DAMN, dude! At this point I'm convinced that when we build a starship for real, Donny's going to be the one to do it. 
Donny's TOS Enterprise Interiors
- Thread starter Donny
- Start date
That was me. You're not changing my mind. Amazing.
Wicked
Hi Donny
I would gladly pay to sit in your modelling and texturing classes if you did them. Awesome work creating reality.
I would gladly pay to sit in your modelling and texturing classes if you did them. Awesome work creating reality.
Someone on this thread recently told me I was insane. Well, I'm starting to believe it. I just spent my Sunday researching, modeling, and texturing this painstakingly accurate celestial globe, complete with accurate constellation positioning and degree tick marks. Why do I do this??
Because you can't abide by anything less for the sake of doing a satisfactory job. I understand that feeling well.
So, tonight, along with the usual showcase screenshots of whatever prop I've finished, I'm gonna give a rundown of my process.
Kirk's hexagonal mirror with sculpted flowery inlays. So unfitting for his person, just like almost everything in his quarters. But a canon TOS prop nonetheless and I must, of course, create a model of it no matter how much I want to come up with a more fitting mirror for James T. (To be fair, Spock, Janice Rand, and Marla McGivers all posess this same mirror in their quarters as well, so we can infer that it's standard Starfleet issue in the 2260s)
I start by collecting reference, which I usually do per set I tackle. I scour the internet and Trekcore for all screenshots as I can of the set, taking specific note of any shots that provide detail, no matter how insignificant, and save them to my library on my hard drive for easy access. These two, in particular, were very helpful for creating this model of the mirror, from "The Enemy Within" and "Wink of an Eye", respectively,:
The "Wink of an Eye" shot even revealed something to me I'd never notice before: the flowerly inlay isn't just carved into a recessed area in between the two diagonal bars, but is actually openwork, in that there are holes in the negative space passing right through. This is evident in the way it is shadowing on the wall behind it.
I then take all good reference shots of a given prop, and compile them into a compilation image, adjusting the values of the images in cases to make details appear more clearly. This also saves me from having to scroll through 10-20 images of a given object, and allows me to see multiple angles of the object at any time:
I set out with a plan in mind on how I'm going to tackle the prop. In the case of the mirror, I can use a traditional polygonal modeling software like 3dsMax to do most of the work: the hexagonal frame around the mirror, the mirror itself, and the diagonal bars that frame the sculpted inlay. But for the flowerly sculpts and the ornate conic shape up top, I know I'll have to sculpt those details in a sculpting program like Zbrush (which, I thankfully have learned how to use in the last two weeks) But that sculpt will result in a 3d mesh consisting of millions of polygon, which 3ds Max and the Unreal game engine (which is my final render) cannot handle very well without severe performance issues in both those programs. So, instead, the final low-poly in-game model's details with be a flat texture using detailed texture information from that high-poly sculpt mesh to simulate the higher-poly 3D geometry in a process called baking. Make sense? No worries. I'll explain a little more later.
I do the easy polygonal work in 3ds max first, as I'm most familiar with the program and it gives me a guide for when it's time to do the sculpting:
I then take that simple polygonal mesh into Zbrush and go to work sculpting, which took 2 days for me to do, because I'm relatively new to the program and some trial and error was involved. I tackled each individual flower/leaf/petal/thingagbob at a time before moving on to the next one, until they're all relatively close to the most detialed reference image I have available, which happens to be the shot from "The Enemy Within" with Evil Shatner going HAM on some overacting. Here's a shot of the final sculpted detail in Zbrush:
I then move on to a texturing program called Substance Painter, which is my most favorite program in the world, to do what also happens to me my most favorite part of the process: texturing. Where I give the model it's color, look, texture, distress, edge wear, etc. Substance Painter uses powerful tool to quickly generate masks for doing things like edge wear, which has taken my models to a level of realism they were lacking before I discovered the program when I was doing the art test for Cryptic Studios.
I start by importing the low poly mesh and do that process I mentioned earlier called "baking", which takes the high-poly model information and projects it onto the textures of the low poly model, giving the illusion of more polygonal detail where there is none. Afterwards, I paint the model itself, specifically painting it's diffuse/albedo values (straight color information), roughness values (how rough/glossy the object is), it's metal values (whether or not the object is metallic), and other fun stuff like height and normal information (like the pattern surrounding the hexagonal mirror), allowing the model to react to light in a way that again gives the illusion of more detail than it actually has (I should note that since all my final rendering is done in a real-time game engine, all models have to be signficantly lower in polygonal detail than they would in say, a CG movie. So in game-engine type rendering, it's crucial to "fake" as much detail as possible without breaking the illusion.) Here's the model in Painter:
I then export all the different textures I made in Substance Painter through baking and painting. Here's a breakdown of all the different images which are then compiled in the game engine into what is referred to as a "material", which then is applied to the model and is computed to give the model it's final look in the game engine (note, the image below has only the information for the sculpted flowery detail. I have not included the texture information for the rest of the model, as it uses a completely separate set of textures).
Here's what they all are:
1) Albedo: As stated before, this is the pure color information
2) Normals: This image is used by the render to determine how light reacts when it's rays hit the object. This map is most crucial in faking geometry that isn't actually on the model itself.
3) Metalness: How metallic the object is in places compared to others, zero (black) being not metallic at all and one (white) being metallic.
4) Roughness: How rough the object is in places compared to others, zero (black) being highly, mirror-like glossy and one (white) being extremely rough/matte.
5) Ambient Occlusion: this is self-shadowing information that aids the game engine add shadowing where geometry does not actually exist, and also aids in shadowing crevices and cavities.
6) Height: This is used by the game engine specifically for this kind of model where a flat surface will be faking sculpted detail. This tells how raised parts of the model should be in comparison to others, zero (black) being flat and one (white) being raised.
7) Opacity: Since I discovered that there is actual negative space in between the sculpted details, this image will tell the game engine which areas are transparent (black) verus opaque (white).
These images are then input into the "material" file which then does all the computations based on the texture information it receives and tells the game engine how to render the object the material is applied to. I then import the low poly model (which ended up only being a little over 1000 polygons versus the 3 milliion polygon model in Zbrush) and assign the material to it, put the model in the game world, and voila! With positioning of some lights, we get to see our final product in game.
Remember you can always click on any of the images I post here to see them in higher resolution
(Also, I've noticed that I could probably up the saturation in the color of the mirror, as it's appear far too silvery/gray in the following shots)
Kirk's hexagonal mirror with sculpted flowery inlays. So unfitting for his person, just like almost everything in his quarters. But a canon TOS prop nonetheless and I must, of course, create a model of it no matter how much I want to come up with a more fitting mirror for James T. (To be fair, Spock, Janice Rand, and Marla McGivers all posess this same mirror in their quarters as well, so we can infer that it's standard Starfleet issue in the 2260s)
I start by collecting reference, which I usually do per set I tackle. I scour the internet and Trekcore for all screenshots as I can of the set, taking specific note of any shots that provide detail, no matter how insignificant, and save them to my library on my hard drive for easy access. These two, in particular, were very helpful for creating this model of the mirror, from "The Enemy Within" and "Wink of an Eye", respectively,:
The "Wink of an Eye" shot even revealed something to me I'd never notice before: the flowerly inlay isn't just carved into a recessed area in between the two diagonal bars, but is actually openwork, in that there are holes in the negative space passing right through. This is evident in the way it is shadowing on the wall behind it.
I then take all good reference shots of a given prop, and compile them into a compilation image, adjusting the values of the images in cases to make details appear more clearly. This also saves me from having to scroll through 10-20 images of a given object, and allows me to see multiple angles of the object at any time:
I set out with a plan in mind on how I'm going to tackle the prop. In the case of the mirror, I can use a traditional polygonal modeling software like 3dsMax to do most of the work: the hexagonal frame around the mirror, the mirror itself, and the diagonal bars that frame the sculpted inlay. But for the flowerly sculpts and the ornate conic shape up top, I know I'll have to sculpt those details in a sculpting program like Zbrush (which, I thankfully have learned how to use in the last two weeks
) But that sculpt will result in a 3d mesh consisting of millions of polygon, which 3ds Max and the Unreal game engine (which is my final render) cannot handle very well without severe performance issues in both those programs. So, instead, the final low-poly in-game model's details with be a flat texture using detailed texture information from that high-poly sculpt mesh to simulate the higher-poly 3D geometry in a process called baking. Make sense? No worries. I'll explain a little more later.I do the easy polygonal work in 3ds max first, as I'm most familiar with the program and it gives me a guide for when it's time to do the sculpting:
I then take that simple polygonal mesh into Zbrush and go to work sculpting, which took 2 days for me to do, because I'm relatively new to the program and some trial and error was involved. I tackled each individual flower/leaf/petal/thingagbob at a time before moving on to the next one, until they're all relatively close to the most detialed reference image I have available, which happens to be the shot from "The Enemy Within" with Evil Shatner going HAM on some overacting. Here's a shot of the final sculpted detail in Zbrush:
I then move on to a texturing program called Substance Painter, which is my most favorite program in the world, to do what also happens to me my most favorite part of the process: texturing. Where I give the model it's color, look, texture, distress, edge wear, etc. Substance Painter uses powerful tool to quickly generate masks for doing things like edge wear, which has taken my models to a level of realism they were lacking before I discovered the program when I was doing the art test for Cryptic Studios.
I start by importing the low poly mesh and do that process I mentioned earlier called "baking", which takes the high-poly model information and projects it onto the textures of the low poly model, giving the illusion of more polygonal detail where there is none. Afterwards, I paint the model itself, specifically painting it's diffuse/albedo values (straight color information), roughness values (how rough/glossy the object is), it's metal values (whether or not the object is metallic), and other fun stuff like height and normal information (like the pattern surrounding the hexagonal mirror), allowing the model to react to light in a way that again gives the illusion of more detail than it actually has (I should note that since all my final rendering is done in a real-time game engine, all models have to be signficantly lower in polygonal detail than they would in say, a CG movie. So in game-engine type rendering, it's crucial to "fake" as much detail as possible without breaking the illusion.) Here's the model in Painter:
I then export all the different textures I made in Substance Painter through baking and painting. Here's a breakdown of all the different images which are then compiled in the game engine into what is referred to as a "material", which then is applied to the model and is computed to give the model it's final look in the game engine (note, the image below has only the information for the sculpted flowery detail. I have not included the texture information for the rest of the model, as it uses a completely separate set of textures).
Here's what they all are:
1) Albedo: As stated before, this is the pure color information
2) Normals: This image is used by the render to determine how light reacts when it's rays hit the object. This map is most crucial in faking geometry that isn't actually on the model itself.
3) Metalness: How metallic the object is in places compared to others, zero (black) being not metallic at all and one (white) being metallic.
4) Roughness: How rough the object is in places compared to others, zero (black) being highly, mirror-like glossy and one (white) being extremely rough/matte.
5) Ambient Occlusion: this is self-shadowing information that aids the game engine add shadowing where geometry does not actually exist, and also aids in shadowing crevices and cavities.
6) Height: This is used by the game engine specifically for this kind of model where a flat surface will be faking sculpted detail. This tells how raised parts of the model should be in comparison to others, zero (black) being flat and one (white) being raised.
7) Opacity: Since I discovered that there is actual negative space in between the sculpted details, this image will tell the game engine which areas are transparent (black) verus opaque (white).
These images are then input into the "material" file which then does all the computations based on the texture information it receives and tells the game engine how to render the object the material is applied to. I then import the low poly model (which ended up only being a little over 1000 polygons versus the 3 milliion polygon model in Zbrush) and assign the material to it, put the model in the game world, and voila! With positioning of some lights, we get to see our final product in game.
Remember you can always click on any of the images I post here to see them in higher resolution
(Also, I've noticed that I could probably up the saturation in the color of the mirror, as it's appear far too silvery/gray in the following shots)
Last edited:
incredible Donny. I couldn't help noticing the TOS walls, do you use Substance Painter for those also? The texturing on them looks slightly skin like?So, tonight, along with the usual showcase screenshots of whatever prop I've finished, I'm gonna give a rundown of my process.
Kirk's hexagonal mirror with sculpted flowery inlays. So unfitting for his person, just like almost everything in his quarters. But a canon TOS prop nonetheless and I must, of course, create a model of it no matter how much I want to come up with a more fitting mirror for James T. (To be fair, Spock, Janice Rand, and Marla McGivers all posess this same mirror in their quarters as well, so we can infer that it's standard Starfleet issue in the 2260s)
I start by collecting reference, which I usually do per set I tackle. I scour the internet and Trekcore for all screenshots as I can of the set, taking specific note of any shots that provide detail, no matter how insignificant, and save them to my library on my hard drive for easy access. These two, in particular, were very helpful for creating this model of the mirror, from "The Enemy Within" and "Wink of an Eye", respectively,:
The "Wink of an Eye" shot even revealed something to me I'd never notice before: the flowerly inlay isn't just carved into a recessed area in between the two diagonal bars, but is actually openwork, in that there are holes in the negative space passing right through. This is evident in the way it is shadowing on the wall behind it.
I then take all good reference shots of a given prop, and compile them into a compilation image, adjusting the values of the images in cases to make details appear more clearly. This also saves me from having to scroll through 10-20 images of a given object, and allows me to see multiple angles of the object at any time:
I set out with a plan in mind on how I'm going to tackle the prop. In the case of the mirror, I can use a traditional polygonal modeling software like 3dsMax to do most of the work: the hexagonal frame around the mirror, the mirror itself, and the diagonal bars that frame the sculpted inlay. But for the flowerly sculpts and the ornate conic shape up top, I know I'll have to sculpt those details in a sculpting program like Zbrush (which, I thankfully have learned how to use in the last two weeks
I do the easy polygonal work in 3ds max first, as I'm most familiar with the program and it gives me a guide for when it's time to do the sculpting:
I then take that simple polygonal mesh into Zbrush and go to work sculpting, which took 2 days for me to do, because I'm relatively new to the program and some trial and error was involved. I tackled each individual flower/leaf/petal/thingagbob at a time before moving on to the next one, until they're all relatively close to the most detialed reference image I have available, which happens to be the shot from "The Enemy Within" with Evil Shatner going HAM on some overacting. Here's a shot of the final sculpted detail in Zbrush:
I then move on to a texturing program called Substance Painter, which is my most favorite program in the world, to do what also happens to me my most favorite part of the process: texturing. Where I give the model it's color, look, texture, distress, edge wear, etc. Substance Painter uses powerful tool to quickly generate masks for doing things like edge wear, which has taken my models to a level of realism they were lacking before I discovered the program when I was doing the art test for Cryptic Studios.
I start by importing the low poly mesh and do that process I mentioned earlier called "baking", which takes the high-poly model information and projects it onto the textures of the low poly model, giving the illusion of more polygonal detail where there is none. Afterwards, I paint the model itself, specifically painting it's diffuse/albedo values (straight color information), roughness values (how rough/glossy the object is), it's metal values (whether or not the object is metallic), and other fun stuff like height and normal information (like the pattern surrounding the hexagonal mirror), allowing the model to react to light in a way that again gives the illusion of more detail than it actually has (I should note that since all my final rendering is done in a real-time game engine, all models have to be signficantly lower in polygonal detail than they would in say, a CG movie. So in game-engine type rendering, it's crucial to "fake" as much detail as possible without breaking the illusion.) Here's the model in Painter:
I then export all the different textures I made in Substance Painter through baking and painting. Here's a breakdown of all the different images which are then compiled in the game engine into what is referred to as a "material", which then is applied to the model and is computed to give the model it's final look in the game engine (note, the image below has only the information for the sculpted flowery detail. I have not included the texture information for the rest of the model, as it uses a completely separate set of textures).
Here's what they all are:
1) Albedo: As stated before, this is the pure color information
2) Normals: This image is used by the render to determine how light reacts when it's rays hit the object. This map is most crucial in faking geometry that isn't actually on the model itself.
3) Metalness: How metallic the object is in places compared to others, zero (black) being not metallic at all and one (white) being metallic.
4) Roughness: How rough the object is in places compared to others, zero (black) being highly, mirror-like glossy and one (white) being extremely rough/matte.
5) Ambient Occlusion: this is self-shadowing information that aids the game engine add shadowing where geometry does not actually exist, and also aids in shadowing crevices and cavities.
6) Height: This is used by the game engine specifically for this kind of model where a flat surface will be faking sculpted detail. This tells how raised parts of the model should be in comparison to others, zero (black) being flat and one (white) being raised.
7) Opacity: Since I discovered that there is actual negative space in between the sculpted details, this image will tell the game engine which areas are transparent (black) verus opaque (white).
These images are then input into the "material" file which then does all the computations based on the texture information it receives and tells the game engine how to render the object the material is applied to. I then import the low poly model (which ended up only being a little over 1000 polygons versus the 3 milliion polygon model in Zbrush) and assign the material to it, put the model in the game world, and voila! With positioning of some lights, we get to see our final product in game.
Remember you can always click on any of the images I post here to see them in higher resolution
(Also, I've noticed that I could probably up the saturation in the color of the mirror, as it's appear far too silvery/gray in the following shots)
See? INSANE!
And the winner of post of the month is …So, tonight, along with the usual showcase screenshots of whatever prop I've finished, I'm gonna give a rundown of my process.
Kirk's hexagonal mirror with sculpted flowery inlays. So unfitting for his person, just like almost everything in his quarters. But a canon TOS prop nonetheless and I must, of course, create a model of it no matter how much I want to come up with a more fitting mirror for James T. (To be fair, Spock, Janice Rand, and Marla McGivers all posess this same mirror in their quarters as well, so we can infer that it's standard Starfleet issue in the 2260s)
I start by collecting reference, which I usually do per set I tackle. I scour the internet and Trekcore for all screenshots as I can of the set, taking specific note of any shots that provide detail, no matter how insignificant, and save them to my library on my hard drive for easy access. These two, in particular, were very helpful for creating this model of the mirror, from "The Enemy Within" and "Wink of an Eye", respectively,:
The "Wink of an Eye" shot even revealed something to me I'd never notice before: the flowerly inlay isn't just carved into a recessed area in between the two diagonal bars, but is actually openwork, in that there are holes in the negative space passing right through. This is evident in the way it is shadowing on the wall behind it.
I then take all good reference shots of a given prop, and compile them into a compilation image, adjusting the values of the images in cases to make details appear more clearly. This also saves me from having to scroll through 10-20 images of a given object, and allows me to see multiple angles of the object at any time:
I set out with a plan in mind on how I'm going to tackle the prop. In the case of the mirror, I can use a traditional polygonal modeling software like 3dsMax to do most of the work: the hexagonal frame around the mirror, the mirror itself, and the diagonal bars that frame the sculpted inlay. But for the flowerly sculpts and the ornate conic shape up top, I know I'll have to sculpt those details in a sculpting program like Zbrush (which, I thankfully have learned how to use in the last two weeks
I do the easy polygonal work in 3ds max first, as I'm most familiar with the program and it gives me a guide for when it's time to do the sculpting:
I then take that simple polygonal mesh into Zbrush and go to work sculpting, which took 2 days for me to do, because I'm relatively new to the program and some trial and error was involved. I tackled each individual flower/leaf/petal/thingagbob at a time before moving on to the next one, until they're all relatively close to the most detialed reference image I have available, which happens to be the shot from "The Enemy Within" with Evil Shatner going HAM on some overacting. Here's a shot of the final sculpted detail in Zbrush:
I then move on to a texturing program called Substance Painter, which is my most favorite program in the world, to do what also happens to me my most favorite part of the process: texturing. Where I give the model it's color, look, texture, distress, edge wear, etc. Substance Painter uses powerful tool to quickly generate masks for doing things like edge wear, which has taken my models to a level of realism they were lacking before I discovered the program when I was doing the art test for Cryptic Studios.
I start by importing the low poly mesh and do that process I mentioned earlier called "baking", which takes the high-poly model information and projects it onto the textures of the low poly model, giving the illusion of more polygonal detail where there is none. Afterwards, I paint the model itself, specifically painting it's diffuse/albedo values (straight color information), roughness values (how rough/glossy the object is), it's metal values (whether or not the object is metallic), and other fun stuff like height and normal information (like the pattern surrounding the hexagonal mirror), allowing the model to react to light in a way that again gives the illusion of more detail than it actually has (I should note that since all my final rendering is done in a real-time game engine, all models have to be signficantly lower in polygonal detail than they would in say, a CG movie. So in game-engine type rendering, it's crucial to "fake" as much detail as possible without breaking the illusion.) Here's the model in Painter:
I then export all the different textures I made in Substance Painter through baking and painting. Here's a breakdown of all the different images which are then compiled in the game engine into what is referred to as a "material", which then is applied to the model and is computed to give the model it's final look in the game engine (note, the image below has only the information for the sculpted flowery detail. I have not included the texture information for the rest of the model, as it uses a completely separate set of textures).
Here's what they all are:
1) Albedo: As stated before, this is the pure color information
2) Normals: This image is used by the render to determine how light reacts when it's rays hit the object. This map is most crucial in faking geometry that isn't actually on the model itself.
3) Metalness: How metallic the object is in places compared to others, zero (black) being not metallic at all and one (white) being metallic.
4) Roughness: How rough the object is in places compared to others, zero (black) being highly, mirror-like glossy and one (white) being extremely rough/matte.
5) Ambient Occlusion: this is self-shadowing information that aids the game engine add shadowing where geometry does not actually exist, and also aids in shadowing crevices and cavities.
6) Height: This is used by the game engine specifically for this kind of model where a flat surface will be faking sculpted detail. This tells how raised parts of the model should be in comparison to others, zero (black) being flat and one (white) being raised.
7) Opacity: Since I discovered that there is actual negative space in between the sculpted details, this image will tell the game engine which areas are transparent (black) verus opaque (white).
These images are then input into the "material" file which then does all the computations based on the texture information it receives and tells the game engine how to render the object the material is applied to. I then import the low poly model (which ended up only being a little over 1000 polygons versus the 3 milliion polygon model in Zbrush) and assign the material to it, put the model in the game world, and voila! With positioning of some lights, we get to see our final product in game.
Remember you can always click on any of the images I post here to see them in higher resolution
(Also, I've noticed that I could probably up the saturation in the color of the mirror, as it's appear far too silvery/gray in the following shots)
Thanks for taking the time and documenting your work. Really impressive and helpful in understanding what it is you are doing.
Oh, woah! I didn't even noticed the black tape! I assume they did this to hide the reflection of bright stage lighting. Looks like they added a slight line of gold trim to the bottom edge of the tape after Season 1 to make it look a little more like a purposeful decorative touch. I may just have to add this!Great walkthrough of your workflow
BTW, insanity would have been including the black tape visible on the finial sides of the mirror glass
Yes! The final sculpted details were projected onto the flat "gridded" plane in between the diagonal bars at the bottom and given the illusion of depth through texture and material work, so no new geometry was required for the final low poly mesh. I removed the grid lines on that plane, however, for the final low poly mesh, because they were just used during the baking process to help average out the bake (if I hadn't have done this, some warping would have occurred). So, actually, the final in-game mesh was actually lower in polygon count than what you see in the 3dsMax shot I provided.
Yes, it's a tiling texture that was done in Painter and given an extra detail "bump map" to simulate paint bumps in Unreal engine. I could probably tweak it a bit, as it's not meant to look like skin, but is an attempt to replicate the fine details of a painted wall.
I'd say it was probably to avoid reflections from the studio lights or the limit of the set wall.
BTW, insanity would have been including the black tape visible on the finial sides of the mirror glass
Wow, I didn't notice that either, but now that it's been pointed out it's blindingly obvious.
OMG! I just noticed that the mirror is hung the other way up in Journey To Babel!
I think the flowers on top and the tassel on the bottom is the way it was supposed to be hung. If the tassel was real it would hang down under the force of gravity, thus designed to be on the bottom.
I think you're right. The edges of the floral triangle seem to represent a hanging cord, splaying out from the circular "nail" at the top.
But since that makes so much sense, I'm not surprised that TOS inverted it. That's futuristic!
