Could KSP Switch To Using Unreal Engine 4?

[jrolson 34]

Now that the Unreal Engine 4 is now free, could they switch from using Unity?

http://www.neowin.net/news/unreal-engine-4-is-now-free-to-download-for-everyone

[Tex_NL 2,203]

Sure they could. But they won't. It would require a complete re-write of the entire game.

[jrolson 34]

Sure they could. But they won't. It would require a complete re-write of the entire game.

Figured that, but the graphics would be mind blowing!

[sal_vager 5,826]

Figured that, but the graphics would be mind blowing!

Why would they? Squad would reuse the existing assets, all that might change is the shaders used.

Unity3D is capable of looking good, and UE4 is just as capable of looking like junk.

[jarmund 136]

As an indie game dev i can say, with the fullest confidence that:

IT CAN!

but

IT WON'T!

I ported a game from one engine to another once (from AndEngine to libGDX) and it was so much painful work of reinventing the wheel for the third time that i wished that my hobby included drilling holes into wooden toy horses just for the sake of having a simple job instead.

[modedit: laaaaanguage!]

Edited March 2, 2015 by technicalfool

[cpast 218]

"Switching engines" is somewhat misleading terminology, because it makes it sound reasonably doable. "Porting" is better -- there is substantial rewrite involved in the change, and a huge amount of the knowledge you gain while writing the game is lost. It's just about never worth it.

[Kerbital 536]

Pitty, because we're now in 2017 and the game is terribly buggy and stutter and performance drops are terrible. I will probably avoid any game based on Unity from now on. Too bad, such an awesome game hampered by awful engine.

[radonek 311]

…except KSP is likely to break any game engine same way. Just because game engine is good at pushing polygons does not mean it can handle physics in multiple moving reference frames at distances that make FPU hide in shame. Just look at space engineers - they have all the fancies, destructible models and whatnot - and universe-wide "speed limit" of 100m/s because their physics just breaks there.

[Gaarst 3,115]

^ This.

KSP in unlike many games as it needs intense physics processing. Even with a "better" engine you will still have the problems that the engine was not build to handle hundreds of parts interacting with each other in many ways at a time.

A dedicated engine would be the best solution, but this is not happening.

[p1t1o 3,224]

What @Gaarst said.

What would be perfect for KSP would be a bespoke-created physics engine for simulating vast volumes of space with high resolution.

One did not exist, and creating one from scratch would have been prohibitively difficult (for starters you could add 2 or 3 years to KSP development time, double the cost to cover the salaries of the people they would have needed to hire to create it, and then add a bunch more cost on top of that because there would be few customers wishing to license the use of their engine.)

Porting from one non-ideal engine to another non-ideal engine would be a lot of work for nothing.

Squad did not make the mistake of "choosing the wrong engine", they made the "mistake" of building a game/sim that went beyond all contemporary expectations of what you could do in a game/sim world.

One should not say "Squad should port KSP to a differnt engine" or "If only Squad didn't choose unity..."

One should say "Gosh it never fails to impress me how much Squad achieved in KSP with the limitations of the physics engines available to them."

Seriously, read about floating point errors and how KSP handles them, its quite clever.

[Sereneti 41]

On 20.1.2017 at 11:27 AM, p1t1o said:

What @Gaarst said.

What would be perfect for KSP would be a bespoke-created physics engine for simulating vast volumes of space with high resolution.

One did not exist, and creating one from scratch would have been prohibitively difficult (for starters you could add 2 or 3 years to KSP development time, double the cost to cover the salaries of the people they would have needed to hire to create it, and then add a bunch more cost on top of that because there would be few customers wishing to license the use of their engine.)

Porting from one non-ideal engine to another non-ideal engine would be a lot of work for nothing.

Squad did not make the mistake of "choosing the wrong engine", they made the "mistake" of building a game/sim that went beyond all contemporary expectations of what you could do in a game/sim world.

One should not say "Squad should port KSP to a differnt engine" or "If only Squad didn't choose unity..."

One should say "Gosh it never fails to impress me how much Squad achieved in KSP with the limitations of the physics engines available to them."

Seriously, read about floating point errors and how KSP handles them, its quite clever.

 

[NSEP 7,149]

Yeah good idea, let everything that is left over from the KSP dev team work their butts of rewriting the entire game just for better polygons.

Switching engines is not a good idea. Especially a game that has a big modding community, like KSP.

Its a no just no thing.

[TheKosanianMethod 337]

Like everyone else is saying, the switch would be too much work for too little reward. Not as many people can work with Unreal, while many can use Unity. Unreal just seems too... unreal.

[Spacetraindriver 990]

KSP is praised for its highly realistic orbital mechanics, not its graphics.  I mean even if they did upgrade, its not like KSP has grass, dust storms, stock clouds, super high poly and definition rocks and trees...I mean most of the game is in space. What would you need that needs graphical improvements in space?

[JAFO 472]

On 1/20/2017 at 8:16 AM, Kerbital said:

the game is terribly buggy and stutter and performance drops are terrible.

Umm.. what?

KSP 1.2.2 is the least buggy, and best performing version ever released..

What have you been smoking? Because I want to be sure I never try it.

[Red Iron Crown 12,473]

8 hours ago, Spacetraindriver said:

KSP is praised for its highly realistic orbital mechanics, not its graphics.  I mean even if they did upgrade, its not like KSP has grass, dust storms, stock clouds, super high poly and definition rocks and trees...I mean most of the game is in space. What would you need that needs graphical improvements in space?

Unity can be quite good graphically, an engine change would not be needed to improve KSP visuals.

[KAL 9000 1,685]

Necro alert!

[Red Iron Crown 12,473]

4 minutes ago, KAL 9000 said:

Necro alert!

Topic is still relevant, and you've been around long enough to know better than to necromoan. 

[KAL 9000 1,685]

Just now, Red Iron Crown said:

Topic is still relevant, and you've been around long enough to know better than to necromoan. 

True, I guess . Was kinda joking.

[Kerbital 536]

5 hours ago, JAFO said:

Umm.. what?

KSP 1.2.2 is the least buggy, and best performing version ever released..

What have you been smoking? Because I want to be sure I never try it.

I'm not smoking anything. Of course, it's the "least buggy" and "the fastest". It's the latest version, duh! But that doesn't mean it's not buggy. I's like Apple saying "this is the fastest Mac ever!". Of course, it's the latest model. LOL. What kind of an argument is that?

I still get the RAM/GC related stutter which has been a known issue for over two years and is directly related to Unity crappiness. MemGraph helps somewhat. Without MemGraph there is a solid 1sec pause every 5-7 seconds. 250+ parts stations with smaller ships docked drop FPS to around 10 near planets. Insane loading times as you have more and more ships in flight. Texture seams and smearing everywhere. KSP is the most buggy game that I'm currently playing.

KSP needs a "performance update". The way Uber did with Planetary Annihilation. They fixed the game from barely playable to 6+ AIs and thousands of units running smooth like butter with all graphics options on. 

Edited January 25, 2017 by Kerbital
splleing

[radonek 311]

3 hours ago, Kerbital said:

I still get the RAM/GC related stutter which has been a known issue for over two years and is directly related to Unity crappiness.

It's a Mono GC crappines and it's a moding issue, KSP itself works perfectly well for me (or at least did before I stuffed with mods too).

4 hours ago, Kerbital said:

KSP needs a "performance update". The way Uber did with Planetary Annihilation. They fixed the game from barely playable to 6+ AIs and thousands of units running smooth like butter with all graphics options on. 

You must be kidding, vanilla PA is unholy mess as far as I can see. Main menu is probably worst UI in known universe, matchmaking is downright painful experience and game runs like a charm, except its a special kind of charm that breaks after an hour or two. True, it's much less buggy then it used to be (it can run whole hour before falling flat on face, can you belive it? whole hour!) but that's hardly comparable to progress that KSP made.

[stenole 20]

On 2.3.2015 at 9:25 PM, jrolson said:

Now that the Unreal Engine 4 is now free, could they switch from using Unity?

There is a 5% royalty on UE4 if you are using it under the standard EULA. So it is definitely not free.

[Kerbital 536]

1 hour ago, radonek said:

It's a Mono GC crappines and it's a moding issue, KSP itself works perfectly well for me (or at least did before I stuffed with mods too).

You must be kidding, vanilla PA is unholy mess as far as I can see. Main menu is probably worst UI in known universe, matchmaking is downright painful experience and game runs like a charm, except its a special kind of charm that breaks after an hour or two. True, it's much less buggy then it used to be (it can run whole hour before falling flat on face, can you belive it? whole hour!) but that's hardly comparable to progress that KSP made.

If KSP run four ships for me as well as PA runs 6-8 AIs and thousands of units I will be a happy camper. I played 3 hour battles with no slowdowns after the update. I have 6700K, 32GB of RAM and two GTX1080 and no other game drops below 20 fps with only few hundreds of objects/parts in view and the CPU is not even at 50% when that happens. I know it's probably mods, but without mods KSP is a 10 hour game at most. It's mods that keeps KSP interesting. My PA is also heavily moded, by the way.

[magnemoe 2,196]

On 20.1.2017 at 0:16 AM, radonek said:

…except KSP is likely to break any game engine same way. Just because game engine is good at pushing polygons does not mean it can handle physics in multiple moving reference frames at distances that make FPU hide in shame. Just look at space engineers - they have all the fancies, destructible models and whatnot - and universe-wide "speed limit" of 100m/s because their physics just breaks there.

This, fun how the same discussion show up in the elder scroll forum all the time too, and how switching engine would solve a lot of problem magically like some windows 7 and 8 computers with an buggy os worked better after upgrading to windows 10. 

For KSP lots of the core mechanic would have to be rebuild. Totally uninteresting unless they make an KSP 2.0 and most would be very unimpressed by the new version as all the changes was under the hood. It would also create as many new errors or more than it fixed. 

 

[Enceos 1,749]

KSP is a semi-indie game with a small team of developers. Instead of working with Unity guys they have to report bugs and invent workarounds. Chris Roberts also hit a wall in the development of Star Citizen and was quick to find out that he needed to work with the Crytek together to overcome their engine limitations. As I see it, the only way KSP can pump up its performance is another inventive workaround or a tight work with Unity engine guys. The second option lies in the hands of the owners of Squad, because it is a financial decision. KSP is their first game and they certainly don't have experience in game development like Chris Roberts which makes them reluctant in funding things they don't understand. It usually takes an experienced, respected, eloquent and persistent developer / engineer to team up with the producer / project manager on a mission to crack the mind of the CEO and persuade him to make a financial decision. One can only hope for such a person to exist.

Edited January 27, 2017 by Enceos