57 replies to this topic

[Nesutizale]

Just found this on twitter and I think it shows why they haven't thought or started developing MWO 2.0 in unreal so far.

Whens talked about the burstfire AC and if it could be used in MWO:

Quote

We can do this in MW5, but probably cant’t for MWO. The network engineers would probably kill me if I tried to push for a weapon with very high projectile counts with bullet spread+camera recoil+gravity on each projectile to also be compatible with the HSR system in MWO.

Sounds like a good reason to port MWO to UE4 using MW5 assets to me....

A bit more complicated than that. Since MW5 doesn’t have to worry about server side authorizing weapon fire, it allows us to throw more robust effects into the weapons since we don’t have to worry about server side tracking every shot made. So even if MWO used UE4, we would still need to keep weapon effects limited to whatever the server can reliably track for server side hit registration. Just one of the benefits of MW5 being made for a single player experience. My personal favorite, missiles follow their path up to their optimal range and then loose their propulsion and get gravity, additional scatter, and hit wind resistance. We just have a lot more toys to play with when we don’t have to worry about server tracking.

I would understand this so that MW5, as it is now, would be to taxing to the servers to allow multiplayer matches like we have in MWO. I hope they find a way to get around that or get a better netcode or servers or whatever it takes to port MWO to MW5.
At least 4v4 PvP would be nice.

On a side not, interesting to know that gravity is beeing used for this. I noticed the curved fire on AC shells but wasn't sure how they did it. Also does that mean if planets have different gravity it will also effect ballistics?

[FRAGTAST1C]

SRMs are also affected by gravity in MW5. You can see them arc down the further they travel. Unless I was off my hinge.

[Nesutizale]

I would guess they have all projectiles and missiles beeing effected by gravity.

What makes me wonder, when performance realy takes such a drastic hit that the servers can't handle realtime combat with each missile or AC shell beeing calculated, would it be so bad to fake things a little bit?

I mean what about doing the following, lets take an SRM 6 as an example.
You pretty much know the spread and how it behaves over distance. So instead of calculating 6 flightpathes, each with its own hit detection that has to be send back and forth between client and server, lets just calculate the spread over distance and its path.
Only if that results in a hit further calculation is done for each missile or even more simple, see how much % of the spread path hits the target and give it damage according to the %.

Yes that wouldn't be accurate and there might be a slight visual disconnect but when the option is to have no MP at all or have one that isn't 100% accurate I think I could life with the less accurate.
Besides MWOs hit registration is buggy allready and not 100% accurate so I think it would be similar.

Overall I think there must be a way to reduce amount of informations beeing send back and forth between client and server and still have some good gameplay.

[MW Waldorf Statler]

with Flamers , my match come to a Stop motion Show , im not Thinking ,thats the UE is the best for PvP MWO 2, seeing all the big problems with ARK

[Nesutizale]

PGI said that the flamers have a bug, that they are working on right now, that causes the problems.

Anyway I am no software engineer, I can't say if this is just a matter of getting more capable people, changeing the design philosophie, a matter of just getting better servers or realy something that just can't be done with the engine as it is.

Just wanted to bring that up as its sad that his kind of informations aren't here in the forums.

[LordNothing]

for projectile drop, if you are using deterministic physics, once you have an initial starting position and velocity vector, you should be able to calculate that shell's position at any point along its trajectory against time. of course that goes out the window with homing weapons.

you could treat a (dumbfire) missile or projectile spread as a field though, which grows as a function of time. when a collision happens you can then figure out how much of that field intersects each collision hull and then compute a fraction of damage each has taken. it sounds hard to figure this out, but you can do a lot of abstractions. fast intersections and volume computations between very simple primitives (spheres, cubes, possibly arbitrary convex hulls) might be able to do the trick. each time the field intersects a thing, fast forward time to the point where z distance between the field position and the thing being hit is the same, figure out what fraction of volume has been intersected. this allows you to allocate damage from the field to the thing being hit. it can go on damaging things beyond the mech if the damage > 1 missiles worth, or until it hits something else that can absorb the damage. or until the field is so diffuse that no possible damage call can be worth more than a missile. its not as clean as tracking every projectile (for example its possible to hit another thing that was directly occluded by the first thing, barring much more complicated volumetric analysis).

and i dont think unreal engine makes the problems go away, though there may be better systems in play to handle them. after all cryengine is mostly dead at this point (barring derivatives), while ue is still actively developed.

Edited by LordNothing, 02 December 2019 - 04:38 AM.

[Nesutizale]

I think your second paragraph is what I had in mind in more complex words
Beside the details on how to do it, I think computers are very capable of doing the calculations in the end, finding creative ways to go around the limits that calculating every single projectile provids is the hardest part. Its a mix of software engineering and math and creative thinking.

As for the engine, no engines don't solve problems. They layout the basic framework to work from and if you are lucky someone allready solved your problem for you and made the solution part of the engine.
From looking at StarCitizen, a game that basicly is Cryengine++ one can see how you can take an existing engine as a startingpoing and then add your solutions on top of that.
IIRC that is also what PGI did with MWO. I think the original engine didn't supported multipart player characters aka mechs with tons of hit locations that are hanging together but each have their own damage registration and values and damage transfere. All that was never part of the Cryengine but it provided a lot of other stuff that you could use.

Personaly I think the Cryengine still looks good and I think it could still compete with other engines today. Its just that unreal editor is so accessible. I think that is the biggest advantage of the unreal engine, that it has a very well done editor, that is easy to understand and has tons of tutorials and a shop to buy and sell assets from.

Edited by Nesutizale, 02 December 2019 - 05:15 AM.

[LordNothing]

Nesutizale, on 02 December 2019 - 05:10 AM, said:

I think your second paragraph is what I had in mind in more complex words
Beside the details on how to do it, I think computers are very capable of doing the calculations in the end, finding creative ways to go around the limits that calculating every single projectile provids is the hardest part. Its a mix of software engineering and math and creative thinking.

As for the engine, no engines don't solve problems. They layout the basic framework to work from and if you are lucky someone allready solved your problem for you and made the solution part of the engine.
From looking at StarCitizen, a game that basicly is Cryengine++ one can see how you can take an existing engine as a startingpoing and then add your solutions on top of that.
IIRC that is also what PGI did with MWO. I think the original engine didn't supported multipart player characters aka mechs with tons of hit locations that are hanging together but each have their own damage registration and values and damage transfere. All that was never part of the Cryengine but it provided a lot of other stuff that you could use.

Personaly I think the Cryengine still looks good and I think it could still compete with other engines today. Its just that unreal editor is so accessible. I think that is the biggest advantage of the unreal engine, that it has a very well done editor, that is easy to understand and has tons of tutorials and a shop to buy and sell assets from.

cryengine is still a fairly modern engine. with active development it would certainly be something. and its not like its an evolutionary dead end either. there are many derivatives. but holding on to an old obscure version like what mwo has been using comes at a cost.

its also a lot easier to migrate to a newer version of the same engine than it is to change engines entirely. if i was doing an f2p that had to run a decade or more id choose an engine with really good long term support options.

Edited by LordNothing, 02 December 2019 - 05:53 AM.

[Anjian]

Nesutizale, on 02 December 2019 - 02:27 AM, said:

Just found this on twitter and I think it shows why they haven't thought or started developing MWO 2.0 in unreal so far.

Whens talked about the burstfire AC and if it could be used in MWO:



I would understand this so that MW5, as it is now, would be to taxing to the servers to allow multiplayer matches like we have in MWO. I hope they find a way to get around that or get a better netcode or servers or whatever it takes to port MWO to MW5.
At least 4v4 PvP would be nice.

On a side not, interesting to know that gravity is beeing used for this. I noticed the curved fire on AC shells but wasn't sure how they did it. Also does that mean if planets have different gravity it will also effect ballistics?

I am not sure why this should be a problem. Unless they want to needle and dime the servers for the lowest cost of operating servers as possible.

Here is a game for example. 14 to 15 players per team. Everything here relies on curved trajectories based on real world historical acceleration and deceleration figures of each shell, which is used to calculate the entry and impact trajectories of the shells against angled armor, also calculating for penetration. You are shooting as much as 12 projectiles simultaneously into the air, each also computed with a random circular error of probability, or dispersion fields of the shell.




Next one is using the Unreal Engine, having a PvP match. You can see missiles and shells curving. This one works off from an iPad.

Edited by Anjian, 02 December 2019 - 07:03 AM.

[MW Waldorf Statler]

Hi Anjian the last video is very different to mechs ,all have only a Single Hitzone im Thinking , not multiple Hitzones thats glitches, overlapping etc ...by a MWO Mech , each Mech has 11 single Hitzones, thats by 24 Mechs 264 Single Hitzones .

what new Crygames can , we can seeing by War of the Rights and Hunt :Showdown;)

https://youtu.be/A84LcBNhXq8

https://steamcdn-a.a...bm?t=1566909105

Edited by MW Waldorf Statler, 02 December 2019 - 08:33 AM.

[Unleashed3k]

MW Waldorf Statler, on 02 December 2019 - 08:31 AM, said:

Hi Anjian the last video is very different to mechs ,all have only a Single Hitzone im Thinking , not multiple Hitzones thats glitches, overlapping etc ...by a MWO Mech , each Mech has 11 single Hitzones, thats by 24 Mechs 264 Single Hitzones .

what new Crygames can , we can seeing by War of the Rights and Hunt :Showdown;)

https://youtu.be/A84LcBNhXq8

https://steamcdn-a.a...bm?t=1566909105

Don't talk about Hunt: Showdown if you haven't played it at all, the first early acces versions had great graphic, but the engine was not able to bring decent framerates, they had to dumb it down over 35% and now walls simply don't render behind 50-100m, this game had a smaller playerbase after release then MW:O has currently, seems you can't judge anything at all...

[MW Waldorf Statler]

Thanks for the Statement Unleashed3K:) ...im never played Hunt , so is interesting to hear what is in the real Game ! ok ,thats good to hear ,thats a good Post !so ist like the Problems from ARK with the UE

Edited by MW Waldorf Statler, 02 December 2019 - 09:20 AM.

[Koniving]

Nesutizale, on 02 December 2019 - 02:27 AM, said:

Just found this on twitter and I think it shows why they haven't thought or started developing MWO 2.0 in unreal so far.

Whens talked about the burstfire AC and if it could be used in MWO:



I would understand this so that MW5, as it is now, would be to taxing to the servers to allow multiplayer matches like we have in MWO. I hope they find a way to get around that or get a better netcode or servers or whatever it takes to port MWO to MW5.
At least 4v4 PvP would be nice.

On a side not, interesting to know that gravity is beeing used for this. I noticed the curved fire on AC shells but wasn't sure how they did it. Also does that mean if planets have different gravity it will also effect ballistics?

RAC...
Quad AC/2...
Etc.

Also graviy is being used in MWO. (Well, actually technically no it's just ballistic trajectories)

Note the ballistic trajectories.
(If it actually was gravity the bullets would slam into the ground considering every map in MWO has over 2.5* Earth's Gravity which is why mechs instantly start falling when the jumpjets turn off, and you hit 30+ feet per second [changed to meters but if it is meters per second you're actually looking at 4-5 times earth's gravity then] which is why things literally slump over, too.)

The mechs would be slower to lose acceleration when the thrusters are turned off (you'd still be going upward briefly), and slower to gain downward acceleration from gravity (you'd experience a brief moment where you'd hit 0 acceleration in any upward/downward direction, i.e. you'd stop in air and then begin to fall... and the jumpjets would have to fight that fall but have a chance to win... at the moment the instant you start going down even 100% jumpjet fuel used after a fall cannot cause you to raise in the air it can only make you hover in place before losing to the over oppressive gravity).

Guess the next thing for me to test is MW5's gravity. I know the jumpjets are terrible...

This video's ancient, but...





These make that statement kinda hard to follow.
Now granted what's dropped is the camera movement and recoil.

And 3 months ago.

[Jackal Noble]

You did it again... Koniving.
We've talked about this.

[Nesutizale]

Well I am allways facinated and scared at what he finds out next
But back to the topic of trajectory. I think there is an importend differance between giveing a gun a fixed path the bullet will follow to haveing actualy gravity and windresistance decide how the bullet acts. One is a very pretictable the other has to constantly be calculated anew. I did say that with the fixed stuff tehy could save a lot of computing power.

As for the Warships example from Anij....Wargameing seams to have found a solution for their engine for this very specific problem. Then again from some of the dev videos I have seen, they seam to have some quite good engineers.
Last video I saw was the "software" solution for Raytraceign that was still workable without RTX cards and that was quite clever thinking if you ask me.
Lets hope that either one at PGI comes up with a solution or that some modders will optimize the game in that regard.

Frankly MW5 needs further opimization. From some testing I made my system lost ~28 frames when effects where set to max and ~37 frames when post processing was at max. That is still quite a lot in my book.

[Anjian]

MW Waldorf Statler, on 02 December 2019 - 08:31 AM, said:

Hi Anjian the last video is very different to mechs ,all have only a Single Hitzone im Thinking , not multiple Hitzones thats glitches, overlapping etc ...by a MWO Mech , each Mech has 11 single Hitzones, thats by 24 Mechs 264 Single Hitzones .

what new Crygames can , we can seeing by War of the Rights and Hunt :Showdown;)

https://youtu.be/A84LcBNhXq8

https://steamcdn-a.a...bm?t=1566909105

Battle of the Titans have multiple hit zones. They just don't have a central torso one. The game doesn't have convergence either, damage occurs through a dispersion cone.

Check the right corner. That is multiple hit zones, for something that is a quad, or is it six legs?





Something that is very cool --- you can slap different kinds of armor on different parts of the mech. You can slap armor that resistant to chemical as in missiles for the legs, you can slap armor that is resistant to kinetic on the body. You can slap a different armor on each of the torso, and also on the weapons. Light armor can gain your more speed, heavy armor can give you more HP but at you get penalties in speed.




Strangely enough, the crotch between the two legs is its own hit box.

Edited by Anjian, 02 December 2019 - 11:21 PM.

[Anjian]

Nesutizale, on 02 December 2019 - 11:46 AM, said:

Well I am allways facinated and scared at what he finds out next
But back to the topic of trajectory. I think there is an importend differance between giveing a gun a fixed path the bullet will follow to haveing actualy gravity and windresistance decide how the bullet acts. One is a very pretictable the other has to constantly be calculated anew. I did say that with the fixed stuff tehy could save a lot of computing power.

As for the Warships example from Anij....Wargameing seams to have found a solution for their engine for this very specific problem. Then again from some of the dev videos I have seen, they seam to have some quite good engineers.
Last video I saw was the "software" solution for Raytraceign that was still workable without RTX cards and that was quite clever thinking if you ask me.
Lets hope that either one at PGI comes up with a solution or that some modders will optimize the game in that regard.

Frankly MW5 needs further opimization. From some testing I made my system lost ~28 frames when effects where set to max and ~37 frames when post processing was at max. That is still quite a lot in my book.

I noticed that the Russians seem pretty good at coding and engineering like Wargaming. The game I used as an example above, Battle of the Titans, are also coded by Russians.

[Nesutizale]

From my old man, who had to work with some russians, he said that most of them had to start with far less advanced tech so to get around the limitations they had to find creative solutions. That resulted in way better usage of the given hardware and much better codeing. That was some 20 years ago but maybe they still work with the philosophy of "it has to run on a calculater too"

[Vellron2005]

I think most people here only focus on the "weapon balance" part of the potential MWO2... but that isn't why we want MWO2..

Weapons work fine in MWO (for the most part, there's always hitreg issues). I don't think it would be any different in MWO2..

I want MWO2 in UE4 engine because the game would look better. AND I want proceduraly generated maps which the ultra-meta guys can't learn by heart and exploit every rock and cranny.

THAT is what is missing greatly from MWO, and why a port to UE4 would be good.

Not to mention that Cryengine has a bunch of hardware issues..

[Nesutizale]

I agree that the weapon balance in MWO works. Not allways the best or it can be better but overall its pretty solid. The problem rises more from the fact that Clantech, for PvP with mixed teams and equal team sizes, had to be brough in line with IS tech. That robbed some of the lore and feeling of the game.

The map generator is indeed something that speaks for using MW5 as a basis for a new MWO / MWO 2.0. The lack of new maps or better mission types is what made MWO stale over time. Sure I had my mechs to level but I think around 80 or so skilled mechs I finaly lost interest. Got to 90+ mechs after that but at least 15 or so are not fully skilled and it wasn't because of bad game balance but because there was no reason why I should. Every mech was one of the two or three basic types of Meta mechs. That is also why I welcome the new Mechlab, finaly it makes sense to have so many different mechs because each chassie will come with its own best config instead of all beeing the same.

As for issues, well UE seams to come with its own bag of issues.

--- Blue bird news ---

Beside all that, I got some more infos about the reasons of why there might be problems with MP with the current way the game is build.

Quote

The biggest difference is we need to account for the simultaneous fire of up to 14 independent weapon systems on a 'Mech with unique weapon trajectories (since they are not casting from the camera like in many other MP games and all 'Mech weapon placement is different.)

Sure we can gain overhead in a number of ways to amp up weapon effects. Heavily restrict how many of certain weapons a 'Mech can use, prevent simultaneous weapon fire/force chain fire to cut down on the number of individual projectiles needing to be tracked at once.
Broken up "abilities" is how a lot of other MPs get away with heavy effects weaponry. But many of those options are pretty much the opposite of what many people want to see in a MechWarrior game.

So we end up having to design around the ability to simultaneously alpha your full arsenal at any time. Something that is fairly unique to the MechWarrior franchise over other shooters out there. But it comes at a cost that does mean we need to be mindful of certain things.

I hope we will, after the release, hear a bit more about the development and their internal thoughts on things like this. I don't know about you guys but I like to hear more of the development process behind MW5.