumps and whatnot on the metal fuselage skin that were different each time, plus that Spitfire screenshot we saw at one point (at high detail with AA/AF enabled) showed that "wrinkles" in the aircraft skin were already reproduced just fine with other techniques (like bump mapping).

In fact, i vaguely remember them stating this was their exact reasoning for not including it...it would destroy the accuracy of their 3d models and would need a complete rework of all models from the ground up to be of any real use.

I think it would be useful to add a true 3d feel to things like gravel surfaces on railroad embankments, masonry and stonework on buildings or maybe waves in the channel, but as for the actual units the needed workload is off the charts and would cause further delays, so i don't really mind them not including it at this point.

http://store.steampowered.com/hwsurvey
Number will go up, as of now Win 7 64bit is the MOST popular OS system in use from the steam survey. Also Microsoft had record win7 sales. DX11 will go prime time in 2011, this is a huge mistake on the part of COD devs to not have dx11 and will come back to bite them.

Btw tesselation - it can be used on terrain for example which will greatly reduce gpu loads but give rolling hills, it can be used on buildings. Even without tessalation, the better shader pipeline and multicore support is worth it. DX10/9 doesnt work very well with multi threaded cores....

edit: The water in the channel would be DX11 Direct Compute, not tesselation. Also tesselation can be scaled and tailored so you can customise it easily. But yes the concept is correct, but it does more in that it adds far more detail to a model than a artist could do by hand in maya etc. Also it "scales" the detail as you get closer. Also DX11 brings advanced particle physics for smoke, clouds etc by using direct compute.

The Kraken
01-27-2011, 06:42 PM
Exactly.

Also, correct me if i'm wrong but isn't DX11 mostly about tesselation?

Well it's a bit more than that and includes some promising concepts (unlike DX10 which was pretty much a dead end), but currently it's still a niche technology that has to mature - not much sense in wasting development resources on that if there are better things to do that everyone profits from.

Tesselation in particular is more of a buzzword than a useful technology at the moment. Too much effort required on the modellers' part for anything but trivial cases (where displacement or even bump maps work nearly equally well), and too slow to provide any performance improvements over established technologies.

As with previously advertised so-called "break-through" technologies (e.g. texture filtering, hardware T&L, antialiasing) it may eventually turn out to be fundamental on the 2nd or 3rd generation of cards that support it. However given that most games these days are ported from consoles, where this obviously isn't available, it's unlikely that we'll see widespread use for this in the game industry beyond some smaller cosmetic improvements to surfaces.

Blackdog_kt
01-27-2011, 06:44 PM
I don't dispute the benefits they could have from using DX11. Obviously, some people developed it because it would actually be good enough for other people to use ;)

What i dispute is that the overall gains (overall as in not only about performance, but general viability of the product) would be so significant to justify an extra few months of delays while they are reworking all of their 3d models.

I mean, if it's already in the engine and it's just turned off because it needs further debugging (as stated on the other thread by one of the devs), then there's nothing stopping them from selling a game in a working state and debugging while the cash comes in, maybe even hiring a few extra people to help do it faster now that they are going to finally start getting paid for their work. In fact, i think it's preferable that not having it for an extra who knows how many months and jeopardizing the entire product's financial viability.

Would it be nice to have it? Well, i don't have a DX11 capable GPU and i don't personally care either way, but sure it would for the people that do. However, i don't think it's going to come back and bite them.

In general, it's a fine balance to release something that works acceptably well right out the box. I equally dislike the practice of companies releasing unfinished games and basing further support and development on initial sales, as i also dislike delaying it ad nauseum to incorporate every new bit of emergent technology, so from where i'm standing they made a well balanced choice.
They left out a half-finished feature that only a minority of the PC demographic can use today, with the aim of working on and completing it during the following months as the amount of people who can use it grows and it would make more sense to include it, while making sure they can release something that will work acceptably for the majority.

All in all, technical specs mean little to me if the feel is not correct and the immersion is missing. This is supposed to be our private little time machine back to the summer of 1940. Sure, graphics play a big part but so does sound and the actual gameplay mechanics and experience in equal measure. It's a compromise between all these things and the capabilities of current hardware.
For example, i wouldn't want it shipped with DX11 if that meant having to axe other features like the new multiplayer modes, the multi-level AI routines or stuff that have to do with the actual flying. It's a game about aircraft first and foremost, so as long as the aircraft look accurate and fly close enough to the real thing i'm not going to mind having a few less polygons for the sake of playability.

In any case, this is my personal opinion and not some kind of gospel to be forced on you so by all means feel free to disagree. It just seems that the majority of potential customers are more interested in having the sum of its parts capture the "time machine" feeling, instead of having one part being excellent to the detriment of the combined total ;)

addman
01-27-2011, 06:54 PM
I don't dispute the benefits they could have from using DX11. Obviously, some people developed it because it would actually be good enough for other people to use ;)

What i dispute is that the overall gains (overall as in not only about performance, but general viability of the product) would be so significant to justify an extra few months of delays while they are reworking all of their 3d models.

I mean, if it's already in the engine and it's just turned off because it needs further debugging (as stated on the other thread by one of the devs), then there's nothing stopping them from selling a game in a working state and debugging while the cash comes in, maybe even hiring a few extra people to help do it faster now that they are going to finally start getting paid for their work. In fact, i think it's preferable that not having it for an extra who knows how many months and jeopardizing the entire product's financial viability.

Would it be nice to have it? Well, i don't have a DX11 capable GPU and i don't personally care either way, but sure it would for the people that do. However, i don't think it's going to come back and bite them.

In general, it's a fine balance to release something that works acceptably well right out the box. I equally dislike the practice of companies releasing unfinished games and basing further support and development on initial sales, as i also dislike delaying it ad nauseum to incorporate every new bit of emergent technology, so from where i'm standing they made a well balanced choice.
They left out a half-finished feature that only a minority of the PC demographic can use today, with the aim of working on and completing it during the following months as the amount of people who can use it grows and it would make more sense to include it, while making sure they can release something that will work acceptably for the majority.

All in all, technical specs mean little to me if the feel is not correct and the immersion is missing. This is supposed to be our private little time machine back to the summer of 1940. Sure, graphics play a big part but so does sound and the actual gameplay mechanics and experience in equal measure. It's a compromise between all these things and the capabilities of current hardware.
For example, i wouldn't want it shipped with DX11 if that meant having to axe other features like the new multiplayer modes, the multi-level AI routines or stuff that have to do with the actual flying. It's a game about aircraft first and foremost, so as long as the aircraft look accurate and fly close enough to the real thing i'm not going to mind having a few less polygons for the sake of playability.

In any case, this is my personal opinion and not some kind of gospel to be forced on you so by all means feel free to disagree. It just seems that the majority of potential customers are more interested in having the sum of its parts capture the "time machine" feeling, instead of having one part being excellent to the detriment of the combined total ;)

+1 to that!

recoilfx
01-27-2011, 07:18 PM
I cant believe they built a 2013 game built on DX9-10 and not DX11. That just doesn't add up.

Why not? DX paths are just different renderers. The developers are right to concentrate on DX9-10, where 90% of the market is at.

It's not like the whole game is based on DX versions. You don't need to dump the whole engine for a few new DX11 APIs.

I am sure that the developers have looked into DX11. Maybe that path is even half done, but they realized that the additional time and money needed isn't worth it for this release.

When the time the market is ready for DX11, they only need to rework their renderer and release the upgrade as an expansion pack. I'd gladly support Oleg with more money.

Azimech
01-27-2011, 07:23 PM
Perhaps weather, AI, FM, DM, online environment updates as more cores / CPU power becomes available.

I sure hope AI, FM and DM are at 100%.

Azimech
01-27-2011, 07:28 PM
I wouldn't want to fly a 109 and have it create overdone bumps and whatnot on the metal fuselage skin that were different each time...

I believe the name Warthog was for a different plane so I agree.

Heliocon
01-27-2011, 07:38 PM
Well it's a bit more than that and includes some promising concepts (unlike DX10 which was pretty much a dead end), but currently it's still a niche technology that has to mature - not much sense in wasting development resources on that if there are better things to do that everyone profits from.

Tesselation in particular is more of a buzzword than a useful technology at the moment. Too much effort required on the modellers' part for anything but trivial cases (where displacement or even bump maps work nearly equally well), and too slow to provide any performance improvements over established technologies.

As with previously advertised so-called "break-through" technologies (e.g. texture filtering, hardware T&L, antialiasing) it may eventually turn out to be fundamental on the 2nd or 3rd generation of cards that support it. However given that most games these days are ported from consoles, where this obviously isn't available, it's unlikely that we'll see widespread use for this in the game industry beyond some smaller cosmetic improvements to surfaces.

Wrong.
1. Tesselation requires relativly little modeling work for the level of detail it gives, thats why its great.
2. Tesselation is dynamic which means it wont render geometry you cant see because of distance, in dx10 you would have to render it, or switch out models are distance. Either option is more work and worse performance.

Napoleon total war was capped to about 12k max soldiers on the battlefield before the engine started to "break down" on even the best pc's. They had to use sprites to reduce the gpu load at a distance.

Shogun 2 total war will have a cap of over 57k soldiers on the field, running DX11, soldiers will be tesselated and not use sprites. Its system requirments are also very similar/the same as NTW, just the gpu is 1 gen up. ie dx11=better performance.

Chivas
01-27-2011, 10:50 PM
I don't understand why its so hard to realize that they were working on DX11 features but were experiencing too many bugs to release these features. It has nothing to do with the developers being short sighted.

bf-110
01-27-2011, 10:56 PM
Not sure if this is off topic,but how flexible is SoW engine?
I mean,if it´s a game that can get outdated in a few years,if it will still be a modern game in 2014,if the game can be upgraded to have better graphics/sounds/etc.

swiss
01-27-2011, 10:59 PM
I don't understand why its so hard to realize that they were working on DX11 features but were experiencing too many bugs to release these features. It has nothing to do with the developers being short sighted.

because some teenagers can't take it if don't receive the newest, flashiest stuff from day one, plus, they are impatient.

Robert
01-27-2011, 11:11 PM
Not sure if this is off topic,but how flexible is SoW engine?
I mean,if it´s a game that can get outdated in a few years,if it will still be a modern game in 2014,if the game can be upgraded to have better graphics/sounds/etc.


Short answer, YES. That's the whole idea behind what Ilya said on the facebook page. CoD will have scalable graphics AND features that are already encoded into the game's engine so when the time is right they can be flipped on like a switch.

In the original IL2/FB there weren't detailed clouds, and perfect landscape. After about four years and PC specs came up to a certain level these features were either turned on or implemented by Oleg's team. A totally modular design feature.

Heliocon
01-28-2011, 03:18 AM
I don't understand why its so hard to realize that they were working on DX11 features but were experiencing too many bugs to release these features. It has nothing to do with the developers being short sighted.

Yes, you are right. I would of had no issue at all with this if the devs had not implied dx11 would be in for release. They should of clarified what will be and will not be in the game at an earlier date. There is a scarcity of information on the game, in fact there is nearly nothing which leads to speculation or wrong assumptions.
Also it conflicts with their claim on the "lifetime" of the new engine...

zauii
01-28-2011, 03:54 AM
I cant believe they built a 2013 game built on DX9-10 and not DX11. That just doesn't add up.

You make dx11 sound a lot more important for the consumer than it actually is, have you actually understood what dx11 does or are you just looking at the numbers..

Chivas
01-28-2011, 04:20 AM
Yes, you are right. I would of had no issue at all with this if the devs had not implied dx11 would be in for release. They should of clarified what will be and will not be in the game at an earlier date. There is a scarcity of information on the game, in fact there is nearly nothing which leads to speculation or wrong assumptions.
Also it conflicts with their claim on the "lifetime" of the new engine...

How does it conflict with the "lifetime" of the new engine. The engine is capable of doing dx11 and probably beyond, they just have to work out the bugs of the DX11 features they have now.

The developers have kept us informed, because they want our feedback. I'm sure they had every intention of the initial release having dx11 aspects, but you can't predict how many bugs will develop, and how long it will take to fix them.

Its a double edge sword keeping the customer in the loop. You get some good feedback, but trashed if something doesn't make the initial release.