Modelling and Rendering the TMP Enterprise
- Thread starter Science Officer
- Start date
Hi Folks,
I hven't had much time to work on the project this week. I managed some work on the rim, but it is still incomplete.
My attention has turned back to the shader and I tried to remove remaining differences between the renders and Cloudster references. I did this by modifying the Phong shader diffuse falloff. The results against references looked good, but just looked wrong for everything else so I reverted back to the look in my previous post.
What I have done is hacked a few colours into the specular highlights - no real attempt to match the model. I changed the hue on 3 of the 10 shades to red, orange and blue. To get a better feel of how they work, I've rendered a 10 second movie moving the camera around the saucer. In the first 5 seconds you can see some colour in the specular highlights (on the left hand side of the frames) - it behaves fairly well. Most importantly it's a fairly subtle effect, which could be a tad stronger.
I've popped this on ImageShack for now, as I messed up trying to set up a YouTube account (having confused it with a Google account, which I'd deleted)
. The file has consequently been downsized from 640x480 pixels.
http://img687.imageshack.us/img687/5530/ussenterprisetexture.mp4
Cheers,
S.O.
I hven't had much time to work on the project this week. I managed some work on the rim, but it is still incomplete.
My attention has turned back to the shader and I tried to remove remaining differences between the renders and Cloudster references. I did this by modifying the Phong shader diffuse falloff. The results against references looked good, but just looked wrong for everything else so I reverted back to the look in my previous post.
What I have done is hacked a few colours into the specular highlights - no real attempt to match the model. I changed the hue on 3 of the 10 shades to red, orange and blue. To get a better feel of how they work, I've rendered a 10 second movie moving the camera around the saucer. In the first 5 seconds you can see some colour in the specular highlights (on the left hand side of the frames) - it behaves fairly well. Most importantly it's a fairly subtle effect, which could be a tad stronger.
I've popped this on ImageShack for now, as I messed up trying to set up a YouTube account (having confused it with a Google account, which I'd deleted)
. The file has consequently been downsized from 640x480 pixels.http://img687.imageshack.us/img687/5530/ussenterprisetexture.mp4
Cheers,
S.O.
Beautiful work, S.O.! 
I love the --ever-so-slight-- blue tint you gave to the lighting and the pearlescent reflections look dead-on to me. Very much looking forward to future postings.
I love the --ever-so-slight-- blue tint you gave to the lighting and the pearlescent reflections look dead-on to me. Very much looking forward to future postings.Thanks.
137th Gebirg - in the animation I used a pure white light, so any hints of blue are due to the shader.
On the subject of blue, I may have made progress following yesterday's post. This concerns the running light movie snapshot. At the time I did those renders (posts 69 and 82), the spotlight required a different hue to the background light to get a purple specular highlight. That is no longer the case becuse the default specular highlight hue is red. Also if you get the light strength right, you also get the blue halo.
Of course there has to be a problem or two!
Because I worked off B&W images, I figured I needed at least 11 shades to reproduce the shade variations in the Aztec pattern. However this did not properly account for colours as it is very likely that two colour highlights at different intensities will produce the same B&W shade. I'm already stuck with the fact that there is insufficient reference material out there to accurately assign colours to the polygons. So I need to work out how many colours are required and assign them to the existing shades the best I can. If neccessary I could split an existing selection and give each a colour.
Although I'm happy with the results in the animation, the saucer still looks a litte dead somehow. The one thing that could help is blurred reflection, but the render hit would make it impractical. It might be the case that as I refine the specular hightlights I'll get a little more play at the edges of the specular hightlights.
Cheers,
S.O.
137th Gebirg - in the animation I used a pure white light, so any hints of blue are due to the shader.
On the subject of blue, I may have made progress following yesterday's post. This concerns the running light movie snapshot. At the time I did those renders (posts 69 and 82), the spotlight required a different hue to the background light to get a purple specular highlight. That is no longer the case becuse the default specular highlight hue is red. Also if you get the light strength right, you also get the blue halo.
Of course there has to be a problem or two!
Because I worked off B&W images, I figured I needed at least 11 shades to reproduce the shade variations in the Aztec pattern. However this did not properly account for colours as it is very likely that two colour highlights at different intensities will produce the same B&W shade. I'm already stuck with the fact that there is insufficient reference material out there to accurately assign colours to the polygons. So I need to work out how many colours are required and assign them to the existing shades the best I can. If neccessary I could split an existing selection and give each a colour.
Although I'm happy with the results in the animation, the saucer still looks a litte dead somehow. The one thing that could help is blurred reflection, but the render hit would make it impractical. It might be the case that as I refine the specular hightlights I'll get a little more play at the edges of the specular hightlights.
Cheers,
S.O.
Very nice S.O. and I think the colors should be a tad more also, but you can get that last refinement only good if you put it to test in simulation of a real shot they did in the TMP drydock sequence when for instance the ship passes the camera. Think that will be the only good reference out there, it really is a shame they did not make good color photos instead of the B&W. My $0.02 
Keep up the good work.
Keep up the good work.
Very nice S.O. and I think the colors should be a tad more also, but you can get that last refinement only good if you put it to test in simulation of a real shot they did in the TMP drydock sequence when for instance the ship passes the camera. Think that will be the only good reference out there, it really is a shame they did not make good color photos instead of the B&W. My $0.02
Keep up the good work.
It looks like a real shot would be the way to go. Although I'd have to reverse engineer the lights (number, colour, position, point source/area etc). But it would yield some useful information.
There are a few HD shots in the drydock sequence that show the colours - especially when the travel pod goes down the side of the secondary hull just prior to docking. Alas not much good too me unless I quicking assign the new stuff to the some of the polys of the DB v1 secondary hull.
I'm reworking highlights with more saturation, but only one hue. The animation is rendering as I write. Might be a little too much this time, but is the closest to what I'd call a pearlescent finish. I'll post an image, when the animation is complete.
Here are some comparison images - left hand side are the ones in the animation from yesterday. The right hand ones have more saturation, but no hue variation. If you look at some of the HD movie shots, you will see this pink specular highlight.
S.O. - thanks for the clarification.
As for the saucer looking a little "dead", that might be due to the fact that the rest of the ship isn't there yet. If you had the other parts there with the big blue (or amber) nav deflector with the warp grilles and spot lights doing their thing, adding to the ambient lighting, I think you'll find the saucer to come alive. The RCS thrusters need their lights, as well as the red/green navigationals and strobes, but other than that, as far as I can tell, it looks completely dead-on.
As for the saucer looking a little "dead", that might be due to the fact that the rest of the ship isn't there yet. If you had the other parts there with the big blue (or amber) nav deflector with the warp grilles and spot lights doing their thing, adding to the ambient lighting, I think you'll find the saucer to come alive. The RCS thrusters need their lights, as well as the red/green navigationals and strobes, but other than that, as far as I can tell, it looks completely dead-on.
I wouldn't actually trust the BD screencaps as far as color goes. They jacked up the color balance to an insane level. I'd look around for screencaps from the HD transfer for SkyHD, they're far more accurate to the original print.
Hi Folks,
USS Mariner - I don't think the HD references are from the blu-ray. I'm using a site wjapsers pointed me to:
http://movies.trekcore.com/gallery/index.php?cat=43
The brightness is a lot higher than on the DE-DVD, but the quality is substantially better. I don't trust my own renders - if I view the website from work, the colours are less saturated than at home. When I look at the earlier set of animation snapshots posted above, I can barely make out the blue and red hues. I put that down to near-white colours washing out on a poor quality monitor.
137th Gebirg - thanks for your comments. To have someone say it looks completely dead-on, is about as good as it gets. However there is plenty of work to do and I probably need to do as Wil suggested and run a test against a scene from the film. I agree that adding in other elements would help bring it to life. But to me it feels like more should happen to the surface as I move the camera. As mentioned before, blurred reflections would help. I use them in stills, but not in animations.
Cheers,
S.O.
USS Mariner - I don't think the HD references are from the blu-ray. I'm using a site wjapsers pointed me to:
http://movies.trekcore.com/gallery/index.php?cat=43
The brightness is a lot higher than on the DE-DVD, but the quality is substantially better. I don't trust my own renders - if I view the website from work, the colours are less saturated than at home. When I look at the earlier set of animation snapshots posted above, I can barely make out the blue and red hues. I put that down to near-white colours washing out on a poor quality monitor.
137th Gebirg - thanks for your comments. To have someone say it looks completely dead-on, is about as good as it gets. However there is plenty of work to do and I probably need to do as Wil suggested and run a test against a scene from the film. I agree that adding in other elements would help bring it to life. But to me it feels like more should happen to the surface as I move the camera. As mentioned before, blurred reflections would help. I use them in stills, but not in animations.
Cheers,
S.O.
S.O. because the paint used on the saucer is the same as on the hull, you can use this sequence, for some more reference.
But you will need to look to the animation, not the stills. From the animation you can see how the different colors behave and how they change with the camera view.
Also, when the ship has passed, the saucer reveals some of the colors, maybe you can compare them to the B&W.
Anyhow, looking good, and don't forget you have fairly bright light on the saucer, when you turn down the intensity (as inside the dock), the specular/reflecting colors become more visible than the underlying brighter paint.
But you will need to look to the animation, not the stills. From the animation you can see how the different colors behave and how they change with the camera view.
Also, when the ship has passed, the saucer reveals some of the colors, maybe you can compare them to the B&W.
Anyhow, looking good, and don't forget you have fairly bright light on the saucer, when you turn down the intensity (as inside the dock), the specular/reflecting colors become more visible than the underlying brighter paint.
Hi Folks,
I needed a bit of time away from the project, thanks in part to toothache and because I was having problems on seeing a way forward.
Now I'm back and I've taken out a hack that was bothering me and made some more changes. The results are a step in the right direction. In the following images I'm lighting the saucer with blue light so that it is similar to the movie snapshot. Until now, I was having trouble generating the various purple and blue shades in the same render. But that has pretty much been resolved - in the one render, you can see purple specular hightlights and a nice blue inversion zone at 7 o'clock.
But there is still the issue of hues and tests here have not been anywhere near so good. In the following render, I've allocated different hues to a number of shades. In white light, it's not to bad and similar to images I've seen of paint jobs done on physical models. But in blue light it goes wrong - there is too much colour in the highlights when you want them to be predominately purple.
So I need a different strategy and the answer might come in a couple of the shaders that come in C4D - DANEL and LUMAS. They are essentially the same - both provide the ability to add three layers of specularity and use anisotrophy. I'm thinking that one layer will contain the existing red highlights and another layer will be used for the different hues. I've got to then work out how they will interact.
But that's for another day.
Cheers,
S.O.
I needed a bit of time away from the project, thanks in part to toothache and because I was having problems on seeing a way forward.
Now I'm back and I've taken out a hack that was bothering me and made some more changes. The results are a step in the right direction. In the following images I'm lighting the saucer with blue light so that it is similar to the movie snapshot. Until now, I was having trouble generating the various purple and blue shades in the same render. But that has pretty much been resolved - in the one render, you can see purple specular hightlights and a nice blue inversion zone at 7 o'clock.
But there is still the issue of hues and tests here have not been anywhere near so good. In the following render, I've allocated different hues to a number of shades. In white light, it's not to bad and similar to images I've seen of paint jobs done on physical models. But in blue light it goes wrong - there is too much colour in the highlights when you want them to be predominately purple.
So I need a different strategy and the answer might come in a couple of the shaders that come in C4D - DANEL and LUMAS. They are essentially the same - both provide the ability to add three layers of specularity and use anisotrophy. I'm thinking that one layer will contain the existing red highlights and another layer will be used for the different hues. I've got to then work out how they will interact.
But that's for another day.
Cheers,
S.O.
I can only imagine the headaches and frustration this project brings you, Olsons paintjob certainly isn't the easiest to emulate but keep it up, I'm sure you'll find a way to make it work.
Very frustrating!
It's the old 80:20 rule (or whatever ratio you want) where 80% of the work takes 20% of the time. You would think adding 3 or 4 hues would be simple. I had a quick go last night using the dual-specular layer idea and it showed a few glimmers of hope. Problem is that specular colours were still stubbornly orange-yellow or red.
I've come across a thread on HobbyTalk that may yield some information.
http://www.hobbytalk.com/bbs1/showthread.php?t=273113&page=18
This is a paint job by a guy called "The TrekModeler" and I think he's got the elements right. I was starting to think that maybe some of the movie shots used a blue filter on the camera lens, but on page 18 of the thread you can see the effect I'm trying to simulate.
I've got to look at some of his demo videos for some clues, but what I really need is a video using one white light pointing at varying angles to the bottom of the primary hull. Better if the specular could remain visible by varying the camera angle too! Then I'd be able to see how this thin iridescent blue layer comes into play.
Can't help you out with any clues to the shaders in C4D, but, when you look at the last movie screenshot you posted, you might get the expression that in the bright part of the spotlight the specular color is magenta, and in the sides the specular color is more bleu/cyan. This you can do with a shader called fallof, shadow/light, if that's avaiable in C4D.
Hi Wil,
I think falloff will enter into this at some point, but not yet. My focus is on the hues.
If you look at the movie snapshot, the panel at 2 o'clock from the phaser bank appears to contain 4 hues:
- The single bright square a third of the way down on the LHS - slightly cyan.
- A 3x1 bar at 4 o'clock to the above square - slightly red.
- Remaining bright sections - purple to dark cyan (either due to natural falloff of the highlight or an additional falloff/gradient).
- Dark sections.
You could say it's a bit like:
Resultant highlight = fn(light colour, material colour, iridescent blue layer, iridescent random colour)
At the moment it looks like my approach has collapsed the final two arguments into one and I need to somehow pick them apart. It needs to yield near purple highlights under blue light and the original iridescent random colour under white light. This is assuming I'm correct about the blue light colour in the movie snapshot!
I really need to work out the behaviour of this thin blue layer under white and blue lighting. Hopefully TrekModeler's paint job images and videos will yield some information I can use.
Please keep the suggestions coming.
Cheers,
S.O.
Well I've just reviewed TrekModelers thread - his paint job is jaw dropping good. Some of his sample photos remind me of Andrew Probert's painting.
There is a lot of information here - I might need to redo the colours for the colour channel as well as the specular highlights. That red specular highlight might be heading for the bin.
There is a lot of information here - I might need to redo the colours for the colour channel as well as the specular highlights. That red specular highlight might be heading for the bin.