Spec/Power/Environment Adventures

[Vraciu]

Spot on Greebo.

If you want to create the uneven specular of a metal which is wavy or distorted, then you do that in the normal map.

I am not talking about wavy panels.  I'm talking about different shades of aluminum and the mottled appearance of same.  This is about material composition/specularity/color not shape.

See what I mean?

Also, perhaps someone can help me but panel waves on the normal map are creating random squares around rivets not waves or dimples.   It looks odd.     The screenshots of the default I posted show this.  I believe Devil posted some from his skin conversions. 

[Devil 505]

I have a theory that the squares are limited only to shallow indented areas being represented by the normal map. I can get nice subtle bumps but not divots.

What I need to test is making raised rivets to see just how the system handles them. If raised rivets look good but recessed ones bad, then my theory would be confirmed.

[Vraciu]

I think Skuzzy is saying that if all the bare metal panels are made of the same material, i.e. aluminium, then they should be the same base shade on the spec, power and environment maps, excluding weathering effects. I'd add that if there are some metal panels made of heat proof metal or armour plate then these might have a different finish and be shaded differently on the maps. Making individual aluminium panels darker or lighter on these maps so they reflect light differently to each other would help the illusion of the skin being made of a patchwork of metal panels but doing this is likely to get your skin rejected. Making the panels different colours on the diffuse map is OK because they plainly were different shades in RL (different gauge and alloys etc).

Here's the ubiquitous photo that people reference for metal composition. 



Is it more effective treating the different panel colors the same by making them all the same shade of gray in the specular map or should I allow the variations to remain?

It seems the consensus is to convert them a single color in the spec maps.   That's what Devil's 109 has. 

[Skuzzy]

There will be some artifacts, in the normal map, due to texture compression.  It is something we have on the list.

[Devil 505]

There will be some artifacts, in the normal map, due to texture compression.  It is something we have on the list.

Is that what is causing the squares with the divots?

[Skuzzy]

Is that what is causing the squares with the divots?

In all likelihood.

[oboe]

Spot on Greebo.

If you want to create the uneven specular of a metal which is wavy or distorted, then you do that in the normal map.

I think I'm getting mixed up.  The normal map replaced the bump map, correct?  And its function is to alter the reflection of light to simulate variations in surface height, such as dimples around rivets, recessed areas, deformations in fabric with ribs underneath, etc?

I wouldn't think that all aluminum is created equally or have identical specularity.   For example, an aluminum panel with a grease stain on it.  The section of the panel with the grease stain would exhibit lower specularity than the clean section of the panel - even though it is made from the same material.   

In other words, weathering (oxidation), grime, exhaust soot buildup - they would each alter the specularity of different areas of the same panel they contaminate, wouldn't they?   Or would these contaminated sections of the same panel constitute different materials for the purposes of specularity?

[Skuzzy]

I think I'm getting mixed up.  The normal map replaced the bump map, correct?  And its function is to alter the reflection of light to simulate variations in surface height, such as dimples around rivets, recessed areas, deformations in fabric with ribs underneath, etc?

I wouldn't think that all aluminum is created equally or have identical specularity.   For example, an aluminum panel with a grease stain on it.  The section of the panel with the grease stain would exhibit lower specularity than the clean section of the panel - even though it is made from the same material.   

In other words, weathering (oxidation), grime, exhaust soot buildup - they would each alter the specularity of different areas of the same panel they contaminate, wouldn't they?   Or would these contaminated sections of the same panel constitute different materials for the purposes of specularity?

Yes oboe, but that also means your diffuse will reflect the differences as well.  I am sure there is the rare case where the diffuse is constant and the specular might need to be changed, but that would be the exception, rather than the rule.

[oboe]

Yes oboe, but that also means your diffuse will reflect the differences as well.  I am sure there is the rare case where the diffuse is constant and the specular might need to be changed, but that would be the exception, rather than the rule.

I think I need a picture to make sure I'm understanding what I've asked and what you're answering.  Here's the default specular map file for the P-51D:


If I look at the wing underside for example, even though its all made of the same material (aluminum), I see all sorts of variation in shades - light for rivet heads, dark for gun muzzle stains, even the hint of a bump mapping effect to show the depressions in areas where the rivets attach to the wing ribs.   You can also really see this effect well on the horizontal stabilizers.  When I look closely at the 3 signal lamp lenses on the lower left wing tip, I can even see shade variations to simulate reflections from the curved glass of the lense.   But the lense glass is all one material, so should are you saying it needs to be all the same shade?

Or looking at the screw heads on the fuselage side where the wing fillet is depicted - you can see the light-dark shading (shadow/glare) which depicts the local deformation of the metal by the screwhead.  Now, I would expect this to also be built into the normal map - but are you saying this technique is not allowed in spec maps?

I also see white panel lines, which give the edge of the panel higher specularity than the rest and thus simulate the glare effect from the panel's edge - but the panel edge and the panel surface are the same material, so I thought I understood you to say they should have the same specularity shade.

I hope I'm just misunderstanding.  But the way I interpreted what you said would cause a skin submitted with spec map similar to this one to be rejected...

[Vraciu]

That's how I took it as well.   

[Greebo]

The default P-51D spec map looks to have been grey-scaled from the diffuse and then had the painted areas crudely blacked out. It is not a very good job as only some of the paintwork has been blacked out, the stuff on the nose was missed. I suspect this may have been done when specular maps were first introduced into AH2 and Waffle, Superfly and the Skinner Team were rushing to create these maps for all the default skins in the game.

If that map was submitted now it would almost certainly be rejected for the issues Oboe and I have mentioned. With new skins in AH3 it seems HTC are clamping down on issues that would have got by in AH2. So ignore what the default skin looks like if it dates from AH2, many of those are not up to the standards required in AH3.

[Skuzzy]

Greebo is on a roll.  Spot on.

I'll also add, we have been replacing default skins with better skins, as they come along.  Not been in a big rush to do so, as it requires a lot of time from Waffle and HiTech to get a default skin replaced.

[Vraciu]

Anyone willing to explain why it would be rejected since it is widely considered the Cadillac of Mustang skins?

It has a few minor issues due to the conversion but overall it's pretty slick.

Wish I had known to ignore it months ago as I was taking a lot of my cues from it thinking that was what was expected...  *facepalm*

[oboe]

Thank you both.  Its probably unfortunate that a default spec map may not comply with the new standards, seeing how the default files are provided to the skinner as a working base and the skinner would probably assume that following that example would be OK.   I do understand how it came about though.  Are the new standards written down anywhere?  I feel like there are probably many standards that I am unaware of and would get tripped up by. 

Greebo's tutorial has been an excellent resource, but I wonder whether that is up-to-date.  It talks about the bump map file, and how it is no longer used except to create the Normal map.  I'm not sure I ever created a bump map file in my earlier skinning days - I think I used to just bake in my surface effects the best I could using  shadow and glare layers described in Fester's old tutorial.    However I have created usable Normal maps by starting with the Diffuse map, eliminating the layers that don't have anything to do with surface deformations or texture, and then applying the nVidia plugin filter.   Is that OK to do?   If I need to create a bump map file as an intermediate step toward the Normal file, I'll need some refresher on how to create it.

Here is another point of confusion I have, and it came up as a result of this discussion on different materials and how they are depicted in the Specular map.  I'll use Greebo's Ki.43 maps, directly from his tutorial.   This is a bare metal Ki.43 with fabric control surfaces and some painted stripes and a painted band around the rear fuselage:



I have two questions.  First, the rivets.  Rivets are all made of the same material and depicted as dark grey dots against the unpainted aluminum fuselage.  In the area of the painted band however, some rivets are shown as white dots.  I assume this is depicting paint chipped off the rivet head, but why then wouldn't it be the same color as the other bare rivet heads shown on the unpainted fuselage?  Why would having the paint chipped off a rivet head alter its specularity to bright white?   It should be the same material now as the other unpainted rivet heads.   So we have two different specular values (dark grey and bright white) for the same material (rivet heads).   Isn't that incorrect, given the standard?

My other question is about the control surface (elevators and rudder) hinge shadows.    Why are shadows being depicted on a specular map?  Again, the surface is the same material whether it is in shadow or not, so why the difference in specular values?  Honestly I don't know where the shadow layers belong now, or if we are not even suppose to create them directly - maybe the graphics engine is supposed to show them based on the sun angle and the Normal map?

I think I do understand correctly that a single material, aluminum for example, may have different specular values based on surface weathering and discoloration, paint, grease or exhaust stains, etc.  But a shadow seems like a wholly different kind of effect...

EDIT: I just realized that last question may contain the explanation for bright white rivet heads - the unpainted rivet heads have weathered for some time, exposed to the elements, and are no longer shiny as they were when newer.  But a rivet head that had been protected by paint until it was recently chipped off would have that shiny newness exposed.... 

[Greebo]

The tutorial is not up to date. I wrote it during the beta as a guide for the then current skinners who were used to bump maps. I've altered it once already but it could do with another update with the bump map references revised and the required bit depth of the various files added.

If your existing skin doesn't have a bump map then yes you use the diffuse map as the starting point. Rename a copy of it xxxx_N.xxx and edit anything out of it that isn't height related. Typically you might leave the panel lines, fastener and rivet layers, raised fabric strips and so on. I tend to do stuff like fabric sag as well but that's just me.

If the level of grey on the paint-chipped rivets is lighter than on the bare metal rivets that would be a mistake on my part, but its not going to be noticeable on the skin.

The shading effect on the rudders is ambient occlusion. This is where light gets bounced around between two adjacent surfaces before being reflected back. The effect is to make the concave junctions of a shape darker and I have been adding it to my diffuse files for ages. My view is that the specular and environment maps control how much of the sun's and the environment's light is reflected back to the viewer and ambient occlusion would reduce some of this reflected light and so should be included in the maps. Note I haven't included it in the power map, which just controls how rough the surface appears.