Press the "ALT" key on your keyboard, wait for the popup telling you that toolbag 2 can't reach the server (you have to be offline). You can then release the key, but when you press "okay", Toolbag 2 will crash.
I have been able to reproduce this crash multiple times.
Marmoset 2 is a joy to work with and I really like where it is headed with the new scene assembly features (move, rotate, scale, duplicate, group meshes etc).
Just curious are there any early whispers/plans for the environment art features to be expanded? For example a functional grid, vertex painting/material blending etc.
Marmoset trumps all other engines for ease-of-use and visual quality but it is missing certain features that would make it a viable alternative to UDK/Cryengine/Unity for non-playable environment art presentation.
In the long term we plan to add functionality aimed at easier management of larger scenes/environments. I can't really give you any specific information about features or dates, but yes, there is some cool stuff that we plan on adding with future updates.
In the long term we plan to add functionality aimed at easier management of larger scenes/environments. I can't really give you any specific information about features or dates, but yes, there is some cool stuff that we plan on adding with future updates.
I have a little downtime at work so I felt compelled to ramble about this for a little;
Right now I'm starting to plan out my next full environment and of course part of that planning process is choosing which engine to present in. Immediately , as soon as I started planning I considered using TB2 because IMHO the shader itself is vastly better than anything out right now. Not only that, but with a PBR workflow, I feel like I am prepping myself for the future and possible future job opportunities. But upon further planning, I realized that at this point using TB2 is probably not worth what you'd be be giving up from other packages.
Here are some feature suggestions that would make TB2 irresistible when
it came down to chose what engine to present in:
+Global Illumination and Bounce Lighting Options
---I have no idea how to implement it- sorry. But It's hard to not use UDK and CE3 because of GI which can really tie an environment together
+Keyframable Object, Camera, and Particle animation (based on local pivot) and/or Vertex-based animation
---This one is big for me. I am always trying to stay ahead of the curve as a 3D Artist and one of the things I think we're going to see a lot of in 2014/15 is dynamic, moving environments. Either presented in HTML5 .gif form or high-res video, I think having a dynamic environment is going to be necessary when creating next gen visuals.
+Vertex Color/Texture Blending
---An amazing way to break up tiling
+Decal Placement
---Well, this one can be worked around pretty easy in TB2 with a plane and additive blending but it would be nice to have a more dynamic way of using decals
+Grid Snapping
---For modualar placement
+Fog
---For... making stuff foggy
That's all I really got off the top of my head. Sorry to mention it if any of this stuff is obvious; I'm just a 3D art nerd and I love discussing this stuff =P
TL;DR: If TB2 can implement some key features from other packages, it will be a no-brainer to use with every project; including large scale dynamic environments!
maybe already mentioned
when rendering out a turntable the sky rotation is snapping back to default.
I can work around it by rotating the camera and mesh instead of the sky but its not as intuitive
I have a little downtime at work so I felt compelled to ramble about this for a little;[ Right now I'm starting to plan out my next full environment and of course part of that planning process is choosing which engine to present in. Immediately , as soon as I started planning I considered using TB2 because IMHO the shader itself is vastly better than anything out right now. Not only that, but with a PBR workflow, I feel like I am prepping myself for the future and possible future job opportunities. But upon further planning, I realized that at this point using TB2 is probably not worth what you'd be be giving up from other packages. Here are some feature suggestions that would make TB2 irresistible when it came down to chose what engine to present in: +Global Illumination and Bounce Lighting Options ---I have no idea how to implement it- sorry. But It's hard to not use UDK and CE3 because of GI which can really tie an environment together +Keyframable Object, Camera, and Particle animation (based on local pivot) and/or Vertex-based animation ---This one is big for me. I am always trying to stay ahead of the curve as a 3D Artist and one of the things I think we're going to see a lot of in 2014/15 is dynamic, moving environments. Either presented in HTML5 .gif form or high-res video, I think having a dynamic environment is going to be necessary when creating next gen visuals. +Vertex Color/Texture Blending ---An amazing way to break up tiling +Decal Placement ---Well, this one can be worked around pretty easy in TB2 with a plane and additive blending but it would be nice to have a more dynamic way of using decals +Grid Snapping ---For modualar placement +Fog ---For... making stuff foggy That's all I really got off the top of my head. Sorry to mention it if any of this stuff is obvious; I'm just a 3D art nerd and I love discussing this stuff =P TL;DR: If TB2 can implement some key features from other packages, it will be a no-brainer to use with every project; including large scale dynamic environments!
This is a decent short list, I’ll add all of these to the suggestion list.
maybe already mentioned when rendering out a turntable the sky rotation is snapping back to default. I can work around it by rotating the camera and mesh instead of the sky but its not as intuitive
Yes, sorry this is a known issue, we just haven’t had time to fix it yet.
Well this is alot of fun, although the local reflection thig is cool it seems to get really blurry for me.
Local reflections are blurred based on the concept of the gloss map, the glossier your surface is, the less blurred the reflections will be. I think on a technical level the reflections are blurred a little extra to avoid ugly artifacts as well.
This might have already been asked, but is there a way to disable mip maps by default? Been finding it slightly annoying having to go through every texture to turn it off.
I got a problem I can't seem to solve.
When I load certain textures, they are offset. For example, if I had a texture of a circle, I end up with a half circle on the left and right side of the texture, since it's tiling. It's not an isolated case.
In this case it's the gloss and metalness maps that gets offset.
If I load the diffuse map into either of those slots, it works.
If I load the gloss or metalness into the diffuse slot, they're offset..so it seems to be the texture itself as it loads that gets offset.
If i click the magnifying glass, the texture is offset. If I click the edit button, it shows up correctly.
All textures were exported straight from substance designer with no extra steps in between.
Didn't find anything on this problem when googling it..
I got a problem I can't seem to solve.
When I load certain textures, they are offset. For example, if I had a texture of a circle, I end up with a half circle on the left and right side of the texture, since it's tiling. It's not an isolated case.
In this case it's the gloss and metalness maps that gets offset.
If I load the diffuse map into either of those slots, it works.
If I load the gloss or metalness into the diffuse slot, they're offset..so it seems to be the texture itself as it loads that gets offset.
If i click the magnifying glass, the texture is offset. If I click the edit button, it shows up correctly.
All textures were exported straight from substance designer with no extra steps in between.
Didn't find anything on this problem when googling it..
It's because grayscale outputs from Designer are not read properly by Toolbag 1 and 2. The only solution for the moment is to plug a gradient map node just before your output node, to export an RGBA texture.
Hey EarthQuake follow up to our discussion about 100% metalness and not receiving shadows. I've done some tests at work with one of our technical artists and a lighting artist. Our engine handles this differently and after some testing I think this is actually a better workflow.
Our engine uses the gloss map to determine how much of the shadow is seen rather than how metal the object is. This seems more physically correct to me and is easier to understand when managing large groups of artists and getting consistency. It's much easier to say to someone that thing is 100% metal, but the metal is not pure chrome so the reflection is a bit glossy. Then the shader just makes the object receive less shadow as it gets closer to chrome/mirror in terms of smoothness. This also makes sense to me personally (not sure if it's scientifically correct) in the real world because my can of red bull in real life is 100% metal, but where the shadow is received it is not a perfect chrome reflection, it has imperfections and a brushed aluminum look.
Maybe someone could argue that there are other substances present in the metal can so it's not truly 100% made out of metal, but I think for game art this makes the most sense where you make the can 100% metalness and then adjust the gloss to get the material properties you are after.
I'd like to request that you guys actually update Marmoset to work this way as I think it's an easier pipeline to understand and author art for and should help maintain the coloured specular for metals while at the same time receiving shadows appropriately.
Hello marmoset peoples!
I ran into this thing the other day and wasn't sure if it was a feature or bug. Andy suggested that I post about it in here.
When changing saturation in post my dither material is washing out in a strange way very quickly. I assume it is something I am doing or maybe a combination of things.
Hey EarthQuake follow up to our discussion about 100% metalness and not receiving shadows. I've done some tests at work with one of our technical artists and a lighting artist. Our engine handles this differently and after some testing I think this is actually a better workflow.
Our engine uses the gloss map to determine how much of the shadow is seen rather than how metal the object is. This seems more physically correct to me and is easier to understand when managing large groups of artists and getting consistency. It's much easier to say to someone that thing is 100% metal, but the metal is not pure chrome so the reflection is a bit glossy. Then the shader just makes the object receive less shadow as it gets closer to chrome/mirror in terms of smoothness. This also makes sense to me personally (not sure if it's scientifically correct) in the real world because my can of red bull in real life is 100% metal, but where the shadow is received it is not a perfect chrome reflection, it has imperfections and a brushed aluminum look.
Maybe someone could argue that there are other substances present in the metal can so it's not truly 100% made out of metal, but I think for game art this makes the most sense where you make the can 100% metalness and then adjust the gloss to get the material properties you are after.
I'd like to request that you guys actually update Marmoset to work this way as I think it's an easier pipeline to understand and author art for and should help maintain the coloured specular for metals while at the same time receiving shadows appropriately.
glossiness doesn't in any way at all change the amount of shadow an object has. in every perceivable way that statement is incorrect.
what glossiness/roughness does is alter the microsurface of an object, changing how diffuse[i/] the light reflected is. it's important to note here that the word diffuse in this context has nothing to do with the texture map of that type. the more diffuse the light is, the more scattered it is. it might appear that there is very slightly less shadowing the rougher a surface is, but it's an illusion created by additional bounce light from the scattered reflections.
Malcolm, I will certainly pass that on to our engineers and see what they think about it (though I have a feeling it will be in line with Lee’s thoughts). In addition to what Lee says above, I think its important to discuss some concepts a bit further (from my understanding at least).
What we typically refer to shadows is actually an absence of light, specifically diffused light. Shadows are not “cast” or “added” or “multiplied”, even if this is what technically happens in 3D rendering.
As Lee points out, objects vary in the amount of light they diffuse, or absorb, and the amount they reflect. Raw metals do not absorb any light, they reflect 100% of light that hits them, so they have a diffuse value of zero (this can vary in real world materials depending on other factors, like oxidization, wear, etc, but this is the basic theory). With a metal material, or any material that has zero diffuse contribution, its impossible to get any darker than zero while “in shadow” as the diffuse contribution is already zero.
In reality, reflections are not shadowed per say, an object simply reflects the other object that happens to be blocking thing light. In this case some more advanced reflections would probably help but screen space reflections do a pretty good job.
This is a simple phenomenon to observe, go into a bathroom or anywhere that has overhead lighting, go up the sink (assuming the faucet is metal) and put your hand between the faucet and the overhead light. The faucet does not suddenly reflect less now that it is “in shadow”, it simply reflects your hand instead of the light. With very glossy materials, the size of that point light happens to be very small, so you will see very little difference in the overall reflection of the faucet (assuming there are more lights in the room, or lots of ambient light). The less glossy a surface is, the larger that point light source is going to be, and the more noticeable to the overall reflection is when its blocked/occluded by another, less bright object.
In practice, TB2 follows this behavior, with dynamic lights, the rougher your surface is, the wider the specular highlight is, and the specular highlight from dynamic lights actually is shadowed or blocked by other objects.
So here is a quick example. For the diving helmet asset I used the metalness thing, so the metallic areas have zero diffuse (side note: its important to mention again this isn't exclusive to the “metalness” method but applies to any material that has zero diffuse contribution). You can see that as shadows are enabled and disabled for the dynamic lights, that the specular reflection from said lights is clearly blocked/shadowed by the little pipe/valve extension thingy
As specular reflections are a view-dependent effect, it may appear that it is not shadowed, but it actually is. You just don’t see “diffuse” shadows, because there is no diffuse lighting with a diffuse value of 0.
Ambient specular reflection does not cast shadows, so there isn’t really any logical way to get shadow data. I guess real time ray traced radiosity would be the solution there, but that's obviously not something available in TB. Again, if you multiply a shadow onto the specular intensity, what you’re really doing is making the object less reflective, not receive less light, which doesn't make much sense.
Ok, so I hope that makes sense and I didn't say anything too stupid. I've argued with various people for the same feature ie: shadows that affect specular the same way they do diffuse, and every time I try, I fail to come up with a real-world situation where this is actually the case.
Hey guys, I "think" we're arguing the same point here. I would imagine in games rendering with chrome/mirror surfaces you don't want to see any shadows/lack of light on your fully smooth surface because it's simulating reflecting all the light back with no diffuse reflected light component. So in our engine when I say removes shadows with gloss what I imagine it does behinds the scene is lessens the diffuse contribution and therefore it looks like it is removing the shadow, but really there is no diffuse contribution to receive light and create contrast between light and lack of light therefore you can't see the shadow. Sorry I am not entirely sure exactly what happens behind the scenes in our shader.
Here's our test case in Marmoset currently
1. Make your albedo white, your metalness 100% and gloss 100% smooth. That should be chrome/mirror, note the lack of shadows/blocked light/diffuse contribution, which ever you would like to call it. This looks correct to me.
2. Make your albedo white, your metalness 100% and gloss 50% smooth. That should be a simulation of a pretty rough metal. The part that doesn't look correct here to me is the lack of shadows/blocked light/diffuse contribution, which ever you would like to call it. Am I misunderstanding this, it appears in real life when I cast a shadow on to a diffused metal object it does appear to receive shadow.
3. Now in Marmoset I can lessen the metalness to an arbitrary amount so I have some diffuse contribution and start to see received shadows, but as I lessen the metalness I am also lessening the amount of speculour colour which starts to make my metals look less metal. I was surprised to find out in the engine we are working with right now that when you make the object 100% metalness you still see received shadows, we did some tests and found out that it is actually our smoothness map which lessens the diffuse contribution and eventually makes the shadows look like they are not there.
I can't actually wrap my head around what it's doing behind the scenes, but the reason I like this workflow is a can make something 100% metal then just worry about dialing in the smoothness and the amount of diffuse contribution is taken care automatically by the shader, also I maintain my coloured spec for rough metal surfaces. I'm a bit confused though because I'm pretty sure our metalness map works the same as Marmoset where it tints the albedo darker and tints the specular hotspots to be the albedo colour, but it looks like there is some additional functionality which is driven by the gloss map. What do you guys think? Or are we saying that most rough metals should not be 100% metalness?
Malcolm: Would it be possible for you to set up a really simple test scene, say with some spheres with varying specular intensity and glossiness, with something between the sphere and the light to cast shadows, and then the same scene in TB2?
I understand if you can't show anything due to NDA, but I feel like this would give me a better understanding of the differences.
Yes that should be possible, sorry I just can't divulge anything about our engine or project we're working on. Will report back once I've got some comparison images.
Quick question, sorry if it has been asked before. But I can not make it work the displacement maps, you know how it's done? or what format it is?
with Normal map and parallax i have not problems, but with the displacement, has been impossible.
Dunno if my questions is silly but,
about the microfiber diffusion and the peach fuzz... What does a map look like for this?
this is what i painted recently for peach fuzz.
it was a quick and dirty approach but i bet it could be done even better by some decent hair alpha or even going further and baking down for real hair/fur systems.
this is what i painted recently for peach fuzz.
it was a quick and dirty approach but i bet it could be done even better by some decent hair alpha or even going further and baking down for real hair/fur systems.
Awesome thank you!
I was really wondering how it worked and couldn't find anything on the interwebs :P
Dunno if my questions is silly but,
about the microfiber diffusion and the peach fuzz... What does a map look like for this?
it's a grayscale map, it just masks the fuzz on or off. you might consider using a very fine, almost blurred noise for things like cotton, and then multiply over the top with cavity map to reduce fuzz inside things like seams, or if the fibres are large enough to see them, then help define them a bit better.
This might have already been asked, but is there a way to disable mip maps by default? Been finding it slightly annoying having to go through every texture to turn it off.
When I load certain textures, they are offset. For example, if I had a texture of a circle, I end up with a half circle on the left and right side of the texture, since it's tiling. It's not an isolated case. In this case it's the gloss and metalness maps that gets offset. If I load the diffuse map into either of those slots, it works. If I load the gloss or metalness into the diffuse slot, they're offset..so it seems to be the texture itself as it loads that gets offset. If i click the magnifying glass, the texture is offset. If I click the edit button, it shows up correctly. All textures were exported straight from substance designer with no extra steps in between. Didn't find anything on this problem when googling it..
In addition to Froyoks suggestion, try resaving the texture in photoshop, a simple resave may work, if not, try re saving as RGB/24bit.
I ran into this thing the other day and wasn't sure if it was a feature or bug. Andy suggested that I post about it in here. When changing saturation in post my dither material is washing out in a strange way very quickly. I assume it is something I am doing or maybe a combination of things.
Interesting, Ive tried something similar out here but I dont think it has to do with the dithered blending. I think its the emissive slot that is causing the problem. Can you turn off emissive and see if you have the same problem?
with Normal map and parallax i have not problems, but with the displacement, has been impossible. Thanks.
To use displacement you need to add a subdivision type (usually flat) as well as displacement. Displacement is a DX11 feature and will not work with older, DX10 video cards.
about the microfiber diffusion and the peach fuzz... What does a map look like for this?
This is not at all a silly question. First off, what the peach fuzz map is is simple a map of the intensity of the peach fuzz effect. Sort of like a specular map, darker values will give you less fuzz, and lighter values will give you more fuzz. This is very useful if you have say, cloth, skin, and rubber/metal/etc, you may want a certain value for skin, a certain value for cloth, and then black for other material types. Now what the content actually looks like is up to you, you can use very simple solid colors, Lee painted in some individual hairs into a fuzz map that seemed to work, you might want a little bit of the cloth pattern overlaid on the fuzz map, really you should just experiment and see what works.
Is it possible to bring in the hdri's from Toolbag 1 into Toolbag2? If so how? thanks Reed
Sorry, this is not possible. The file format for TB1 skies is not supported in TB2 and we have no plans to support that format at this time (the data is completely different). Unfortunately the source files for the skies that we created for Darkest of Days at 8monkey Labs have been lost so those skies can not be re-created in TB2. There are a lot of freely available HDR sky panoramas out there on the internet these days though, and it is very easy to load standard .hdr files and such, and place sky lights with the new sky editor. Overall the sky system is much more powerful in TB2 so I would encourage you to try it out. The getting started tutorial goes over the new sky system: http://www.marmoset.co/toolbag/learn/getting-started
Hey all, i couldn't find anything on this, and tried all settings, file formats, new drivers but no change.
The Problem is that everything in tb2 seems darker, Even the ui is hard to read, and not like the examples i saw in the tb2 tutorial. The captured images come out washed and light, no matter what i tried.
i am really liking the viewport renders of tb2, so its a shame i cant get them captured right.
Any tips ?, i was hoping someone already experienced it.
This is with default camera presets, sky settings all at 1 (brightness etc.), also selective brightness.
Yes, that works thanks for the tip . Gamma correction of 0.45 seems to do the trick. But i still hope there is a way to get to the core of it since it affects the whole program and also lighting.
Hey all, i couldn't find anything on this, and tried all settings, file formats, new drivers but no change. The Problem is that everything in tb2 seems darker, Even the ui is hard to read, and not like the examples i saw in the tb2 tutorial. The captured images come out washed and light, no matter what i tried. i am really liking the viewport renders of tb2, so its a shame i cant get them captured right. Any tips ?, i was hoping someone already experienced it. This is with default camera presets, sky settings all at 1 (brightness etc.), also selective brightness.
This is very odd and certainly should not be happening. Could you please give me some more information: What OS are you using? What video card are you using? Have you updated to the latest video drivers? If not, try that. Do you have a custom color profile or anything like that? IE: spyder color correction software
I'm wondering if any of you who are getting the long loading time bug with highpoly meshes could humour me by sending me your dxdiag's in txt format. I don't expect I'll learn anything useful but nonetheless I'd like to see if there's any common things between our systems. Thanks.
Can somebody tell me how to set up alpha hair in marmoset? Because i've seen people render their hair with no problem, but i wonder why mine didn't work?
Sorry if this has been asked before, but is there a way to export a mesh in a tbscene to obj, or whatever other format? (Yeah I know 99.99% of the time this is useless.)
Do you plan to have the 3D connexion mouses work in tb2 ?
Just curious as I have one and probably know this feature if plan is something at the bottom of the list.
But might be cool to move with the 3D mouses into tb2
@JimmyRustler - Place your alpha texture in the alpha channel on your diffuse. Then on your material use the transparency drop down to select either cutout or dither.
I'm wondering if any of you who are getting the long loading time bug with highpoly meshes could humour me by sending me your dxdiag's in txt format. I don't expect I'll learn anything useful but nonetheless I'd like to see if there's any common things between our systems. Thanks.
Can somebody tell me how to set up alpha hair in marmoset? Because i've seen people render their hair with no problem, but i wonder why mine didn't work?
Put your mask in the alpha of your diffuse, then set the transparency module to dithered or cutout. Dithered is for soft blending, so it should be perfect for hair. It may be a bit noisy in the viewport but will look better with exported images. Cutout is basic alphatesting. Also be sure to try the anisotropic reflection mode (or add aniso to the secondary spec). This works really well for hair, especially if your uvs for your hair all run in the same direction, then you can use the simple direction slider in the aniso shader. Otherwise you will need a direction map.
Sorry if this has been asked before, but is there a way to export a mesh in a tbscene to obj, or whatever other format? (Yeah I know 99.99% of the time this is useless.)
Sorry, no, there is not a feature to export meshes.
Do you plan to have the 3D connexion mouses work in tb2 ? Just curious as I have one and probably know this feature if plan is something at the bottom of the list. But might be cool to move with the 3D mouses into tb2
Unfortunately no this is not something we support, if there is a lot of user interest for support we will consider adding it, or maybe if someone sends us one.
when I have UV islands, I get visible seams within Marmoset2.02 - even without a texture assigned.
is there a fix for that?
I tried it with exported meshes from Zbrush, Topogun and Blender.
no matter where I export the meshes, -tried with different ones- I can replicate this issue.
edit;
after a few experiments I found the missing setting in Blender to fix the visible seams on the mesh within Marmoset.
the option "Loop Normals" during OBJ export has to be activated.
apparently the name for the export function got changed from "Include Normals" to "Loop Normals" in one of the recent Developer Builds.
for now only Lowpolys exported from Zbrush will give me the visible seam issue.
So far i've been importing all of my maps with sRGB checked. my normal map looks too light without it in the map previewer but the texture looks better without sRGB checked. Which should I do?
what you see in the mip previewer doesn't really matter in that regard.
So for your normal keep sRGB unchecked so it actually looks correct where it matters, the viewport.
when I have UV islands, I get visible seams within Marmoset2.02 - even without a texture assigned.
is there a fix for that?
I tried it with exported meshes from Zbrush, Topogun and Blender.
no matter where I export the meshes, -tried with different ones- I can replicate this issue.
edit;
after a few experiments I found the missing setting in Blender to fix the visible seams on the mesh within Marmoset.
the option "Loop Normals" during OBJ export has to be activated.
apparently the name for the export function got changed from "Include Normals" to "Loop Normals" in one of the recent Developer Builds.
for now only Lowpolys exported from Zbrush will give me the visible seam issue.
Looks like you sorted it out. In general meshes from zbrush have a lot of issues with smoothing/normals etc because Zbrush’s exporter is really limited. Generally its best to run them through another 3d app beforehand.
So far i've been importing all of my maps with sRGB checked. my normal map looks too light without it in the map previewer but the texture looks better without sRGB checked. Which should I do?
Normal maps should always be linear space (sRGB off).
Gloss maps should generally be linear space off, but its not a big deal if you use sRGB/Gamma space.
Specular maps generally are Gamma space.
Diffuse/Albedo maps are generally gamma space.
Replies
Press the "ALT" key on your keyboard, wait for the popup telling you that toolbag 2 can't reach the server (you have to be offline). You can then release the key, but when you press "okay", Toolbag 2 will crash.
I have been able to reproduce this crash multiple times.
In the long term we plan to add functionality aimed at easier management of larger scenes/environments. I can't really give you any specific information about features or dates, but yes, there is some cool stuff that we plan on adding with future updates.
I have a little downtime at work so I felt compelled to ramble about this for a little;
Right now I'm starting to plan out my next full environment and of course part of that planning process is choosing which engine to present in. Immediately , as soon as I started planning I considered using TB2 because IMHO the shader itself is vastly better than anything out right now. Not only that, but with a PBR workflow, I feel like I am prepping myself for the future and possible future job opportunities. But upon further planning, I realized that at this point using TB2 is probably not worth what you'd be be giving up from other packages.
Here are some feature suggestions that would make TB2 irresistible when
it came down to chose what engine to present in:
+Global Illumination and Bounce Lighting Options
---I have no idea how to implement it- sorry. But It's hard to not use UDK and CE3 because of GI which can really tie an environment together
+Keyframable Object, Camera, and Particle animation (based on local pivot) and/or Vertex-based animation
---This one is big for me. I am always trying to stay ahead of the curve as a 3D Artist and one of the things I think we're going to see a lot of in 2014/15 is dynamic, moving environments. Either presented in HTML5 .gif form or high-res video, I think having a dynamic environment is going to be necessary when creating next gen visuals.
+Vertex Color/Texture Blending
---An amazing way to break up tiling
+Decal Placement
---Well, this one can be worked around pretty easy in TB2 with a plane and additive blending but it would be nice to have a more dynamic way of using decals
+Grid Snapping
---For modualar placement
+Fog
---For... making stuff foggy
That's all I really got off the top of my head. Sorry to mention it if any of this stuff is obvious; I'm just a 3D art nerd and I love discussing this stuff =P
TL;DR: If TB2 can implement some key features from other packages, it will be a no-brainer to use with every project; including large scale dynamic environments!
when rendering out a turntable the sky rotation is snapping back to default.
I can work around it by rotating the camera and mesh instead of the sky but its not as intuitive
This is a decent short list, I’ll add all of these to the suggestion list.
Yes, sorry this is a known issue, we just haven’t had time to fix it yet.
Local reflections are blurred based on the concept of the gloss map, the glossier your surface is, the less blurred the reflections will be. I think on a technical level the reflections are blurred a little extra to avoid ugly artifacts as well.
More shots coming soon !
Cheers
And also, latest wip funs!
love the feel of the fabric by the way mazz423
When I load certain textures, they are offset. For example, if I had a texture of a circle, I end up with a half circle on the left and right side of the texture, since it's tiling. It's not an isolated case.
In this case it's the gloss and metalness maps that gets offset.
If I load the diffuse map into either of those slots, it works.
If I load the gloss or metalness into the diffuse slot, they're offset..so it seems to be the texture itself as it loads that gets offset.
If i click the magnifying glass, the texture is offset. If I click the edit button, it shows up correctly.
All textures were exported straight from substance designer with no extra steps in between.
Didn't find anything on this problem when googling it..
It's because grayscale outputs from Designer are not read properly by Toolbag 1 and 2. The only solution for the moment is to plug a gradient map node just before your output node, to export an RGBA texture.
Our engine uses the gloss map to determine how much of the shadow is seen rather than how metal the object is. This seems more physically correct to me and is easier to understand when managing large groups of artists and getting consistency. It's much easier to say to someone that thing is 100% metal, but the metal is not pure chrome so the reflection is a bit glossy. Then the shader just makes the object receive less shadow as it gets closer to chrome/mirror in terms of smoothness. This also makes sense to me personally (not sure if it's scientifically correct) in the real world because my can of red bull in real life is 100% metal, but where the shadow is received it is not a perfect chrome reflection, it has imperfections and a brushed aluminum look.
Maybe someone could argue that there are other substances present in the metal can so it's not truly 100% made out of metal, but I think for game art this makes the most sense where you make the can 100% metalness and then adjust the gloss to get the material properties you are after.
I'd like to request that you guys actually update Marmoset to work this way as I think it's an easier pipeline to understand and author art for and should help maintain the coloured specular for metals while at the same time receiving shadows appropriately.
I ran into this thing the other day and wasn't sure if it was a feature or bug. Andy suggested that I post about it in here.
When changing saturation in post my dither material is washing out in a strange way very quickly. I assume it is something I am doing or maybe a combination of things.
glossiness doesn't in any way at all change the amount of shadow an object has. in every perceivable way that statement is incorrect.
what glossiness/roughness does is alter the microsurface of an object, changing how diffuse[i/] the light reflected is. it's important to note here that the word diffuse in this context has nothing to do with the texture map of that type. the more diffuse the light is, the more scattered it is. it might appear that there is very slightly less shadowing the rougher a surface is, but it's an illusion created by additional bounce light from the scattered reflections.
What we typically refer to shadows is actually an absence of light, specifically diffused light. Shadows are not “cast” or “added” or “multiplied”, even if this is what technically happens in 3D rendering.
As Lee points out, objects vary in the amount of light they diffuse, or absorb, and the amount they reflect. Raw metals do not absorb any light, they reflect 100% of light that hits them, so they have a diffuse value of zero (this can vary in real world materials depending on other factors, like oxidization, wear, etc, but this is the basic theory). With a metal material, or any material that has zero diffuse contribution, its impossible to get any darker than zero while “in shadow” as the diffuse contribution is already zero.
In reality, reflections are not shadowed per say, an object simply reflects the other object that happens to be blocking thing light. In this case some more advanced reflections would probably help but screen space reflections do a pretty good job.
This is a simple phenomenon to observe, go into a bathroom or anywhere that has overhead lighting, go up the sink (assuming the faucet is metal) and put your hand between the faucet and the overhead light. The faucet does not suddenly reflect less now that it is “in shadow”, it simply reflects your hand instead of the light. With very glossy materials, the size of that point light happens to be very small, so you will see very little difference in the overall reflection of the faucet (assuming there are more lights in the room, or lots of ambient light). The less glossy a surface is, the larger that point light source is going to be, and the more noticeable to the overall reflection is when its blocked/occluded by another, less bright object.
In practice, TB2 follows this behavior, with dynamic lights, the rougher your surface is, the wider the specular highlight is, and the specular highlight from dynamic lights actually is shadowed or blocked by other objects.
So here is a quick example. For the diving helmet asset I used the metalness thing, so the metallic areas have zero diffuse (side note: its important to mention again this isn't exclusive to the “metalness” method but applies to any material that has zero diffuse contribution). You can see that as shadows are enabled and disabled for the dynamic lights, that the specular reflection from said lights is clearly blocked/shadowed by the little pipe/valve extension thingy
As specular reflections are a view-dependent effect, it may appear that it is not shadowed, but it actually is. You just don’t see “diffuse” shadows, because there is no diffuse lighting with a diffuse value of 0.
Ambient specular reflection does not cast shadows, so there isn’t really any logical way to get shadow data. I guess real time ray traced radiosity would be the solution there, but that's obviously not something available in TB. Again, if you multiply a shadow onto the specular intensity, what you’re really doing is making the object less reflective, not receive less light, which doesn't make much sense.
Ok, so I hope that makes sense and I didn't say anything too stupid. I've argued with various people for the same feature ie: shadows that affect specular the same way they do diffuse, and every time I try, I fail to come up with a real-world situation where this is actually the case.
Here's our test case in Marmoset currently
1. Make your albedo white, your metalness 100% and gloss 100% smooth. That should be chrome/mirror, note the lack of shadows/blocked light/diffuse contribution, which ever you would like to call it. This looks correct to me.
2. Make your albedo white, your metalness 100% and gloss 50% smooth. That should be a simulation of a pretty rough metal. The part that doesn't look correct here to me is the lack of shadows/blocked light/diffuse contribution, which ever you would like to call it. Am I misunderstanding this, it appears in real life when I cast a shadow on to a diffused metal object it does appear to receive shadow.
3. Now in Marmoset I can lessen the metalness to an arbitrary amount so I have some diffuse contribution and start to see received shadows, but as I lessen the metalness I am also lessening the amount of speculour colour which starts to make my metals look less metal. I was surprised to find out in the engine we are working with right now that when you make the object 100% metalness you still see received shadows, we did some tests and found out that it is actually our smoothness map which lessens the diffuse contribution and eventually makes the shadows look like they are not there.
I can't actually wrap my head around what it's doing behind the scenes, but the reason I like this workflow is a can make something 100% metal then just worry about dialing in the smoothness and the amount of diffuse contribution is taken care automatically by the shader, also I maintain my coloured spec for rough metal surfaces. I'm a bit confused though because I'm pretty sure our metalness map works the same as Marmoset where it tints the albedo darker and tints the specular hotspots to be the albedo colour, but it looks like there is some additional functionality which is driven by the gloss map. What do you guys think? Or are we saying that most rough metals should not be 100% metalness?
I understand if you can't show anything due to NDA, but I feel like this would give me a better understanding of the differences.
In the image below, the hair looks great from the front but becomes completely transparent from behind (simple planes)
with Normal map and parallax i have not problems, but with the displacement, has been impossible.
Thanks.
-Reed-
about the microfiber diffusion and the peach fuzz... What does a map look like for this?
Is it possible to bring in the hdri's from Toolbag 1 into Toolbag2? If so how?
thanks
Reed
this is what i painted recently for peach fuzz.
it was a quick and dirty approach but i bet it could be done even better by some decent hair alpha or even going further and baking down for real hair/fur systems.
I was really wondering how it worked and couldn't find anything on the interwebs :P
it's a grayscale map, it just masks the fuzz on or off. you might consider using a very fine, almost blurred noise for things like cotton, and then multiply over the top with cavity map to reduce fuzz inside things like seams, or if the fibres are large enough to see them, then help define them a bit better.
Currently there is no way to do this, but I will add it to the suggestion list.
This is an interesting idea, but probably not feasible due to large differences in UE3/UDK lighting and material system and what TB2 does.
In addition to Froyoks suggestion, try resaving the texture in photoshop, a simple resave may work, if not, try re saving as RGB/24bit.
Interesting, Ive tried something similar out here but I dont think it has to do with the dithered blending. I think its the emissive slot that is causing the problem. Can you turn off emissive and see if you have the same problem?
To use displacement you need to add a subdivision type (usually flat) as well as displacement.
Displacement is a DX11 feature and will not work with older, DX10 video cards.
This is not at all a silly question.
First off, what the peach fuzz map is is simple a map of the intensity of the peach fuzz effect. Sort of like a specular map, darker values will give you less fuzz, and lighter values will give you more fuzz. This is very useful if you have say, cloth, skin, and rubber/metal/etc, you may want a certain value for skin, a certain value for cloth, and then black for other material types.
Now what the content actually looks like is up to you, you can use very simple solid colors, Lee painted in some individual hairs into a fuzz map that seemed to work, you might want a little bit of the cloth pattern overlaid on the fuzz map, really you should just experiment and see what works.
Sorry, this is not possible. The file format for TB1 skies is not supported in TB2 and we have no plans to support that format at this time (the data is completely different).
Unfortunately the source files for the skies that we created for Darkest of Days at 8monkey Labs have been lost so those skies can not be re-created in TB2.
There are a lot of freely available HDR sky panoramas out there on the internet these days though, and it is very easy to load standard .hdr files and such, and place sky lights with the new sky editor. Overall the sky system is much more powerful in TB2 so I would encourage you to try it out. The getting started tutorial goes over the new sky system:
http://www.marmoset.co/toolbag/learn/getting-started
The Problem is that everything in tb2 seems darker, Even the ui is hard to read, and not like the examples i saw in the tb2 tutorial. The captured images come out washed and light, no matter what i tried.
i am really liking the viewport renders of tb2, so its a shame i cant get them captured right.
Any tips ?, i was hoping someone already experienced it.
This is with default camera presets, sky settings all at 1 (brightness etc.), also selective brightness.
Yes, that works thanks for the tip
Ok thanks, looks like there may be a bug with saturation and emissive then, I’ll look into it.
This is very odd and certainly should not be happening. Could you please give me some more information:
What OS are you using?
What video card are you using?
Have you updated to the latest video drivers? If not, try that.
Do you have a custom color profile or anything like that? IE: spyder color correction software
Just curious as I have one and probably know this feature if plan is something at the bottom of the list.
But might be cool to move with the 3D mouses into tb2
Yeah we may be interested in this as well.
Put your mask in the alpha of your diffuse, then set the transparency module to dithered or cutout.
Dithered is for soft blending, so it should be perfect for hair. It may be a bit noisy in the viewport but will look better with exported images.
Cutout is basic alphatesting.
Also be sure to try the anisotropic reflection mode (or add aniso to the secondary spec). This works really well for hair, especially if your uvs for your hair all run in the same direction, then you can use the simple direction slider in the aniso shader. Otherwise you will need a direction map.
Sorry, no, there is not a feature to export meshes.
Unfortunately no this is not something we support, if there is a lot of user interest for support we will consider adding it, or maybe if someone sends us one.
is there a fix for that?
I tried it with exported meshes from Zbrush, Topogun and Blender.
no matter where I export the meshes, -tried with different ones- I can replicate this issue.
edit;
after a few experiments I found the missing setting in Blender to fix the visible seams on the mesh within Marmoset.
the option "Loop Normals" during OBJ export has to be activated.
apparently the name for the export function got changed from "Include Normals" to "Loop Normals" in one of the recent Developer Builds.
for now only Lowpolys exported from Zbrush will give me the visible seam issue.
So for your normal keep sRGB unchecked so it actually looks correct where it matters, the viewport.
Looks like you sorted it out. In general meshes from zbrush have a lot of issues with smoothing/normals etc because Zbrush’s exporter is really limited. Generally its best to run them through another 3d app beforehand.
Hehe yeah, thats similar to my laptop as well. Nice isn’t it?
Normal maps should always be linear space (sRGB off).
Gloss maps should generally be linear space off, but its not a big deal if you use sRGB/Gamma space.
Specular maps generally are Gamma space.
Diffuse/Albedo maps are generally gamma space.
Just butting in to make a couple of smaller feature requests, if that's okay
A save prompt when closing up + a 'File -> Open recent' would be spleeendid
Thanks!