pearl paint doesn't behave anisotropically, it's an incidence effect, so I don't think that's the right way to go
imagine this is a microscopic cross section of the way the pearl paints went on, this is what the normal map is trying to copy:
(the effect is actually more within the paint, but the principle is the same)
First you need to know if he used this paint (pearl paint), pearl paint does not flip flop, it just has all the colors of the rainbow, displaying them like a sea shelve. Painting squares on a ship would indeed give a random coloring. But they come just in one flavor, displaying all colors. You can not buy a green one, not a bleu one etc. It would display ALL colors within every painted square. From the description Olson gives, he used flip flop paint, but they behave different as your microscopic cross section.
my $0.02
First you need to know if he used this paint (pearl paint), pearl paint does not flip flop, it just has all the colors of the rainbow, displaying them like a sea shelve. Painting squares on a ship would indeed give a random coloring. But they come just in one flavor, displaying all colors. You can not buy a green one, not a bleu one etc. It would display ALL colors within every painted square. From the description Olson gives, he used flip flop paint, but they behave different as your microscopic cross section.
my $0.02
flip paint then, it's just semantics, I know the paints he used but it was 2 years ago when I set this material up, so do excuse me if I forget some terms, I did study a couple of research papers to understand how it works and get the flip angles correct so I'm pretty confident I've got the paint accurate
seems like people are a bit skeptical that anyone can understand this effect, in reality it's not that complicated, and it's certainly not the un-reproducable enigma everyone makes it out to be
Isn't a issue of understanding it, it is implementing it with the tools available. And yes, I do not care about the names etc, just the effect and the ability to control the details of that effect. The stuff I am using is limited in the controls available, reason why I think I am approaching it the wrong way. Guess it is a lack of knowledge of the available methods within max.
I do appreciate the time and effort everyone is putting into it.
MKF, have youn tried using a Gradient Ramp in normal mode? (like lennier suggested in chris' thread)?
Gives you basically a fallof that's more customizable than 2 slots that you have to nest over and over again
It would keep you from having to mix between red green and blue patterns by hand, just drop them in as nodes of the gradient and put them at whatever angle you want them to show
This doesn't really do what irml says, but it might make what you are doing easier. Haven't really read through irml's posts yet, but off hadn I can't really think of a way to modify normals at the moment.
NO I have not, sorta missed lennier's comment. I have to take a look at that I ignored gradient ramps since I did not think I could use them to create the incidence as a non fixed point. Erm, hopefully that makes sense? lol Just woke up and decided to refresh the forums. Anyhow I didnt think I could get them to do what I wanted so I will have to give it a go and see how it works out.
What I would like to do is tie that aniso rotation map to the individual pearl effects within a layered material.
IE layer mat (doesnt exist as a base mat in max that I know of) with say however many layers there are of pearlish stuff each with their own aniso spec reflective etc settings stacked in how they would appear.
I thought one way would be to mess with that auto shader but it just hangs max when not used as diffuse.
ok when I set this up it was a few years ago, I read the details from the guy who painted the model describing how he did it, it didn't quite translate into 3D like you'd expect so I had to change a few things, but it's more or less the same
to start with the model was painted matte white, so that bit is easy:
the next layer is the aztec pattern, this one uses blue
if you look carefully you can see the tweaked normals trick in action here - each panel is flipping colours slightly differently, this is important because the guy painted the model one panel at a time, the paint was the same but it was applied slightly differently for each panel, so maybe through the thickness of the paint being different or through the angle each panel was sprayed at being different, the iridescent effect is slightly different for each panel, and that's what tweaking the normals will simulate:
next is the random panels (these have the same normals trick too), you can do as many layers here as you like, the guy did dozens but it gets a bit impractical so I just did one layer for each colour:
flip paint then, it's just semantics, I know the paints he used but it was 2 years ago when I set this material up, so do excuse me if I forget some terms, I did study a couple of research papers to understand how it works and get the flip angles correct so I'm pretty confident I've got the paint accurate
seems like people are a bit skeptical that anyone can understand this effect, in reality it's not that complicated, and it's certainly not the un-reproducable enigma everyone makes it out to be
The way I understand you did this, is by using the normals of the model, making the effect (flip) random. As you say, make each panel flip slightly later or sooner. In LW you call this a normal map, in 3dsmax this is a falloff map, it takes the normal of the plane to flip.
However, this paint does not work this way, it does not flip because the way or rotation it was painted on, it flips because the way it hits the (two different) particles in the paint and are reflected to your eyes. One can even calculate which color is shown.
Sometimes the one sort of particle wins in reflecting, sometimes the other particle wins, depending on the amount of light AND where this light is coming from.
Here is a good example: http://www.youtube.com/watch?v=L2RGR-MaXdg
But this does not change the fact that you did a good job.
I do have an incidence effect like that; but the angle the paint goes on at, the number of layers of paint, the thickness of the paint etc can slightly change the way it behaves. This is what I've tried to duplicate, but don't let the exaggerated tests and examples fool you - in reality it's quite subtle.
*EDIT* - I should explain... at the time I was originally working on this I did factor in the light direction, but I came to the conclusion it wasn't necessary for two-tone paint in the way it was being used here.
*EDIT 2* - Also remember your example video is a different paint, multi-colour for a start, but more importantly it's not a clear coat.
I do have an incidence effect like that; but the angle the paint goes on at, the number of layers of paint, the thickness of the paint etc can slightly change the way it behaves. .
Yes it does change according to thickness, and some colors are painted on top of eachother giving it yet another color.
This is what I've tried to duplicate, but don't let the exaggerated tests and examples fool you - in reality it's quite subtle..
No, the way you do stuff is superb, just trying to clear some things, not all members in here do understand what is going on, and I had the impression this thread was started to learn something, or try to tune in.
*EDIT 2* - Also remember your example video is a different paint, multi-colour for a start, but more importantly it's not a clear coat.
Yes I know, but it is a good clear sample of flip flopping (where is the sun, where are the eyes, what is the light intensity in combination with the angle). You can however get the same paint as clear coat and paint it on a black vehicle to get the same result, ofcourse the majority of the car will be black, but you get the same flip flop in the "specularity".
Yes I know, but it is a good clear sample of flip flopping (where is the sun, where are the eyes, what is the light intensity in combination with the angle). You can however get the same paint as clear coat and paint it on a black vehicle to get the same result, ofcourse the majority of the car will be black, but you get the same flip flop in the "specularity".
sorry I should have explained this in the last post, I'm not saying you were wrong on the paint just because it's a clear coat, but I was saying that because it's only visible on specular highlights this negates light direction - because of the way specular highlights work and only the two colours to flip between I worked out it wouldn't be necessary to factor in light direction - just the incidence effect on its own worked because any further flipping based on light direction would have taken place outside the highlight (hope that makes sense)
obviously there's confusion between different 3D apps, I wasn't sure if you were telling me I'd got it wrong or not, it's fine if you do think I've got it wrong, I was just trying to explain why I think I've got it right, I'm not pretending the normals trick is physically accurate or anything like that, but it's visually similar to my references which is good enough for me
getting there. Altered the order of things but, it oddly kept flipping the other colours! Seems set now. The gradients were odd, and did not animate well.
obviously there's confusion between different 3D apps
The good news is, plugging in some random normal vectors/colors in a "normal bump" shader in the bump slot seems to effectively bend the normals the same way it seems to work for you in lightwave.
I can even tell it to not affect diffuse.
The bad news is, there seems to be no way to trick 3ds max into not doing the parallax stuff around edges in the normal map.
From some angles, the borders of the panels seem blurred.
here's a test with just a reflective (actual reflections, not just specular) material and a normal map, the material has the same reflectivity everywhere, no anisotropy.
Funny I set up a material like this last night and set to render but it crashed in the night. Here is a new test of it with no enviro and just 2 lights. Simplified it over what I had last night since yours looked so good. I know the flip is fail and the colours are off but it is just a test of the normal in the bump slot. Earlier I had tried to do this with the displacement thinking it would produce a better effect, it doesn't.
I was so absorbed with getting the colour effect I was loosing perspective. I should have worked to get the proper reflective properties over the colour then worked on the colour. The aniso was working but I think this mixed with aniso should prove workable.
NOW I am going to put a halt to my exp for now until I can locate my TMP dvd or pick up the blurays (that or clips in HD) I have TWOK on my dvr but by then ILM had matte coated the ship. (though nothing saying replicating that hull isnt of importance too)
I am going to attempt this in scanline though, least the normal bump effect.
A question these normal maps of random squares, are you all generating these with filter tricks in PS or something or are you manually placing the colours? When I do it in the latter it is too regular and looks like crap when used.
Mmm affects the diffuse too much BUT when looking at the other films this might be a good thing? Anyhow seems to look better in scanline! -_-;' only took 5min to vs the what 2 hours? oh well this should help point those using standard mats to where they need to go as well. Flowcharty thing later on too.
NOW I am going to put a halt to my exp for now until I can locate my TMP dvd or pick up the blurays (that or clips in HD) I have TWOK on my dvr but by then ILM had matte coated the ship. (though nothing saying replicating that hull isnt of importance too)
you just reminded me of something, apparently the ship was too reflective to film for TMP so they dusted it down with talcum powder before shooting, if this is true then another layer on top is required at the end, but at least this time it's a simple layer - i.e. white matte diffuse (probably oren-nayar) and a plain incidence gradient for the mask
Well I had read somewhere that after TMP they, ILM matte coated the whole ship. When you say layers max does not manage materials like this, I have to set the surface params with the gloss and reflection settings with aniso as a means to control the reflection brightness as a whole. (well as per what aniso is) I do have a roughness setting to control too likely this might create the effect. I have to look into it. I figured I would have to work based on how it appears rendered out and match it to screen examples as best I can.
I found a few hd youturd vids man the filming of the ship in TMP least on the bluray was pretty bad. Out of focus and a lot of filth on the lens. (one scene it looks like pencil erasure bits) lol
But yeah I will have to take in account screen and actual evidence. Sadly all my refs are post ILM alterations to the ship. I think I have 2 or 3 really grainy or badly moried images of her in TMP trim. (all shiny)
The only thing so far I have not found a way is to replicate that layering your able to do in LW. Controlling each of the bits to be of various strength (eg the random panels) It really makes me consider seeing if max 2012 has been updated to match it when they added that ugly flowchart material editor. There are some options I have to see how I can apply them that might enable me to get around it. But I am hitting a wall of knowledge atm. That normal bump shader, all the results in google or even the built in help files just referred to game modeling, and nothing about using that shader to create effects! Really frustrating. I can bet I coulda gotten that data if I was a paid subscriber to that elitist max forum over at discreet. -_-;'
BUT anyhow I have learned many new tricks that I can now use to replace other BS methods and to add to what I already know. Big question is if I will remember it to use it later! Unlike animation and rigging I have to keep referring to the same tuts over and over to remember how to do crap. :O
the random normals were just three solid colors mixed together, R/G/B, in a composite map, masked with different random white "Tile" procedurals with high color variance
127/127/255 should be a neutral "up" normal
Do not mean to sound like a dick BUT, yeah because everyone has 1300$ to blow on a renderer.
Sorry no,
I also do not use anything outside of the stock app for compatibility. Already been screwed by "plugins" that die (app ver) or meshes that use plugs without a note in the read me or anything and cost a bundle etc so on. My rule is to use what the app comes with. Least in regards to max where it is bundled as such.
the random normals were just three solid colors mixed together, R/G/B, in a composite map, masked with different random white "Tile" procedurals with high color variance
127/127/255 should be a neutral "up" normal
So your saying that purplish bitmap is shader generated? I'll have to look into that. Ok tiles you mean the checker shader or tile since tile I can only assign grout. LAWL A screen cap would help a lot since I cannot get the effect granted only doing this in the 5 or so min I have available to me atm.
AH NM I figured it out in like 1min after taking a break from it DURRR TILE shader.
Posts
First you need to know if he used this paint (pearl paint), pearl paint does not flip flop, it just has all the colors of the rainbow, displaying them like a sea shelve. Painting squares on a ship would indeed give a random coloring. But they come just in one flavor, displaying all colors. You can not buy a green one, not a bleu one etc. It would display ALL colors within every painted square. From the description Olson gives, he used flip flop paint, but they behave different as your microscopic cross section.
my $0.02
seems like people are a bit skeptical that anyone can understand this effect, in reality it's not that complicated, and it's certainly not the un-reproducable enigma everyone makes it out to be
I do appreciate the time and effort everyone is putting into it.
Gives you basically a fallof that's more customizable than 2 slots that you have to nest over and over again
It would keep you from having to mix between red green and blue patterns by hand, just drop them in as nodes of the gradient and put them at whatever angle you want them to show
This doesn't really do what irml says, but it might make what you are doing easier. Haven't really read through irml's posts yet, but off hadn I can't really think of a way to modify normals at the moment.
What I would like to do is tie that aniso rotation map to the individual pearl effects within a layered material.
IE layer mat (doesnt exist as a base mat in max that I know of) with say however many layers there are of pearlish stuff each with their own aniso spec reflective etc settings stacked in how they would appear.
I thought one way would be to mess with that auto shader but it just hangs max when not used as diffuse.
The way I understand you did this, is by using the normals of the model, making the effect (flip) random. As you say, make each panel flip slightly later or sooner. In LW you call this a normal map, in 3dsmax this is a falloff map, it takes the normal of the plane to flip.
However, this paint does not work this way, it does not flip because the way or rotation it was painted on, it flips because the way it hits the (two different) particles in the paint and are reflected to your eyes. One can even calculate which color is shown.
Sometimes the one sort of particle wins in reflecting, sometimes the other particle wins, depending on the amount of light AND where this light is coming from.
Here is a good example:
http://www.youtube.com/watch?v=L2RGR-MaXdg
But this does not change the fact that you did a good job.
*EDIT* - I should explain... at the time I was originally working on this I did factor in the light direction, but I came to the conclusion it wasn't necessary for two-tone paint in the way it was being used here.
*EDIT 2* - Also remember your example video is a different paint, multi-colour for a start, but more importantly it's not a clear coat.
Yes it does change according to thickness, and some colors are painted on top of eachother giving it yet another color.
No, the way you do stuff is superb, just trying to clear some things, not all members in here do understand what is going on, and I had the impression this thread was started to learn something, or try to tune in.
Yes I know, but it is a good clear sample of flip flopping (where is the sun, where are the eyes, what is the light intensity in combination with the angle). You can however get the same paint as clear coat and paint it on a black vehicle to get the same result, ofcourse the majority of the car will be black, but you get the same flip flop in the "specularity".
obviously there's confusion between different 3D apps, I wasn't sure if you were telling me I'd got it wrong or not, it's fine if you do think I've got it wrong, I was just trying to explain why I think I've got it right, I'm not pretending the normals trick is physically accurate or anything like that, but it's visually similar to my references which is good enough for me
getting there. Altered the order of things but, it oddly kept flipping the other colours! Seems set now. The gradients were odd, and did not animate well.
The good news is, plugging in some random normal vectors/colors in a "normal bump" shader in the bump slot seems to effectively bend the normals the same way it seems to work for you in lightwave.
I can even tell it to not affect diffuse.
The bad news is, there seems to be no way to trick 3ds max into not doing the parallax stuff around edges in the normal map.
From some angles, the borders of the panels seem blurred.
here's a test with just a reflective (actual reflections, not just specular) material and a normal map, the material has the same reflectivity everywhere, no anisotropy.
http://dl.dropbox.com/u/21007284/tmptest.mp4
I was so absorbed with getting the colour effect I was loosing perspective. I should have worked to get the proper reflective properties over the colour then worked on the colour. The aniso was working but I think this mixed with aniso should prove workable.
NOW I am going to put a halt to my exp for now until I can locate my TMP dvd or pick up the blurays (that or clips in HD) I have TWOK on my dvr but by then ILM had matte coated the ship. (though nothing saying replicating that hull isnt of importance too)
I am going to attempt this in scanline though, least the normal bump effect.
A question these normal maps of random squares, are you all generating these with filter tricks in PS or something or are you manually placing the colours? When I do it in the latter it is too regular and looks like crap when used.
I found a few hd youturd vids man the filming of the ship in TMP least on the bluray was pretty bad. Out of focus and a lot of filth on the lens. (one scene it looks like pencil erasure bits) lol
But yeah I will have to take in account screen and actual evidence. Sadly all my refs are post ILM alterations to the ship. I think I have 2 or 3 really grainy or badly moried images of her in TMP trim. (all shiny)
The only thing so far I have not found a way is to replicate that layering your able to do in LW. Controlling each of the bits to be of various strength (eg the random panels) It really makes me consider seeing if max 2012 has been updated to match it when they added that ugly flowchart material editor. There are some options I have to see how I can apply them that might enable me to get around it. But I am hitting a wall of knowledge atm. That normal bump shader, all the results in google or even the built in help files just referred to game modeling, and nothing about using that shader to create effects! Really frustrating. I can bet I coulda gotten that data if I was a paid subscriber to that elitist max forum over at discreet. -_-;'
BUT anyhow I have learned many new tricks that I can now use to replace other BS methods and to add to what I already know. Big question is if I will remember it to use it later! Unlike animation and rigging I have to keep referring to the same tuts over and over to remember how to do crap. :O
It's called 'blend', you can mix together as many as you like.
ah apparently it does, srsly shows how my brain just isn't what it used to be to miss that connection! tnks it should help me a lot.
127/127/255 should be a neutral "up" normal
its late and I should have been to bed ages ago! lol Ill give it a go when I get home tomorrow.
have a look at vray's blend material. you can work in layers.
Sorry no,
I also do not use anything outside of the stock app for compatibility. Already been screwed by "plugins" that die (app ver) or meshes that use plugs without a note in the read me or anything and cost a bundle etc so on. My rule is to use what the app comes with. Least in regards to max where it is bundled as such.
AH NM I figured it out in like 1min after taking a break from it DURRR TILE shader.
IAâm just watching and learning with you all until now, cause I never did a good texture work but... I played a bit with blend in the past, and I believe youAâre right Steve. It will works. Here e goes two good examples of complex textures with the blend material:
http://www.cgshelf.com/vertex_paint.php
http://www.tutorialized.com/view/tutorial/Switch-Materials-in-mid-animation/41531
ItAâs not called AâblendAâ it is called AâcompositeAâ.
Blend is for mixing different materials, like rust on a plate. Composite is for stacking materials.
There is actually a button in 3dsmax for making these TMP aztecs:
help.jpg
It is called HELP , Have Enterprise Layer Production :fishslap:
Here my try at it:
To show you, I put the fallof in the diffuse channel:
Base material white:
1.jpg
Bleu cover:
2-1.jpg
Green aztec:
3.jpg
Red plates (red on bleu, and red on green):
4.jpg
Gold plates (gold on bleu, and gold on green):
5.jpg
All crap :
6.jpg
Maps only in specular slot:
7.jpg
test:
aztectest.jpg
compare:
aztectest1.jpg
Needs a lot fiddling with the colors :argh:
Video: