46 replies to this topic

[Jetfire]

I see no reason at all to believe that the engine is the cause of any of the problems people have stated. In the end any engine could do the job, it is just a matter of writing very efficient code to allow for maximum throughput between users and the servers. You can change the engine all you want but unless the engine simply contains very little server authenticated data your netcode is going to be the primary barrier and you have to write that for any engine.

If you just use the server to transmit user data but do not authenticate it then you can do a lot more but the downside is that it is very easy to hack the game.

[Kiiyor]

Anjian, on 24 March 2015 - 11:43 PM, said:

Its impossible to change the game engine right now. If it were being built in scratch, I would love them to use the DAGOR engine used in War Thunder.

I get graphics like these and yet with significantly higher FPS, with 32 combatants on the field, not counting all the AI tanks which are replicas of the ones player use. Which means the engine has plenty of room for additional combatants. Along with dynamic lighting and weather, destructible terrain, huge water fountains where bombs drop on the water, camouflage scaling and degradation. If you want your mech looking its got weathered paint or has taken some hard punches. What's more, they are achieving maps at least 4k x 4k, which is 4x bigger than the maps in MWO, which is at max, 2k x 2k.

WT does indeed look amazing. I'd love it if MWO could be squeezed into that engine.

It's worth noting however that the tanks in WT have almost no moving parts compared to mechs in MWO. Each mech in MWO has dozens of points of articulation, and fatalistically complicated animation and damage skeletons. Every tiny moving piece needs a render call, and most need a server call. Most mechs have also have more weapons than WT tanks (sometimes dozens more) all with different firing modes, flight paths, splash damage radius and damage ticks, all requiring rendering, all requiring server calls, all requiring damage calculations using several different models - armour damage, internal damage, component damage, location damage - all of which also require draw calls, all requiring server calls.

How often in WT are more than 5 or 6 tanks firing at once? Now think of your average fustercluck in MWO, and the tank battalion's worth of ordnance being thrown downrange by mechs every few seconds.

MWO isn't spectacularly optimized, but people always underestimate just how expensive mechs are to render, and how much strain they put on the MWO servers. WT still slows down if things get particularly crazy.

If there were a way to get WT fidelity and still have gorgeous mechs, i'd be a very happy camper. While I think there are better engines out there, I don't begrudge PGI for the mileage they have squeezed out of Cryengine.

[LordNothing]

i dont think the problem is the engine. in fact i think the engine is being far underutilized. its capable of much more than what pgi has got it to do so far. if it was another engine we would be wondering why they dont switch to cryengine.

Edited by LordNothing, 25 March 2015 - 04:06 AM.

[carl kerensky]

NextGame, on 25 March 2015 - 12:57 AM, said:

If you were changing engine right now, you might as well make a completely new game.

Why not then? Really. All I have seen is regression visually in MWO and excuse after excuse from PGI that this engine has limitations. They just remove beautiful features and alter core physic gamepay. In early CB this engine looked fantastic and gameplay was epic. You felt like you were in the mech. I guess PGI decided to go the COD/fast twitch gameplay route and degrade graphics. A perfect combo for which I do not play anymore. No thanks. I am now waiting for them to A) come to there senses and make it priority to get that mech feel back or another developer creates a Mechwarrior game that makes you feel like your piloting a huge machine.

[Wrathful Scythe]

And who says that another engine will help?
It might be easier but PGI has more experience with the cryengine. They would need to learn everything from scratch if they decide to change the engine now. And a badly written game still runs like garbage no matter what engine you use.

We have what we have and need to live with it.

[GreyNovember]

OP, I'm genuinely curious here.

What factors make you believe that another engine will solve all these issues you claim Cryengine to have? Please, enlighten me.

[Kotzi]

No pro here so just my 2 cents. AFAIK is an engine the tools and hardware of any game. The better the tools the more you can do with them, if you know how to handle them though. I dont think that the cryengine is one of the bad ones. I would rather say that pgi cant handle the engine and their skills in programming and design are mediocre. UI2 is less convinient than its predecessor for example. Last time they changed the grouping tool so that you would have to click the ok button for each invite. Only for an affirmation window. This is the opposite of what you want to achieve, i want it easier not more annoying.

And in regards to the change of the graphical part of the engine i think it is to lower the requirements. You want to address as many potential customers as possible.

[Lily from animove]

nitra, on 24 March 2015 - 08:27 PM, said:

But But star citizen is made from the engine !! how can it not be the greates thing evar !!



I think when they started making mwo they really didnt have to many choices in engines.

and cryengine 3 was probably the most advanced engine at the time.


i would have rather seen a custom engine cause that is what battle tech really needs. but that is undertaking all its own.

so i would argue that giving the time frame and what was available pgi made a pretty good choice for the engine.

Hmm, i wait for the day, when SC tries to make 40+ intense playerbattles. doubt the engine will survive that.

[Anjian]

Kiiyor, on 25 March 2015 - 03:28 AM, said:

WT does indeed look amazing. I'd love it if MWO could be squeezed into that engine.

It's worth noting however that the tanks in WT have almost no moving parts compared to mechs in MWO. Each mech in MWO has dozens of points of articulation, and fatalistically complicated animation and damage skeletons. Every tiny moving piece needs a render call, and most need a server call. Most mechs have also have more weapons than WT tanks (sometimes dozens more) all with different firing modes, flight paths, splash damage radius and damage ticks, all requiring rendering, all requiring server calls, all requiring damage calculations using several different models - armour damage, internal damage, component damage, location damage - all of which also require draw calls, all requiring server calls.

Not really when it comes to the moving parts. As a matter of fact, if you watch how the tracks and each individual wheel articulate on the tank, as each float and follow the contour of the terrain, each WT tank has more moving parts compared to a mech. Some open air tanks have moving crew members which you can see that are rapidly turning wheels to turn their cannons. The flak cannons --- some vehicles have four of them --- have their own separated articulated recoil, and they do it rapidly. One tank has a pail hanging on its back and you can see it swing back and forth as you roll past the terrain.

The damage models of each WT tank is far more intricate. In fact, they are designed to simulate real physical damage.

There is no HP.
Each tank is much more complex polygonally than a mech. There is a huge variety of surfaces, slopes, angles and curves, each recreated from real museum tanks. Each surface has different armor thickness. This is why it matters where you hit a tank, because a shell can be deflected, or you can send it right into the driver's hatch. A shot on the back can set the engine on fire, a shot on the track can immobile the tank. And if you hit the right spot, it goes boom in one shot. And if you hit all the wrong spots, you can fire more than 10 shots and the tank will still get away.

A WT tank has "internal" hit spots. When a shell is fired, a lot of things happen.
The game follows actual physical ballistics --- weight of the shell, forces of gravity, acceleration, air resistance --- to draw a flight patch from the gun to the target. No shot is a straight line, it bends with an apogee. You need to aim higher on slower shells to compensate for range.

Then when the shell hits a tank, a lot of things have to happen to determine a pentration or not.
Where the shell hits on the tank, at what velocity and range.
The slope, curvature, and thickness of that particular armor facet.
If there are any obstacles like side armor, skirts, etc,. These are broken away as the tank gets more hit.

If the shell penetrates or not, if it penetrates, more variables comes into play.
If the shell is a solid shot.
If the shell has high explosive in it. All shells do not have the same amount of HE. The amount of HE in each shell is determined by historical data (US and British shells have lower HE component than German and Soviet shells).
The fuse timer inside the shell.

This will determine where the shell explodes, the blast and fragmentation radius of the shell inside the tank.
How the blast radius affects the tank, determines on the tank's internal hit areas. Which are:
The four or five crew members and where they sit or stand.
The location of the ammo racks
The location of the fuel tank.
The location of the engine and the transmission.

Do you know you get less of a chance of being ammo racked if you reduce the amount of ammo you save into the tank?

The fate of the tank can result in a wide variety of results, ranging from crew members being knocked out, or just the tank going up in one shot. Sometimes it takes multiple shots to take out a tank, sometimes all it takes is one shot.

The rinse and repeat that across more than 32 tanks in the battlield.

Quote

How often in WT are more than 5 or 6 tanks firing at once? Now think of your average fustercluck in MWO, and the tank battalion's worth of ordnance being thrown downrange by mechs every few seconds.

The same game engine is also used to simulate 32 fighters and bombers twisting in the air at high speed, shelling out rapid fire from 4 to 10 rapid fire cannons or machine guns, each dispensing hundreds of rounds in seconds. The same goes with the AA tanks.

On top of that, the game has to simulate carpet bombing attacks from large bombers, blast radius and all, which maybe falling on those tanks.

Quote

MWO isn't spectacularly optimized, but people always underestimate just how expensive mechs are to render, and how much strain they put on the MWO servers. WT still slows down if things get particularly crazy.

I have not really seen WT slow down, not even in dramatic air battles. The only time it slows down are PC graphics lag from heavy textures. To do what WT does, not just requires very powerful servers, but some incredible netcoding, and I am very impressed with the netcode in War Thunder.

Quote

If there were a way to get WT fidelity and still have gorgeous mechs, i'd be a very happy camper. While I think there are better engines out there, I don't begrudge PGI for the mileage they have squeezed out of Cryengine.

I don't really think they have done that well. Those bad maps that get your mech feet stuck, those obstruction actually consume polygons and textures. This results in a map that consumes resources, both server and PC, only to make your gaming experience worst.

Those overly complicated map structures --- consuming vital polygon and texture budgets --- also contribute to environment hitting, because your gun arm is low, and how you shoot doesn't reflect how you see it. The result is hitting your environment instead. This consumption of graphics budget doesn't improve your gaming experience at least.

Those mechs with large torsos that make them easier to hit, yet they also consume on the graphics budget. That's a double negative.

I for one thinks that some of the maps need a complete pass and reevaluation, and many of the existing mechs as well. There is no point in making mechs huge and easy to target, and huge mechs consume more of the graphics budget even more. The maps and the mechs have the feeling they are rushed out, and not enough time and effort, maybe because of money constraints, to really optimize them.

Edited by Anjian, 25 March 2015 - 07:25 AM.

[Jetfire]

Lily from animove, on 25 March 2015 - 07:06 AM, said:

Hmm, i wait for the day, when SC tries to make 40+ intense playerbattles. doubt the engine will survive that.

Given the intense rework that has basically branched their engine from CryEngine years ago I see nothing to make me suspect this will be an issue but somewhere between 40 and 150 I expect there will arise a very real limit.

[Anjian]

Personally once they have committed to that game engine, they have to stick with it, unless they plan on overhauling the game completely, like a major reboot.

I think the main problem PGI has is their rushing. They rush maps, they rush their mech designs to meet deadlines (remember the one mech a month thing --- that is just a freaking awful thing to sustain). All to satisfy their user base who got addicted to the thrill of monthly updates.

The other problem is that they don't seem to have the money, developer resources and time to go over the maps and mechs for further experience and optimization passes once the content is released.

[Loc Nar]

I can only imagine how crytek feels seeing what pgi has done/is doing with their engine. Predictable results from shovelware devs that cling to 'minimum viable product' office culture. It's all they know.

[DONTOR]

Will I ever be able to go more than 171KPH? If the answer is no, than I too am extremely dissapointed in Cryengine.

[Tarl Cabot]

Anjian, on 24 March 2015 - 11:43 PM, said:

Its impossible to change the game engine right now. If it were being built in scratch, I would love them to use the DAGOR engine used in War Thunder. I get graphics like these and yet with significantly higher FPS, with 32 combatants on the field, not counting all the AI tanks which are replicas of the ones player use. Which means the engine has plenty of room for additional combatants. Along with dynamic lighting and weather, destructible terrain, huge water fountains where bombs drop on the water, camouflage scaling and degradation. If you want your mech looking its got weathered paint or has taken some hard punches. What's more, they are achieving maps at least 4k x 4k, which is 4x bigger than the maps in MWO, which is at max, 2k x 2k.

Atm only one map have rolling hills but it is also a small map, and another has some trees but they appear to be more of an afterthought. Those two maps are small and more conducive to a quick 8vs8 or even 4vs4.

What they are missing are more Earth-like terrain that actual battles have been fought on, and that is why many are so down about the current maps themselves. Tis like at times it is being thought about too much when designing maps. And the ones with actual buildings try to make it appear as the outskirts of a city.

[GeneralArmchair]

Cryengine isn't the best choice for multi-player games. MWO barely gets away with it by using such small teams. Space toilet simulator's decision to use the engine was based entirety on it being pretty and it being a household name. It is really questionable if they'll be able to deliver a tenth of what they've promised.

[Felio]

I think you mean achieved with Cryengine.

[Anjian]

Tarl Cabot, on 25 March 2015 - 08:35 AM, said:

Atm only one map have rolling hills but it is also a small map, and another has some trees but they appear to be more of an afterthought. Those two maps are small and more conducive to a quick 8vs8 or even 4vs4.

What they are missing are more Earth-like terrain that actual battles have been fought on, and that is why many are so down about the current maps themselves. Tis like at times it is being thought about too much when designing maps. And the ones with actual buildings try to make it appear as the outskirts of a city.

I don't know what you mean, since they have nine maps already, and some of them are huge that it takes some time to race from one point to another. At least four of these maps feature hills, two are urban, two are representation of actual battle locations (Kursk), and one upcoming map based on the Battle of Berlin where the Reich made its last stand.

The Dagor engine used in War Thunder has also been used in RPGs, and at least one space game.

http://star-conflict.com/

Edited by Anjian, 25 March 2015 - 09:30 AM.

[Summon3r]

i am quite positive cryengine is not the issue at hand in MWO in any way shape or form, like not even close, like not even in this universe close. pee gee i maybe?

[Heffay]

As an F2P game, MWO has to make sure the game works for as many possible systems as they can. The minimum game requirements are incredibly flexibly. If they decided to ramp up the minimum requirements, they could do a lot more. However, they'd have to walk away from a significant portion of their player base.

I remember them saying once that 30% of their players play the game at 1024x768. I mean you can DO a lot if you require a minimum of 8GB of RAM, i7's and a 970 or higher graphics card, but then your potential player base evaporates.

Star Citizen is willing to walk away from the marginal systems, but they are a rather unique snowflake.

[Sarlic]

The engine is not the problem.

The engineers are the problem. It's a extremely difficult engine to customize. PGI customized it...