[WIP] Dangerously Shaded Stock Refresh

[DangerouslyDave]

On 1/11/2016 at 10:28 AM, NathanKell said:

...the issue here is that the combustion chamber is where it gets hottest, but that's very strongly regeneratively cooled. The nozzle throat has a bit less cooling and glows, and the glow gradually fades out as you go down the nozzle because the flow expands and cools, thus doubly reducing heating. 

Huh! I'll have to go searching for some image examples of the heat falloff from the nozzle throat to the combustion chamber. Is it just a hard line?

After more trouble than I care to admit, I finally got the engine gimbal to work properly in game. Thanks again to Ven for posting his Unity folder online. Looking through his cfg files was the secret sauce.

Question to the collective knowledge base: is there support for any sort of soft deformation in-game? For example, in the model above I'd like to be able to constrain the top of the main fuel line going into the tank to a single point in space, and have the tube deform as the engine gimbals. I imagine it's a pipe dream, but it's worth asking.

[CobaltWolf]

1 minute ago, DangerouslyDave said:

Huh! I'll have to go searching for some image examples of the heat falloff from the nozzle throat to the combustion chamber. Is it just a hard line?

After more trouble than I care to admit, I finally got the engine gimbal to work properly in game. Thanks again to Ven for posting his Unity folder online. Looking through his cfg files was the secret sauce.

Question to the collective knowledge base: is there support for any sort of soft deformation in-game? For example, in the model above I'd like to be able to constrain the top of the main fuel line going into the tank to a single point in space, and have the tube deform as the engine gimbals. I imagine it's a pipe dream, but it's worth asking.

Oof. You just convinced me that it would be worth it to make animated gimbals.

[NathanKell]

Afraid I don't know about deformation.

Regarding emissives...pinging @ferram4 who probably knows better than I. I *think* there's a fairly sharp fade in (from non to emissive) from the bottom of the combustion chamber to the throat, then a gradual fade out from the throat to some bit down the nozzle proper, that place being somewhere before the nozzle proper ends (and before the extension begins, if there is one). It's also vaguely possible that the start of the extension might glow, although I don't know for sure--that would be the first area not regeneratively cooled, and the exhaust might still be hot enough.

That assumes a regeneratively-cooled nozzle (with or without a radiatively-cooled extension). For the ablatively-cooled AJ10s, it might well have been different.

[blowfish]

11 hours ago, DangerouslyDave said:

Question to the collective knowledge base: is there support for any sort of soft deformation in-game? For example, in the model above I'd like to be able to constrain the top of the main fuel line going into the tank to a single point in space, and have the tube deform as the engine gimbals. I imagine it's a pipe dream, but it's worth asking.

This might work: attach the vertices in the top of the fuel line to a bone which is not parented under the gimbal.

EDIT: I'm actually mostly sure this will work now.  I believe it's the same way the gimbal on the Panther works.

Edited January 14, 2016 by blowfish

[ferram4]

So, as for visible emission along the length of the rocket combustion chamber-throat-nozzle length... it depends on the structure.  Strictly speaking, the gas will be hottest in the combustion chamber and coolest at the end of the nozzle (for relative definitions of cool), but that won't directly control emission.  The next step up is heat flux into the walls, which will be highest at the throat (large mass flow and flow density, good amount of flow velocity encouraging convective heat transfer), next highest in the combustion chamber (not much movement through most of it, but a lot of heat), lowest at the nozzle exit (lowest temperature, low density flow), and in between those points it should smoothly vary.

Next is trying to figure how hot any of those parts get, based on how much heat is taken away.  For most rockets, the throat itself actually probably shouldn't glow, given that there's a lot of structure around it (generally, the bottom of the cylinder just before the expanding section is where the throat is, hidden away) and that it is regeneratively or ablatively cooled.  The expanding portion of the nozzle right after it, however, probably should, though that should fade away as the nozzle expands and the gases cool.  I suspect that @NathanKell is correct that the glow should increase when transitioning to a nozzle extension, if and only if there is no ducting of turbopump exhaust into the nozzle to cool the extension, like on the F-1.  Perhaps some bit of glow just above the nozzle throat, right where the combustion chamber starts going into the converging section of the nozzle before the throat, but it should be very subtle.

In any case, I would argue that if you can get the heating animation correct, emissives should start right after the throat for standard operation, but overheat should have the throat and combustion chamber start to glow more and more.  Also, note that if your design violates any of the above assumptions in the structure of the rocket engine, then thing should probably glow differently.

[RaendyLeBeau]

19 hours ago, DangerouslyDave said:

Thanks again to Ven for posting his Unity folder online.

hmm sorry my question . someone can give me a hint for VEN's link ?

thanks in advance, cheers

 

Edited January 14, 2016 by RaendyLeBeau

[Nertea]

23 hours ago, DangerouslyDave said:

Question to the collective knowledge base: is there support for any sort of soft deformation in-game? For example, in the model above I'd like to be able to constrain the top of the main fuel line going into the tank to a single point in space, and have the tube deform as the engine gimbals. I imagine it's a pipe dream, but it's worth askin

It's not too tough. Just create a few bones or dummy objects to act as them. Parent the ones you want to use to link to the gimballed mesh to the GimbalBone. Then, split off the parts you want to deform (typically just the pipes) and skin them as you would any kind of human/organic/deformable mesh in your modeler. Unity should import that mesh as a SkinnedMesh, and voila, deforming pipes! I'm not sure what your modeler is so I can't be much more specific, as 95% of the work is done in there.

Nice texture work btw

[riocrokite]

@DangerouslyDave

alternatively to skinning you can do something similar to what Shadowmage have done with his engines (SSTU mod). In other words two lines connected with a thicker center that hides endings of those lines pointing at each other (or pointing at the center section, dunno exactly). Or a thin line ending in a thicker line (so you can get some angling without endings sticking out too much). Hope that helps:)

 

 

[DangerouslyDave]

12 hours ago, Nertea said:

It's not too tough. Just create a few bones or dummy objects to act as them. Parent the ones you want to use to link to the gimballed mesh to the GimbalBone.

This worked perfect, thanks! I didn't know that KSP would work with skinning and bones that way. Here's how it looks in game:

@NathanKell @ferram4 Thanks for the breakdown of the heat emissives! Here's my revised attempt:

I've left the end of the nozzle completely dark, as there is an aspirator on this model.

Lastly, here's a link to a snapshot of my working document for the project roadmap:

http://imgur.com/fKo1T42

My goal is to first complete a full set of 1m parts, including an upgrade path for each of the parts listed in the image above. I'll reevaluate whether to proceed with the 2m and 3m parts from there, or whether to switch gears to utilities and science parts. If you have any suggestions for substitutions or rearrangements, please don't hesitate to post them. Also please note that the parts listed are for visual guides only, so some might seem over/underpowered for the parts they're replacing. My idea right now is that the Tier 1 (Refresh) parts will be slightly "lower", and the Tier 2 (Upgrade) parts will by slightly "higher" in stats compared to the stock parts. I'll probably be requesting help with balancing everything when we get closer to that stage. Thoughts?

 

[DangerouslyDave]

17 hours ago, RaendyLeBeau said:

hmm sorry my question . someone can give me a hint for VEN's link ?

thanks in advance, cheers

 

Whoops, forgot to answer. He provides a link to his Unity folder on the first page of the Dev thread:

http://forum.kerbalspaceprogram.com/index.php?/topic/83696-105-stock-part-revamp-update-191/

[*Aqua*]

On ‎10‎.‎01‎.‎2016 at 8:46 AM, DangerouslyDave said:

I'm not sure if I like this look. The metal looks like aluminum. Does this engine look like that in RL?

Also I have the impression that they usually have a less shiny apperance.

Spoiler

 

Russians like to paint their rocket engines. O_O

Spoiler

 

Also for emissives I found these pictures:
 

Spoiler

I hope it can be of help for you.

[TiktaalikDreaming]

22 hours ago, DangerouslyDave said:

This worked perfect, thanks! I didn't know that KSP would work with skinning and bones that way. Here's how it looks in game:

@NathanKell @ferram4 Thanks for the breakdown of the heat emissives! Here's my revised attempt:

I've left the end of the nozzle completely dark, as there is an aspirator on this model.

Lastly, here's a link to a snapshot of my working document for the project roadmap:

http://imgur.com/fKo1T42

My goal is to first complete a full set of 1m parts, including an upgrade path for each of the parts listed in the image above. I'll reevaluate whether to proceed with the 2m and 3m parts from there, or whether to switch gears to utilities and science parts. If you have any suggestions for substitutions or rearrangements, please don't hesitate to post them. Also please note that the parts listed are for visual guides only, so some might seem over/underpowered for the parts they're replacing. My idea right now is that the Tier 1 (Refresh) parts will be slightly "lower", and the Tier 2 (Upgrade) parts will by slightly "higher" in stats compared to the stock parts. I'll probably be requesting help with balancing everything when we get closer to that stage. Thoughts?

 

Wow, that's good excrements.  I had no idea the constraints etc would import through Unity and out to KSP so nicely.  Thanks @Nertea.  I'd been doing what @riocrokite suggested, aka, fudging it.  Fine _mostly_, until it looks odd.  This is MUCH better.  And those engines are looking fabulous.

[TiktaalikDreaming]

18 hours ago, *Aqua* said:

I'm not sure if I like this look. The metal looks like aluminum. Does this engine look like that in RL?

Also I have the impression that they usually have a less shiny apperance.

  Reveal hidden contents

 

 

 

  Reveal hidden contents

  Reveal hidden contents

 

http://heroicrelics.org/ussrc/engines-a-7-indoors/dsc11488.jpg.html

We're all used to a much more detailed outside of nozzles than the early systems of basically a dual wall with booze pumped between the walls.

As to paintings, a lot of museum pieces will have extra paint to prevent corrosion and sometimes to highlight different systems, eg: blue for oxidizer, red for fuel etc.  A lot of that wouldn't be there on live production pieces.  I'm not saying the Russians don't paint their engines, but a museum piece isn't demonstrating it clearly.

[DangerouslyDave]

18 hours ago, *Aqua* said:

I'm not sure if I like this look. The metal looks like aluminum. Does this engine look like that in RL?

Also I have the impression that they usually have a less shiny apperance.

Thanks for the feedback! I agree that I may have missed the mark on the first outing trying to texture these things. In fact, one of the things coming up on my to-do list is to create a material library to use across all the parts so I can keep a consistent look. I'll post a set of images once I have something cohesive down to see what people think.

I've been using heroicrelics, historicspacecraft, and b14643 as my main reference sources. They seem to show a variety of different materials, from matte to very glossy, even used on different builds of the same engine. I guess that's chalked up to aging and, as @TiktaalikDreaming said, some intervention on the museums' part to make them more archival.

33 minutes ago, TiktaalikDreaming said:

We're all used to a much more detailed outside of nozzles than the early systems of basically a dual wall with booze pumped between the walls.

I may wind up reducing some of the detail even as I move on to more modern engines. My goal isn't to make photo-realistic parts, but to use the real-life analogues as a guide while preserving the stock KSP look. To my eye, that means simplifying some of the forms and doing away with some of the fine details. In general, I'll be trying to at least include the major fuel and exhaust lines.

Here's the next engine I'm working on. A replacement for the LV-909  based on the Bell 8096.

I'm looking for some feedback regarding the design on this one. As you can see, the combustion chamber and turbo pump extend well beyond the attachment ring. The stack node will be in-line with the ring, and the collision mesh will only include geometry from the ring down (the footprint is nearly identical to the stock LV-909). My idea here is that for designs where the engine is attached to a 1m fuel tank, the engine will look similar to stock, with only the nozzle showing. However, if you wind up attaching to something like a strut/girder segment, the engine will be visible. The only thing missing from the design at this point is some sort of foil/cloth cover so that you can't see the engine clipping through the fuel tank if it's stacked directly below one.

I can't find clear reference of the bottom of an Agena stage. Any suggestions as to how the cover should look? (Keeping in mind that the engine will gimbal while the exhaust will remain stationary.)

[Phineas Freak]

@DangerouslyDave here you can find some good info and pictures about the Agena.

[jonrd463]

14 hours ago, DangerouslyDave said:

I may wind up reducing some of the detail even as I move on to more modern engines. My goal isn't to make photo-realistic parts, but to use the real-life analogues as a guide while preserving the stock KSP look. To my eye, that means simplifying some of the forms and doing away with some of the fine details. In general, I'll be trying to at least include the major fuel and exhaust lines.

I agree 100% with this design philosophy. If detail is needed, let the texture do the work. Given the shader support, bumps, normals, and other mapping tricks will do well for getting the surface features on while preserving poly counts. Besides, Stock KSP maintains a certain stylistic elegance in its simplicity. I can see perhaps upping the detail to be more in line with real parts in a Realism Overhaul branch of the mod, but the detail level I'm seeing is just fine for the stock game.

As for the engine finishes, brand new Merlins are pretty shiny. What I'd like to see is perhaps an evolution of the heat shield shader. Start the engines off nice, clean, and pristine, and once they've done a burn through an atmosphere, have them be all sooty and dull, especially for KeroLox. See the returned Falcon 9 boost stage for reference.

This:

vs. this:

All in all, looking fantastic, @DangerouslyDave!

Edited January 16, 2016 by jonrd463

[sashan]

Here's a good simulation of temperatures. That's NERVA, but Laval nozzle is still same. Parts with active cooling shouldn't glow tho.
 
Also, here is radiatively cooled Merlin:

And a good video of Draco firing: https://www.youtube.com/watch?v=Z5j5exxNZFw
Basically, the narrowest part of the throat is the hottest, unless cooled.
I really like your engines, will you release them as a separate branch? Ven's revamp already adds some selection for engines, but still not enough for my tastes. We expecially need small vacuum engines and small high-thrust atmospheric ones.

Edited January 16, 2016 by sashan

[sashan]

17 hours ago, DangerouslyDave said:

I'm looking for some feedback regarding the design on this one.

I really like it, but could you extend outer ring upwards? SO it'll be partially covered with shell aerodynamical engine. It'll make all those nice parts visible and at the same time it'll look aerodynamic. Clipping it? DOn't like it much, aside from unrealistic problem of "where is that fuel actually" it won't look good as clipping would be visible.
IMO the cover should look like any interstage, with those vertical groves.

[DangerouslyDave]

5 hours ago, jonrd463 said:

As for the engine finishes, brand new Merlins are pretty shiny. What I'd like to see is perhaps an evolution of the heat shield shader. Start the engines off nice, clean, and pristine, and once they've done a burn through an atmosphere, have them be all sooty and dull, especially for KeroLox. See the returned Falcon 9 boost stage for reference.

This is a fantastic idea! The thought briefly crossed my mind but I didn't think it would be possible. I'll definitely be taking a close look at how do achieve this. Although I don't have much programming experience. Can you/anyone point me in the right direction?

2 hours ago, sashan said:

I really like your engines, will you release them as a separate branch? Ven's revamp already adds some selection for engines, but still not enough for my tastes. We expecially need small vacuum engines and small high-thrust atmospheric ones.

Thanks for the images, that's a helpful visualization of what NathanKell and ferram4 were describing. I'm definitely open to releasing the parts in branches. It's still pretty early in the process, and I'll need some help down the line with balancing the parts and figuring out how everyone wants things packaged up.

2 hours ago, sashan said:

I really like it, but could you extend outer ring upwards? SO it'll be partially covered with shell aerodynamical engine. It'll make all those nice parts visible and at the same time it'll look aerodynamic...
...IMO the cover should look like any interstage, with those vertical groves.

I'm having a bit of trouble making out exactly what you mean, but the part is kind of locked in with regards to the outer ring, as I'm trying to keep the bounding box the same as stock. Your point about making more of the details visible makes sense, though. How's this as a compromise:

Pardon the messy mockup, but it gets the idea across. Basically, I squash the engine vertically, bringing more of the details into the visible part below the ring. It moves things pretty far away from the RL analogue, though (especially considering I already had to move some of the upper assembly parts around to fit them between the stabilization bars). Thoughts?

In the meantime, I went ahead and did a quick material pass on this guy:

I don't know if I'm totally happy with this one. I'm having some difficulty breaking the parts up with materials while making sure everything reads as metal. I'm thinking I need to define 3 types of metal (varying glossy/matte,light/dark) and a couple glossiness levels of paint and stick to them. After this engine is at a good resting point, I think that's next (along with fuel tanks).

[NathanKell]

Please at least save a copy of the 80xx form of the engine! (And keep the mounting ring as a second mu added via MODEL, so we can get rid of it )

Really, though, if you want that form-factor, RD-0105/9 would serve better (or S1.5400). The Bell engine was very, very tall for its width.

 

Oh also: yes, totally possible to either darken the texture based on a module-passed shader var (the heat shield approach) or blend between two different textures (asmi's approach for BobCat's Soyuz RM, it blended between the regular diffuse texture and a scorched one).

[DangerouslyDave]

After further thought, I'm kind of attached to the look of the full Bell engine. Once the cover is in place and/or its part of a larger assembly, it shouldn't look so long in the tooth. I've done a quick sculpt and normal overlay as a test for the cover when radially attached. Please ignore the seam, this is for mockup only.

Whats the general opinion of something like this, versus something closer to a stock interstage look, as @sashan suggested?

[Jimbodiah]

Meh, all the details are lost. Why animate everything if you will hide it?

[Bedi]

10 hours ago, DangerouslyDave said:

This is a fantastic idea! The thought briefly crossed my mind but I didn't think it would be possible. I'll definitely be taking a close look at how do achieve this. Although I don't have much programming experience. Can you/anyone point me in the right direction?

I'm not the modder, but is't the stock heatshields do the similar thing? Maybe looking into heatshields will help you get some ideas...

Edit: oh sorry @jonrd463 already mentioned it in his suggestion

 

Edited January 17, 2016 by Bedi

[*Aqua*]

It shouldn't be that difficult. Basically you need two textures for it: The "clean" and the "dirty" one.

You put both on the heatshield, the dirty one on a lower layer and the clean one in the layer above it. (You can archieve that by duplicating the surface and moving the copy before the other one. But be aware of possible z fighting.) On start you set the the clean one to 100% opacity. While the heatshield burns up you decrease the opacity so that the burned up surface begins to show through. You can tie the opacity to the ablator_current / ablator_maximum ratio or use a more refined math.

There may already be a function in Unity which can combine two textures to archieve the same effect.

[sashan]

15 hours ago, DangerouslyDave said:

'm having a bit of trouble making out exactly what you mean,

Take that outer structural ring that is supposed to be attached to fuel tank, and extend its upper edges exactly vertically with unchanged radius to cover the top of the engine. Then, add attach node right where that extruded "tube" ends.