While I haven't touched the Constitution model since my last post, I did start modelling a shuttle craft for it (Started it around the release of Blender 2.9, but ended up distracted and have only got back to it this past month).
So far I have finished the warp nacelles, and I'm currently slowly integrating the main hull pieces together (modeled it in 3 pieces).
One thing I'm currently undecided on is to have the warp nacelles parallel to each other, or have them toed in so they are parallel to the hull?
Very slowly progressing on the shuttle. So far I have finished integrating the the bottom hull and the side walls of the top hull into one mesh, and I've finished modelling the warp pylons/wings (I ended up going with the warp nacelles parallel to each other).
I've finished most of the work on the rear landing gear, but I may redo some of it to add some details I missed (mainly the teardrop shaped indent).
I also changed how I was doing the curve on the warp nacelles corrugated rear caps. I was using a boolean object to cut out the curve, but I realised that I could get a similar result using vertex groups and a shrinkwrap modifier. The great thing about this method is it doesn't create unneeded vertices, which should mean that there will be minimal or no cleanup when I apply the modifier stack.
Next part I'll be working on is integrating the middle hull part into the rest of the main hull. I should have done this earlier as it's going to be a real pain to do cleanly , since I've already freezed the subd of the main hull part.
Very slowly progressing on the shuttle. So far I have finished integrating the the bottom hull and the side walls of the top hull into one mesh, and I've finished modelling the warp pylons/wings (I ended up going with the warp nacelles parallel to each other).
I've finished most of the work on the rear landing gear, but I may redo some of it to add some details I missed (mainly the teardrop shaped indent).
I also changed how I was doing the curve on the warp nacelles corrugated rear caps. I was using a boolean object to cut out the curve, but I realised that I could get a similar result using vertex groups and a shrinkwrap modifier. The great thing about this method is it doesn't create unneeded vertices, which should mean that there will be minimal or no cleanup when I apply the modifier stack.
Next part I'll be working on is integrating the middle hull part into the rest of the main hull. I should have done this earlier as it's going to be a real pain to do cleanly , since I've already freezed the subd of the main hull part.
Haven't touched this in a while, but I do have some newer wip images, so might as well post them.
Modeled the rear details and added the engines. Not sure if I'll keep the glowing look, or go with something that better matches the full size prop - it appears to have rectangular tubes in the show, but the restored prop has the glowing panels.
Close up of the details behind the vent grill. It's mainly all geometry, but the holes and lights are procedural materials - the grill that covers it is also a procedural material.
Did some initial modeling of the side door but I might end up keeping it closed on the finished model - plus I might have to remodel it, as it possibly intersects the side geometry when it opens and closes.
From memory the main things left to do on the model is some minor modelling, a whole lot of beveling, and creating all the textures.
Slowly plodding along with this. I've pretty much finished all of the exterior modelling, just need to apply the mirror modifier and reconnect the asymmetrical bits... well until I find something else to model or remodel .
Ended up modelling a simple interior, but probably won't do much more with it for a while (got a few other models I want to work on). It's rather cramped height wise and with the models current scale if you were to stand up straight your head would be sticking out of the hull.
I've also started on drawing the decals, I do need to find a better font for the smaller text but considering you can't really see it on the show it probably doesn't matter too much.
Finished adding all the details (hopefully haven't missed any), and applied the mirror modifier on the main hull and warp engine. Realised that I hadn't removed the step on the other nacelle, so did that. Really the only thing left on the exterior is to fiddle a bit with the materials (the full size model appears to have a rather glossy/reflective paint on it) and rig the landing gear.
So today got an urge to render the shuttle in the shuttle bay (the shuttle bay is still very much wip, I haven't touched it in ages), which in the end led me to doing a bit more material work.
I've also been experiencing a rendering bug with the alphas of my decals, for whatever reason it ignores some of the underlying meshe's material settings, messing up the specular highlights. Here's a quick example from the viewport renderer showing what going on.
That's a good looking bay. That sucks about the decals. What if you were to move the shader layers on top of them in the hierarchy, would that maybe fix the issue?
That's a good looking bay. That sucks about the decals. What if you were to move the shader layers on top of them in the hierarchy, would that maybe fix the issue?
I have a general idea what causes the render errors. The error seems to be caused by the interaction between the hull material (which uses the clearcoat option on the principle shader); and the transparency on the decal material (the decals are separate meshes). It seems to only show up if a light with a large radius, or large emission mesh, is used to light the model.
The clearcoat itself isn't the cause, as I can replicate the render error by mixing two shaders together using a frensel Layer Weight node, to kinda replicate the clearcoat material.
I have two solutions to fix the render error.
The first solution is to have the decals texture part of the hull material; instead of a separate mesh. This works, but means I lose the flexibility of having the decals as separate meshes, plus I would have to create a new UV map for the decals.
The other much simpler solution (which I'm currently using), is to remove the clearcoat and just reduce the roughness on the principle shader, so I have a nice glossy material.
Ended up adding some headlights after reading that the model was supposed to have them. While setting up the lights found out that blender doesn't always like transmitting light through transparent/translucent materials, so had to fiddle with the material to make the light rays ignore it.
Done a bit more work on the interior. Main things being the chairs, sphere monitors, and blocking out the details on the control panels. I need to collect some more references before I can continue, and also need to decide how close I'll keep the details to their originals (I'm looking at you reel-to-reel player).
Also got inspired by backstept's aztec material, and ended up fiddling around in the shader editor to make my own version using fresnel layer weights. Still needs a bit more fiddling with, but I'm reasonable happy with it. At some point I'll need to apply it to a full model to see if it'll work outside of the test scene.
Here's the node setup.
Almost forgot to mention that the colours are based on PixelMagic's colour chart in backstept's refit connie thread.
While setting up the lights found out that blender doesn't always like transmitting light through transparent/translucent materials, so had to fiddle with the material to make the light rays ignore it.
I'm using cylces, but Eevee likely doesn't have the same issue, since it uses a more simplified lighting model and transparent materials are treated differently to Cycles.
The following image better illustrates the issue. The left side is using a Glass BSDF mixed with a transparent BSDF, and is using the Is Shadow Ray from a Light Path node as the mix factor. While the right is just using a Glass BSDF shader.
I've found that the effect is more noticeable on small lights, and that Indirect Light Clamping, in the Render Properties, has an impact on it. Turning it off or upping its amount will increase the amount of light transmitted, but will create more fireflies and noise. It also requires a much longer render time to get a noise free image - I upped the sample count to over 8000 (been using 512 for almost all my images) and still had a horrible noisy image.
I've also found that in general it's better to use light objects instead of emission objects, as the light objects produce a lot less noise.
Posts
So far I have finished the warp nacelles, and I'm currently slowly integrating the main hull pieces together (modeled it in 3 pieces).
One thing I'm currently undecided on is to have the warp nacelles parallel to each other, or have them toed in so they are parallel to the hull?
I've finished most of the work on the rear landing gear, but I may redo some of it to add some details I missed (mainly the teardrop shaped indent).
I also changed how I was doing the curve on the warp nacelles corrugated rear caps. I was using a boolean object to cut out the curve, but I realised that I could get a similar result using vertex groups and a shrinkwrap modifier. The great thing about this method is it doesn't create unneeded vertices, which should mean that there will be minimal or no cleanup when I apply the modifier stack.
Next part I'll be working on is integrating the middle hull part into the rest of the main hull. I should have done this earlier as it's going to be a real pain to do cleanly , since I've already freezed the subd of the main hull part.
Great attention to details.
That's really good!
Various Work: U.S.S. Constellation - Matt Jefferies Concept Shuttle
Modeled the rear details and added the engines. Not sure if I'll keep the glowing look, or go with something that better matches the full size prop - it appears to have rectangular tubes in the show, but the restored prop has the glowing panels.
Close up of the details behind the vent grill. It's mainly all geometry, but the holes and lights are procedural materials - the grill that covers it is also a procedural material.
Did some initial modeling of the side door but I might end up keeping it closed on the finished model - plus I might have to remodel it, as it possibly intersects the side geometry when it opens and closes.
From memory the main things left to do on the model is some minor modelling, a whole lot of beveling, and creating all the textures.
Ended up modelling a simple interior, but probably won't do much more with it for a while (got a few other models I want to work on). It's rather cramped height wise and with the models current scale if you were to stand up straight your head would be sticking out of the hull.
I've also started on drawing the decals, I do need to find a better font for the smaller text but considering you can't really see it on the show it probably doesn't matter too much.
While searching for info on the shuttle's chairs, I found a blog post with some information on the shuttle's colours, so now it should be closer in colour to the actual model - one thing I hadn't noticed was that the underside was a darker grey than the top (thought it was just the lighting).
https://collectormodel.com/round2-models/5013-star-trek-modeling-a-brief-history-of-the-shuttlecraft-galileo-pt-4/
I've also been experiencing a rendering bug with the alphas of my decals, for whatever reason it ignores some of the underlying meshe's material settings, messing up the specular highlights. Here's a quick example from the viewport renderer showing what going on.
I have a general idea what causes the render errors. The error seems to be caused by the interaction between the hull material (which uses the clearcoat option on the principle shader); and the transparency on the decal material (the decals are separate meshes). It seems to only show up if a light with a large radius, or large emission mesh, is used to light the model.
The clearcoat itself isn't the cause, as I can replicate the render error by mixing two shaders together using a frensel Layer Weight node, to kinda replicate the clearcoat material.
I have two solutions to fix the render error.
The first solution is to have the decals texture part of the hull material; instead of a separate mesh. This works, but means I lose the flexibility of having the decals as separate meshes, plus I would have to create a new UV map for the decals.
The other much simpler solution (which I'm currently using), is to remove the clearcoat and just reduce the roughness on the principle shader, so I have a nice glossy material.
The ground texture in the second picture is based on a tutorial by Ryan King Art
https://youtube.com/watch?v=PIZ_wi3yFUM&list=PLsGl9GczcgBs6TtApKKK-L_0Nm6fovNPk&index=74
With this the exterior should be complete (well until I find something else to model on it ).
Done a bit more work on the interior. Main things being the chairs, sphere monitors, and blocking out the details on the control panels. I need to collect some more references before I can continue, and also need to decide how close I'll keep the details to their originals (I'm looking at you reel-to-reel player).
Also got inspired by backstept's aztec material, and ended up fiddling around in the shader editor to make my own version using fresnel layer weights. Still needs a bit more fiddling with, but I'm reasonable happy with it. At some point I'll need to apply it to a full model to see if it'll work outside of the test scene.
Here's the node setup.
Almost forgot to mention that the colours are based on PixelMagic's colour chart in backstept's refit connie thread.
Various Work: U.S.S. Constellation - Matt Jefferies Concept Shuttle
The following image better illustrates the issue. The left side is using a Glass BSDF mixed with a transparent BSDF, and is using the Is Shadow Ray from a Light Path node as the mix factor. While the right is just using a Glass BSDF shader.
I've found that the effect is more noticeable on small lights, and that Indirect Light Clamping, in the Render Properties, has an impact on it. Turning it off or upping its amount will increase the amount of light transmitted, but will create more fireflies and noise. It also requires a much longer render time to get a noise free image - I upped the sample count to over 8000 (been using 512 for almost all my images) and still had a horrible noisy image.
I've also found that in general it's better to use light objects instead of emission objects, as the light objects produce a lot less noise.
According to this article they were planned but never actually installed; same as the warp engine lights.