Developer Insights #18 - Graphics of Early Access KSP2

[RocketRockington]

7 hours ago, aNERVASman said:

no there are playable tutorials too

That means that any changes to the game may incur an even bigger development cost because a tutorial may need to be validated or even  updated to match it - another reason not to do your tutorials too early.  Also I seriously doubt they did all thier tutorials before 2019, when they couldn't even hold their studio together to ship a game.

[Jefftheguyperson]

Excellent explanation of the working going on under the hood! It makes me feel much more comfortable about the direction of the game.

On 3/10/2023 at 4:00 PM, Intercept Games said:

As a sidenote, despite how much we love you modders, this change will definitely break most visual mods (sorry modders, sometimes we must hurt the ones we love). 

Do you mean that visual mods will be broken permanently? Or just that it makes visual mods more difficult to implement into the game?

[Periple]

15 minutes ago, Jefftheguyperson said:

Do you mean that visual mods will be broken permanently? Or just that it makes visual mods more difficult to implement into the game?

Any visual mods made on the current system will break! You can always port them or make new ones afterwards. (In fact it ought to be easier to make them with HDRP!)

Edited March 15, 2023 by Periple

[Tw1]

On 3/11/2023 at 8:00 AM, Intercept Games said:

The next-gen CBT system will be the topic of a future dev blog which will contain a much more detailed look at what we’re building

Will this produce identical planets to PQS+? Or will there be changes?

[ElliotWizerd]

I know one thing that peopel do not think about mods and vanila gameplay is that a moder do not need to have the same type of capabilitys as like run on many type of software like on a console. the vanila game is coded so it can be run on everything. that dose make the hole difftent type of level to make it to work with no problem

[ElliotWizerd]

16 hours ago, RocketRockington said:

That means that any changes to the game may incur an even bigger development cost because a tutorial may need to be validated or even  updated to match it - another reason not to do your tutorials too early.  Also I seriously doubt they did all thier tutorials before 2019, when they couldn't even hold their studio together to ship a game.

i belive that they have already made them in that i minde so to not take up recurse from the team

[ElliotWizerd]

4 hours ago, Tw1 said:

Will this produce identical planets to PQS+? Or will there be changes?

yes it will i think. the redering in the background is going to change alot as i remeber but there is always way to make the asets usebale in a diffren rendering engine

[Jefftheguyperson]

19 hours ago, Periple said:

Any visual mods made on the current system will break! You can always port them or make new ones afterwards. (In fact it ought to be easier to make them with HDRP!)

Awesome, that's great to hear. Looking forward to seeing what people can create, in that case!

[Space Kadet]

On 3/10/2023 at 11:00 PM, Intercept Games said:

As many of you have noticed, KSP2’s performance isn’t amazing at the start of Early Access

that statement is kind of like me saying my grandmother whos been dead for years 'has some health problems'....

On 3/10/2023 at 11:00 PM, Intercept Games said:

The machine I’m using here for performance analysis is a laptop with i7-8650U CPU, Mobile Nvidia GTX 1060 6gb GPU and 16gb RAM. It has a slower GPU than our current min spec, so we’re not expecting it to make a playable framerate yet. 

i've tried this on i5 11500, 2060super with 2gig and got 15 fps at launch, ive then tried it on both i911900k 4070ti 64gig. then i9 13900k 4090 both 64gig and got 15 to 18 fps at the same place. I built the big rigs specifically for ksp. 

can you explain why there a 8 fps difference between your below minimum spec laptop and my rigs? Aslo why the fps problem that was blatent in the pre alpha footage is still present 3 years down the line?

[Palshife]

On 3/19/2023 at 9:35 AM, Space Kadet said:

can you explain why there a 8 fps difference between your below minimum spec laptop and my rigs? Aslo why the fps problem that was blatent in the pre alpha footage is still present 3 years down the line?

They explained. The old system was extended, but has bottlenecks and is being replaced. The new system will be faster and prettier.

[InterstellarDrifter]

Hmmm, my 12900k 4090 ddr5 system runs ksp2 pretty well, most of the time my fps sits at or just below my 60fps@4K frame cap, even at launch..pre-patch, and especially on 0.1.1     

I think some might be describing an experience outside of expected use cases, like 200 part, multi control point monstrosities.

For instance watch this silly goose(https://www.youtube.com/watch?v=wJG03fnymUY) try to launch a 180 part satellite constellation, lol. Then wonders why FPS tanks...huuuhhh

KSP does not "play" like an arcade game like most expect it to. It presents like this weird juvenile arcady space game, but in reality is more like solidworks/blender, than a game, it more like "runs" like a solver/renderer/CFD with similar limitations. So, its not really their fault.  Because of this cartoony feel to this simulator, people expect that they can go build whatever their heart desires because the VAB will allow that freedom. But then are often met with much friction(such as, bad fps and craft performance) because its simulating reality to its best, reasonable extent.  KSP is much more enjoyable when approached with realistic expectations and understanding of its limitations, especially with somewhat raw software.

 

[dr.phees]

On 3/28/2023 at 2:55 AM, InterstellarDrifter said:

outside of expected use cases, like 200 part, multi control point monstrosities.
 

Framing large craft as silly will not make KSP 2 run better. Leave it to others how they want to play the game.

Edit:

By the way: That "silly" launch is something that I have also done regularly in KSP 1: Launch a couple of relay satellites in one launch (usually 3 to 6). Nothing special here, works fine, and is basically realistic. Maybe not the 6 boosters, but also not that far off. If you want to stay strictly realistic with low part counts, then play Juno. If you want complex crafts with may parts (the science parts alone add up to quite some part count on science missions!), KSP was the go-to game.

And regarding performance: KSP 1 is already doing what you describe here as "simulating reality to its best". And doing more of it. And better. And with better performance.

I don't understand why we cannot be realistic about this: KSP 2's performance is bad compared to KSP 1. Devs obviously don't want to compare it to KSP 1, hence the missing 'really low' graphics settings. OK. Maybe KSP 2 looks awesome, compared to KSP 1 with mods, each to his own. But pretending that KSP 2 was doing a better job of simulating physics is just not true.

Edited March 31, 2023 by dr.phees

[Space Kadet]

On 3/21/2023 at 6:18 AM, Palshife said:

They explained. The old system was extended, but has bottlenecks and is being replaced. The new system will be faster and prettier.

thats not an explination of why a blatent bug from 3 years before launch was still present. and saying 'oh now you have noticed we will make it newier and prettier' after 3 years of them supposidly playing the game every day and not seeing the several bugs it causes.....

but its not like they launched the game in that state at full price....

[Space Kadet]

On 3/28/2023 at 3:55 AM, InterstellarDrifter said:

Hmmm, my 12900k 4090 ddr5 system runs ksp2 pretty well, most of the time my fps sits at or just below my 60fps@4K frame cap, even at launch..pre-patch, and especially on 0.1.1     

I think some might be describing an experience outside of expected use cases, like 200 part, multi control point monstrosities.

For instance watch this silly goose(https://www.youtube.com/watch?v=wJG03fnymUY) try to launch a 180 part satellite constellation, lol. Then wonders why FPS tanks...huuuhhh

KSP does not "play" like an arcade game like most expect it to. It presents like this weird juvenile arcady space game, but in reality is more like solidworks/blender, than a game, it more like "runs" like a solver/renderer/CFD with similar limitations. So, its not really their fault.  Because of this cartoony feel to this simulator, people expect that they can go build whatever their heart desires because the VAB will allow that freedom. But then are often met with much friction(such as, bad fps and craft performance) because its simulating reality to its best, reasonable extent.  KSP is much more enjoyable when approached with realistic expectations and understanding of its limitations, especially with somewhat raw software.

 

Wied flex but ok. For me at launch , litterally the hour it launched, i9 11900 with a 4090 and 128gb ram was sitting at under 18fps all the way to almost space the ship was capsule chute tank and engine. and the reason i didnt get there was the capsule and engine came away from the tank for no reason before the crash. ironically the other rig i have runs just 3 fps slower and its an i5 11500, 32gb and a 2060 super.   the fps isint tanking because of the parts, nor is it tanking because of the power of the pc, its tanking because of bugs in the KSP 1 code the copied and didnt understand to do with the rendering of the ground.

the problem is that that bug was visible in the pre alpha footage that was released years ago, and if they havent fixed it in 3 years thats a scary image of how little effort was being put into KSP. Couple that with the money grab of asking full price for a game in this state, specifically against steam rules, that shows a dead horse being flogged for all its worth.  then the 'slow down' of updates, on a game that specifically they do not leagally have to update for the players.... what was made and sold sofar is what people paid for, no more. and it needs an extremly highend pc to be bareable.

Ill admit i dont know how it runs after the patches because i refunded that game after 3 hours of trying to get to space with everything falling appart, crashing to desktop, or more kracken hits than my entire time in KSP1 which is in the 10's of thousands of hours.  but looking at the comments below every steam update, it aint getting better for most.

[JoeSchmuckatelli]

@Dakota... Can we get a CBT / HDRP update into the comms pipeline?

 

Would really like to know more now that they have been playing with this for several weeks. 

[TomKerbal]

2 hours ago, Space Kadet said:

...its tanking because of bugs in the KSP 1 code the copied and didnt understand to do with the rendering of the ground.

Are you sure ? How do you know? Well, that would be great news for me, because I was thinking they threw all the good old code away and started from scratch, what would be a big mistage (from the view of an old-schooled developer).

Edited April 29, 2023 by TomKerbal

[cvusmo]

There's hardly any KSP 1 code if any within KSP 2. If anything the programmers used KSP 1 as a template to base KSP 2 off of. But they sure as heck didn't copy and paste a bunch of KSP 1 code in KSP 2 code. There's also a lot of features written in the code that isn't in the game yet. Why they do this? Because you need your namespaces setup, code organized, and when you add Science Mode or Colonies (ie the Content Updates) the content they've worked on and are adding, already has a common language with the base game, thus making it work far easily.

As far as how they render the terrain it's pretty neat. KSP 2 has all custom vegetation/terrain that they didn't have in KSP 1.

So no, KSP 2 isn't a copy and paste of KSP 1. They're very different.

[Bej Kerman]

36 minutes ago, TomKerbal said:

How do you know?

Like 99% of assertions on this forum, they do not. If anyone claims they know the developers did X or Y, just dismiss it. It's worthless blah blah unless it's backed up with evidence.

[cvusmo]

1 minute ago, Bej Kerman said:

Like 99% of assertions on this forum, they do not. If anyone claims they know the developers did X or Y, just dismiss it. It's worthless blah blah unless it's backed up with evidence.

You cannot share anything in this forum or anywhere of any decompiled dynamic link libraries. 

The US does allow for the decompilation of software for interoperability purposes (see: 17 U.S.C. § 1201 (2017) – Circumvention of Copyright Protection Systems - § 1201(f)(2)). It means that I can decompile the .DLLs to fix it on my end. But I cannot share the information that I decompile because it isn't my intellectual property.

So no, unfortunately I legally cannot back up any claims I make with evidence.

[Bej Kerman]

26 minutes ago, cvusmo said:

You cannot share anything in this forum or anywhere of any decompiled dynamic link libraries. 

The US does allow for the decompilation of software for interoperability purposes (see: 17 U.S.C. § 1201 (2017) – Circumvention of Copyright Protection Systems - § 1201(f)(2)). It means that I can decompile the .DLLs to fix it on my end. But I cannot share the information that I decompile because it isn't my intellectual property.

So no, unfortunately I legally cannot back up any claims I make with evidence.

I was mainly on about the person you were replying to.

[cvusmo]

4 minutes ago, Bej Kerman said:

I was mainly on about the person you were replying to.

EDIT: It's all good. I'm sharing the information because I'd rather people know the laws and rules about it, than assume things. Cheers!

It's also in the forum rules:

 

2.2 Forbidden content

Messages that involve the following content may result in moderator action.

1. Discussing or facilitating copyright infringement or software theft (software piracy);

Edited April 29, 2023 by cvusmo