I'd like some more information about the way marmoset is interpreting the specular intensity and sharpness from textures. although it may be gamma corrected properly grey values in these textures does not seem to translate to any intuitive mid point in relation to the interface sliders. for example - grey in the rgb of the spec texture seems to translate to a slider value of about 0.3 and the sharpness seems even crazier.
I'd like some more information about the way marmoset is interpreting the specular intensity and sharpness from textures. although it may be gamma corrected properly grey values in these textures does not seem to translate to any intuitive mid point in relation to the interface sliders. for example - grey in the rgb of the spec texture seems to translate to a slider value of about 0.3 and the sharpness seems even crazier.
The specular map is gamma corrected, which means the true 50% gray is around 0.3 as you suspect.
The gloss map is more complicated. The sharpness/gloss setting sets the max value, which relates to the max resolution the material will use for gloss. The shader uses 256, 128, 64, 32, 16, 8 etc resolution cube maps which are progressively blured to create the blurry reflections.
The problem with it is that, the difference between 8 and 32 is a lot more noticeable than 128 and 256. So the slider is weighted to combat that, its a purely subjective thing. The slider has a curve on it to try to give you an even change while adjusting the setting.
The gloss map itself is not gamma corrected, which means an RGB value of .5 and an sharpness value of 256 will give you a sharpness value of 128, which in turn would be equivalent to about 2/3rds on the slider.
So while you're authoring your gloss map content, you don't get benifit of the weighted slider compensating for the perceived difference, so you just have to sort of eyeball it. This is one of the reasons I prefer to block out my diffuse, spec and gloss values with basic flat color values before I start doing any detailing work for my materials.
Hey, first of all let me say that I'm really loving Marmoset so far, it's great to have a tool that makes it so easy to display things in real time. Now, onto my question- is there anything I can do about loss of texture quality due to mipmapping? I've imported multiple meshes, but some are larger than others causing a visible disparity in resolution between them. Here's a screenshot to illustrate what I mean-
The texture on the table is crisp, but the textures on the cards are somewhat blurry. Using an increased texture resolution on the cards doesn't help, nor does using the 'enlargement' setting for the screenshot, since the textures are automatically blurred in Marmoset due to what I assume is mipmapping. (i.e., when I zoom in on the cards in Marmoset the textures display at a higher resolution) I haven't found a way to change or disable settings for this kind of thing, so is it something I just have to deal with? Maybe it's kind of a nitpicky thing, I just want to make sure that my assets look as nice as they can. Thanks!
Chandler: Sorry but we currently don't have any way to adjust mip-mapping bias, though this is something we plan on looking at in the future.
You may be able to force some mip-mapping related options in your video card drivers, but that would affect everything in your scene. Maybe try cranking up the Anisotropic filtering and see if that helps?
Also, do you happen to have an AMD card? When we save out high res screenshots this should be corrected for on nvidia cards, but it doesn't work right with AMD cards.
In other news, we're looking for adventurous folks who have experience with unity to test some Toolbag -> Unity tools we're working on. Check out this thread if you're interested.
Ah, okay. I do indeed have an AMD card, which was a bad choice considering it doesn't play well with Maya either. Thankfully this is a fairly minor problem, everything else has been working fine for me. Thanks for the response!
How do I have to set up a material to make it like glass? I tryed with a grey alpha on the part that I want transparent but it didn't work, the part with grey alpha is all dotted...
Is there something else I need to do? how can Ido it? someone could explain it to me plese? I really don't know what to do!
Thanks!!
I'm working on some industrial CAD models, but I'm not particularly a CAD modeler so I'd love to try out the Marmoset Toolbag instead of Keyshot which I've been using.
I've been able to export from Alibre to an STL, then use Meshlab to turn the STL into an OBJ, and import that OBJ into the Toolbag. My problem now is that the UVs don't seem to like Toolbag, they just come out a single colour. Since Keyshot is built for this kind of model, I presume it's settings allow the application of a normal, bump etc. map overtop of a UV that is I'm guessing collapsed at 0,0?
Is there a workaround for this OTHER than unwrapping the objects manually in something like 3DS Max? Or does this problem rule out Marmoset for CAD renderings?
How do I have to set up a material to make it like glass? I tryed with a grey alpha on the part that I want transparent but it didn't work, the part with grey alpha is all dotted...
Is there something else I need to do? how can Ido it? someone could explain it to me plese? I really don't know what to do!
Thanks!!
The best way to do glass is with the ComplexRefraction Channel Model(very top of the material tab). There isn't much documentation on this feature, but this is the basics of how to use it:
A. You need a 32 bit diffuse texture, the RGB defines the color like a standard material, and the alpha channel is a mask to mask the refractive effect(ie for dust or non-glass parts of your material.
B. You need a 32 bit specular channel, the RGB defines the color like a standard material, and the alpha channel is a mask for the transparency of the material(ie: how transperent or opaque the glass is).
C. The normal map takes a standard normal map.
D. There is no way to use a gloss map, so you will have to set the gloss value with the sharpness slider
E. Refraction strength adjusts how visible the refraction effect is, though this can look pretty bad if you crank it up too high.
I'm working on some industrial CAD models, but I'm not particularly a CAD modeler so I'd love to try out the Marmoset Toolbag instead of Keyshot which I've been using.
I've been able to export from Alibre to an STL, then use Meshlab to turn the STL into an OBJ, and import that OBJ into the Toolbag. My problem now is that the UVs don't seem to like Toolbag, they just come out a single colour. Since Keyshot is built for this kind of model, I presume it's settings allow the application of a normal, bump etc. map overtop of a UV that is I'm guessing collapsed at 0,0?
Is there a workaround for this OTHER than unwrapping the objects manually in something like 3DS Max? Or does this problem rule out Marmoset for CAD renderings?
You will need a proper uv layout to load any sort of textures onto your mesh. Toolbag does not currently support any sort of procedural texturing that would work without UVs, sorry.
If your mesh does have uvs of some sort, it sounds like they are getting lost in file conversion. Any chance you could post some images of what your uvs look like?
You will need a proper uv layout to load any sort of textures onto your mesh. Toolbag does not currently support any sort of procedural texturing that would work without UVs, sorry.
If your mesh does have uvs of some sort, it sounds like they are getting lost in file conversion. Any chance you could post some images of what your uvs look like?
This is what I've done so far.
The pieces seem to have UVs since the materials use texture maps for bump, normals, opacity etc. I suppose it is simply being lost in the the transfers. Their fairly simple objects that I could model in max, that's just not available at my office.
I'm trying to create a light beam effect, but the alphas don't seem to be working right when there's a mesh behind it. What am I doing wrong? I'm using Alpha blend mode with Phong Env channel
The pieces seem to have UVs since the materials use texture maps for bump, normals, opacity etc. I suppose it is simply being lost in the the transfers. Their fairly simple objects that I could model in max, that's just not available at my office.
Can you show me what the UVs look like? Just a screenshot of what the OBJ or whatever file you're trying to load into Toolbag look like. Though if you do not have access to max or a different 3d modeling program that you can do uvs in, and uvs aren't converting from your CAD files, there isn't much you can do with your model in Toolbag.
In future versions we plan on putting in some more support for showing off high poly models, ie: not requiring uvs, but even then it will likely only be simple color values for the materials, not image maps etc, as there isn't really any way to apply textures without uvs.
You could play around with doing some very simply uvs for your mesh and apply basic flat color textures to get different material looks.
I'm trying to create a light beam effect, but the alphas don't seem to be working right when there's a mesh behind it. What am I doing wrong? I'm using Alpha blend mode with Phong Env channel
This is typical alpha sorting issues, I could give you some tips here but I think the better thing to do would be to add in two spotlights, place them in the correct position, turn on the volumetric haze option and tweak the light angle and sharpness to get the desired effect. This way the it will actually cast light as well.
Can you show me what the UVs look like? Just a screenshot of what the OBJ or whatever file you're trying to load into Toolbag look like. Though if you do not have access to max or a different 3d modeling program that you can do uvs in, and uvs aren't converting from your CAD files, there isn't much you can do with your model in Toolbag.
In future versions we plan on putting in some more support for showing off high poly models, ie: not requiring uvs, but even then it will likely only be simple color values for the materials, not image maps etc, as there isn't really any way to apply textures without uvs.
You could play around with doing some very simply uvs for your mesh and apply basic flat color textures to get different material looks.
I think the Keyshot rendering program is faking UVs on the objects. When I dig in to its materials I have the option to texture it by UVs which ends up with pretty much the same result as Toolbag. The option that adds the texture is called Cube Map, I'm not sure if that means it's projecting a mapped cube onto the object (if that's a thing) or unwrapping it into a cube UV... or something.
Either way it seems to be doing it in Keyshot. I COULD easily model these in max in my spare time... but I'm not sure how worth it that is, heh. Perhaps I'll just make do with Keyshot for now
It seems like Alpha-to-Coverage is not working with screenshot output in version 1.08
It appears to be working here for me, can you be more specific about the problem you are having?
Is it not showing up in the alpha channel at all for you?
Is something showing up, but the content is wrong? (If so can you post some images?)
It appears to be working here for me, can you be more specific about the problem you are having?
Is it not showing up in the alpha channel at all for you?
Is something showing up, but the content is wrong? (If so can you post some images?)
I figured out the problem. High Dynamic Range was enabled. It's looking fine without it
I figured out the problem. High Dynamic Range was enabled. It's looking fine without it
Ok cool, yeah the HDR format doesn't support alpha channels. You can get around this by saving and HDR file, and then saving a second shot from the same angle with HDR turned off, and grabbing the alpha channel from the LDR image.
Hello, I need to improve the quality of the shadows and I read that using "set render.shadows.res 4096" improves them, but its not working for me, it says invalid variable.
Hello, I need to improve the quality of the shadows and I read that using "set render.shadows.res 4096" improves them, but its not working for me, it says invalid variable.
Can someone help me? Thank you.
I believe this was a setting in older versions. From what I understand, the latest version of toolbag sets the quality setting to the max your video card can handle, so unfortunately there is no way to increase the quality of the shadows further, sorry.
I believe this was a setting in older versions. From what I understand, the latest version of toolbag sets the quality setting to the max your video card can handle, so unfortunately there is no way to increase the quality of the shadows further, sorry.
But it seems the resolution its too low and I have a high end pc with a dual 480gtx, so my graphic card can support that, I have to modify values on the nvidia panel or something?
But it seems the resolution its too low and I have a high end pc with a dual 480gtx, so my graphic card can support that, I have to modify values on the nvidia panel or something?
Thank you very much for the fast response
Try changing your mesh scale around a bit, try exporting a copy scaled 10x and one scaled 0.1x and see if that makes any difference. Scale used to affect shadow quality but I'm not sure if it does anymore(and I'm on a laptop here with no toolbag to test).
Also, what sort of lights are you using? Sky light or spotlights? Can you post some images of your problem too?
Try changing your mesh scale around a bit, try exporting a copy scaled 10x and one scaled 0.1x and see if that makes any difference. Scale used to affect shadow quality but I'm not sure if it does anymore(and I'm on a laptop here with no toolbag to test).
Also, what sort of lights are you using? Sky light or spotlights? Can you post some images of your problem too?
Scaling it dosnt seem to affect, its scaled based on the reference human model, im using sky light, but using spot lights dosnt fix it either, maybe a little, but you still can see it jagged.
here is two images:
normal and normal with one spotlight.
The main problem is in the alphas:
The green light is the spotlight, marmoset cant render alpha shadows?
Am I doing something wrong? Thank you for your help.
Yeah I'm not sure there is anything you can do about the jagged shadows in the first couple images, though I will ask our programmers. I don't think its anything you are doing wrong though, just a limitation of the tool.
The third image, perhaps alpha shadows are not working with dynamic lights, which wouldn't surprise me as there are a few other issues with dynamic lights. I will take a look at this tomorrow and get in a proper bug report if I can't get it to work.
I found out that using in the material editor the option of alpha-testing with a threshold of more than 0.01 starts to show the alpha shadows, but the resolution problem is still there !
So maybe knowing a method to increase the shadow resolution could help us there.
Dvonio: Looks like you fixed the alpha issue. Yeah, alpha test with some value higher than 0 is required to get alpha objects to cast shadows properly.
As far as the general shadow quality issue there isn't any way to improve it at the moment, sorry! However, in the long term the shadow system will likely be revised.
The best way to do glass is with the ComplexRefraction Channel Model(very top of the material tab). There isn't much documentation on this feature, but this is the basics of how to use it:
A. You need a 32 bit diffuse texture, the RGB defines the color like a standard material, and the alpha channel is a mask to mask the refractive effect(ie for dust or non-glass parts of your material.
B. You need a 32 bit specular channel, the RGB defines the color like a standard material, and the alpha channel is a mask for the transparency of the material(ie: how transperent or opaque the glass is).
C. The normal map takes a standard normal map.
D. There is no way to use a gloss map, so you will have to set the gloss value with the sharpness slider
E. Refraction strength adjusts how visible the refraction effect is, though this can look pretty bad if you crank it up too high.
I just upgraded my computer. So I will need a new fingerprint for it. However my question is, I previously had a DX 10 card which could not do tesslation and the skin shader was kind of bleh. Is there a big change with skin shader with the DX 11 cards in marmoset?
The fairy nvidia 2012 tech demo had fairly impressive tesslated hair and skin shaders. I guess this is beyond the current builds capabilities?
So, messing with the sky tool a bit. I'm using a regular image which is not panoramic. In the skytool, if I set it to open as a Panorama option, it looks fine. There's a seam, but it's manageable. However, whenever I save out the skybox, it seems to revert to the "best guess" option when I decide to open it in the actual scene.
I just upgraded my computer. So I will need a new fingerprint for it. However my question is, I previously had a DX 10 card which could not do tesslation and the skin shader was kind of bleh. Is there a big change with skin shader with the DX 11 cards in marmoset?
The fairy nvidia 2012 tech demo had fairly impressive tesslated hair and skin shaders. I guess this is beyond the current builds capabilities?
It really depends which version of the skin shader you're talking about. In early builds of toolbag the skin shader was very simple with no options. The current version of the skin shader has lots of a conigurable settings to load a depth map for translucency, adjust normal smoothing, etc, so you can really do some nice things with it, but you've gotta play with all of the settings. Unfortunately its not documented very well, but try it out and let me know if you have any questions and I'll do what I can to help. It should be the same with DX10 and DX11.
You can use displacement mapping, bake a high res(very smooth) hair down to simpler hair lowpoly hair to get tessellated smooth hair. Toolbag doesn't support tessellation just for smoothing though, so you have to have some source content, I'm not sure how difficult this would be to set up for hair, really depends on how you plan on doing it.
So, messing with the sky tool a bit. I'm using a regular image which is not panoramic. In the skytool, if I set it to open as a Panorama option, it looks fine. There's a seam, but it's manageable. However, whenever I save out the skybox, it seems to revert to the "best guess" option when I decide to open it in the actual scene.
"Best Guess" option:
"Panorama" option:
I have to say I'm not sure what your exact problem is here. You mention that it resets to "best guess" when you load it, but both of your screenshots are from the Sky Tool. After you process your image in the Sky Tool, you will save out a .env file, and then load that .env file in the Light tab in Toolbag.
If you're having issues with how your sky lighting is loading inside of Toolbag, please give me some more information and most some images of the sky in Toolbag as well.
I see that someone was talking about removing glossines channel from file to the interface. If someone is interested I created a script which are creating textures inside photoshop from group of layers. Also is copying group of layers with name "g" to the alpha channel.
hello I have the full license of Marmoset 1.08 and when i render out a snapshot it wont include the logo and it is checked.
Any ideas why this is happening?
hello I have the full license of Marmoset 1.08 and when i render out a snapshot it wont include the logo and it is checked.
Any ideas why this is happening?
What file type are you saving out? Does it happen when you try different file types?
To get around this:
1. Open your image in photoshop.
2. Go to: ..\MarmosetToolbag\base\interface\toolbag\
3. Open watermark.tga in photoshop
4. Copy and paste watermark.tga (after you've resized/edited/etc) onto your image.
5. Set the blend mode to screen, adjust the postion and opacity to your liking.
6. Optionally, invert the watermark and set the blend mode to multiply if you have a bright background.
There currently is no way to adjust the position of the watermark in toolbag, so this is a more flexible way to do it anyway.
@akacg, the toolbag logo only shows up some of the time for me, try changing (under Output tab) Enlargement value, it doesn't show up for me with 0x but does with 2x enlargement and above.
It did work for me when I changed the enlargement value but it was kinda small. But thanks a lot I did not know.
How can i easily merge multiple images together? For example the 360 version. My images saves out as a TGA and dont seperate from the background. Are they supposed to? when i open my images in photoshop the images are flattened.
Never mind I figured it out. But still having trouble getting the logo to show when I take a snap shot. But I do have the other way to go about that. Thanks everyone
When I rotate the camera in the viewport the lighting changes on the model. Is this normal? I assumed the direction the lighting hits the model would only be affected if I rotate the mesh, not the camera. Notice how the lighting becomes darker on the front of the model as I rotate the camera.
Also - The center of the model always seems to receive more light that the outer areas. The only way I can figure out how to fix this issue is to add extra lights alongside the model and that doesn't work perfectly.
When I rotate the camera in the viewport the lighting changes on the model. Is this normal? I assumed the direction the lighting hits the model would only be affected if I rotate the mesh, not the camera. Notice how the lighting becomes darker on the front of the model as I rotate the camera.
Also - The center of the model always seems to receive more light that the outer areas. The only way I can figure out how to fix this issue is to add extra lights alongside the model and that doesn't work perfectly.
Am I misunderstanding something here?
A. The diffuse lighting should not rotate when you rotate the camera, unless your shift key is stuck down or something (shift+drag left mouse rotates the sky lighting).
B. Specular lighting will change as you rotate the camera, as it is a dynamic effect that is based on your camera's angle in relation to the mesh and light source.
C. I'm not seeing your problem with the center of your model receiving more light, it looks even to me on the side shot, and on the angled shot, the specular reflection is bouncing off the some of the details in the center more, but that is due to the angle of the light/detail there.
I've always had this larger problem with Marmoset. I asked once but never got a response, so I'm asking again.
When I use point lights in Marmoset, the edges of meshes within other meshes become visible. For instance, I can see the entire roundness of an eye within a head.
I have had this problem since Marmoset 1.4 (I believe that's when I purchased it). I've had two different graphics cards, both ATI cards. Right now my marmoset and ati card are up to date.
As I try to push myself, Marmoset is becoming unusable because of this. Any suggestions anyone?
ysalex have you tried scaling everything up by 100%? I just tried to recreate what you have there, and after a few zoom in and outs the z-fighting desists for me. And I have a dx11 geforce card so not the best to test against. Is it just eyes with multi layered alpha shells? Tried toning down the translucency of the skin?
----
Is there any possibility we could get alpha to coverage to scale up with the Enlargement on renders?
ysalex: can you send your scene+model+materials to: toolbag@marmoset.co, mention Joe/EQ said he was going to take a look at it so it will get forwarded to me.
I've seen similar issues but I think it may be a scale problem as firestarter suggests, so I would love to have a look at the files. You can also upload them somewhere (dropbox etc) and send me a link via PM and I will take a look.
firestarter: I'm not sure I understand your request about a2c, what is the problem you're having with it exactly? Could you provide some images too as well?
Ok, it's not a 'problem', it's more than likely my failing in understanding the capabilities of a2c.
So the image on the right is the 2x enlargement, it's basically scaling up the default size render for the alpha, rather than using a higher resolution. Supersampling does smooth it a touch of course. I left it off for the sake of this.
Ok, it's not a 'problem', it's more than likely my failing in understanding the capabilities of a2c.
So the image on the right is the 2x enlargement, it's basically scaling up the default size render for the alpha, rather than using a higher resolution. Supersampling does smooth it a touch of course. I left it off for the sake of this.
Thanks that makes sense, a quick word with one of our programmers and it looks like we're probably at the mercy of video drivers for this.
This is an unfortunate side effect of alpha-to-coverage not playing well with Toolbag's up-sampling. Toolbag uses interleaved up-sampling, meaning to get a 4x scale we move the camera by 4 sub-pixel increments and interleave the resulting images. But because the AtoC pattern shifts along with the camera, the pattern just gets big and gross.
In short: AtoC is a driver hack and there's not much we can do to help it :-/.
Replies
The specular map is gamma corrected, which means the true 50% gray is around 0.3 as you suspect.
The gloss map is more complicated. The sharpness/gloss setting sets the max value, which relates to the max resolution the material will use for gloss. The shader uses 256, 128, 64, 32, 16, 8 etc resolution cube maps which are progressively blured to create the blurry reflections.
The problem with it is that, the difference between 8 and 32 is a lot more noticeable than 128 and 256. So the slider is weighted to combat that, its a purely subjective thing. The slider has a curve on it to try to give you an even change while adjusting the setting.
The gloss map itself is not gamma corrected, which means an RGB value of .5 and an sharpness value of 256 will give you a sharpness value of 128, which in turn would be equivalent to about 2/3rds on the slider.
So while you're authoring your gloss map content, you don't get benifit of the weighted slider compensating for the perceived difference, so you just have to sort of eyeball it. This is one of the reasons I prefer to block out my diffuse, spec and gloss values with basic flat color values before I start doing any detailing work for my materials.
I hope this helps.
The texture on the table is crisp, but the textures on the cards are somewhat blurry. Using an increased texture resolution on the cards doesn't help, nor does using the 'enlargement' setting for the screenshot, since the textures are automatically blurred in Marmoset due to what I assume is mipmapping. (i.e., when I zoom in on the cards in Marmoset the textures display at a higher resolution) I haven't found a way to change or disable settings for this kind of thing, so is it something I just have to deal with? Maybe it's kind of a nitpicky thing, I just want to make sure that my assets look as nice as they can. Thanks!
You may be able to force some mip-mapping related options in your video card drivers, but that would affect everything in your scene. Maybe try cranking up the Anisotropic filtering and see if that helps?
Also, do you happen to have an AMD card? When we save out high res screenshots this should be corrected for on nvidia cards, but it doesn't work right with AMD cards.
In other news, we're looking for adventurous folks who have experience with unity to test some Toolbag -> Unity tools we're working on. Check out this thread if you're interested.
Is there something else I need to do? how can Ido it? someone could explain it to me plese? I really don't know what to do!
Thanks!!
I've been able to export from Alibre to an STL, then use Meshlab to turn the STL into an OBJ, and import that OBJ into the Toolbag. My problem now is that the UVs don't seem to like Toolbag, they just come out a single colour. Since Keyshot is built for this kind of model, I presume it's settings allow the application of a normal, bump etc. map overtop of a UV that is I'm guessing collapsed at 0,0?
Is there a workaround for this OTHER than unwrapping the objects manually in something like 3DS Max? Or does this problem rule out Marmoset for CAD renderings?
The best way to do glass is with the ComplexRefraction Channel Model(very top of the material tab). There isn't much documentation on this feature, but this is the basics of how to use it:
A. You need a 32 bit diffuse texture, the RGB defines the color like a standard material, and the alpha channel is a mask to mask the refractive effect(ie for dust or non-glass parts of your material.
B. You need a 32 bit specular channel, the RGB defines the color like a standard material, and the alpha channel is a mask for the transparency of the material(ie: how transperent or opaque the glass is).
C. The normal map takes a standard normal map.
D. There is no way to use a gloss map, so you will have to set the gloss value with the sharpness slider
E. Refraction strength adjusts how visible the refraction effect is, though this can look pretty bad if you crank it up too high.
You will need a proper uv layout to load any sort of textures onto your mesh. Toolbag does not currently support any sort of procedural texturing that would work without UVs, sorry.
If your mesh does have uvs of some sort, it sounds like they are getting lost in file conversion. Any chance you could post some images of what your uvs look like?
This is what I've done so far.
The pieces seem to have UVs since the materials use texture maps for bump, normals, opacity etc. I suppose it is simply being lost in the the transfers. Their fairly simple objects that I could model in max, that's just not available at my office.
Can you show me what the UVs look like? Just a screenshot of what the OBJ or whatever file you're trying to load into Toolbag look like. Though if you do not have access to max or a different 3d modeling program that you can do uvs in, and uvs aren't converting from your CAD files, there isn't much you can do with your model in Toolbag.
In future versions we plan on putting in some more support for showing off high poly models, ie: not requiring uvs, but even then it will likely only be simple color values for the materials, not image maps etc, as there isn't really any way to apply textures without uvs.
You could play around with doing some very simply uvs for your mesh and apply basic flat color textures to get different material looks.
This is typical alpha sorting issues, I could give you some tips here but I think the better thing to do would be to add in two spotlights, place them in the correct position, turn on the volumetric haze option and tweak the light angle and sharpness to get the desired effect. This way the it will actually cast light as well.
I think the Keyshot rendering program is faking UVs on the objects. When I dig in to its materials I have the option to texture it by UVs which ends up with pretty much the same result as Toolbag. The option that adds the texture is called Cube Map, I'm not sure if that means it's projecting a mapped cube onto the object (if that's a thing) or unwrapping it into a cube UV... or something.
Either way it seems to be doing it in Keyshot. I COULD easily model these in max in my spare time... but I'm not sure how worth it that is, heh. Perhaps I'll just make do with Keyshot for now
It appears to be working here for me, can you be more specific about the problem you are having?
Is it not showing up in the alpha channel at all for you?
Is something showing up, but the content is wrong? (If so can you post some images?)
I figured out the problem. High Dynamic Range was enabled. It's looking fine without it
Ok cool, yeah the HDR format doesn't support alpha channels. You can get around this by saving and HDR file, and then saving a second shot from the same angle with HDR turned off, and grabbing the alpha channel from the LDR image.
Can someone help me? Thank you.
I believe this was a setting in older versions. From what I understand, the latest version of toolbag sets the quality setting to the max your video card can handle, so unfortunately there is no way to increase the quality of the shadows further, sorry.
But it seems the resolution its too low and I have a high end pc with a dual 480gtx, so my graphic card can support that, I have to modify values on the nvidia panel or something?
Thank you very much for the fast response
Try changing your mesh scale around a bit, try exporting a copy scaled 10x and one scaled 0.1x and see if that makes any difference. Scale used to affect shadow quality but I'm not sure if it does anymore(and I'm on a laptop here with no toolbag to test).
Also, what sort of lights are you using? Sky light or spotlights? Can you post some images of your problem too?
There's always Blender? :P
Scaling it dosnt seem to affect, its scaled based on the reference human model, im using sky light, but using spot lights dosnt fix it either, maybe a little, but you still can see it jagged.
here is two images:
normal and normal with one spotlight.
The main problem is in the alphas:
The green light is the spotlight, marmoset cant render alpha shadows?
Am I doing something wrong? Thank you for your help.
The third image, perhaps alpha shadows are not working with dynamic lights, which wouldn't surprise me as there are a few other issues with dynamic lights. I will take a look at this tomorrow and get in a proper bug report if I can't get it to work.
EDIT:
I found out that using in the material editor the option of alpha-testing with a threshold of more than 0.01 starts to show the alpha shadows, but the resolution problem is still there !
So maybe knowing a method to increase the shadow resolution could help us there.
Thank you.
As far as the general shadow quality issue there isn't any way to improve it at the moment, sorry! However, in the long term the shadow system will likely be revised.
Thanks!! I'll try it!
The fairy nvidia 2012 tech demo had fairly impressive tesslated hair and skin shaders. I guess this is beyond the current builds capabilities?
"Best Guess" option:
"Panorama" option:
It really depends which version of the skin shader you're talking about. In early builds of toolbag the skin shader was very simple with no options. The current version of the skin shader has lots of a conigurable settings to load a depth map for translucency, adjust normal smoothing, etc, so you can really do some nice things with it, but you've gotta play with all of the settings. Unfortunately its not documented very well, but try it out and let me know if you have any questions and I'll do what I can to help. It should be the same with DX10 and DX11.
You can use displacement mapping, bake a high res(very smooth) hair down to simpler hair lowpoly hair to get tessellated smooth hair. Toolbag doesn't support tessellation just for smoothing though, so you have to have some source content, I'm not sure how difficult this would be to set up for hair, really depends on how you plan on doing it.
I have to say I'm not sure what your exact problem is here. You mention that it resets to "best guess" when you load it, but both of your screenshots are from the Sky Tool. After you process your image in the Sky Tool, you will save out a .env file, and then load that .env file in the Light tab in Toolbag.
The sky tutorial should explain how it all works: http://www.marmoset.co/toolbag/learn/skytool
If you're having issues with how your sky lighting is loading inside of Toolbag, please give me some more information and most some images of the sky in Toolbag as well.
link
Any ideas why this is happening?
What file type are you saving out? Does it happen when you try different file types?
To get around this:
1. Open your image in photoshop.
2. Go to: ..\MarmosetToolbag\base\interface\toolbag\
3. Open watermark.tga in photoshop
4. Copy and paste watermark.tga (after you've resized/edited/etc) onto your image.
5. Set the blend mode to screen, adjust the postion and opacity to your liking.
6. Optionally, invert the watermark and set the blend mode to multiply if you have a bright background.
There currently is no way to adjust the position of the watermark in toolbag, so this is a more flexible way to do it anyway.
I'm getting this problem when importing to Marmoset from Maya using the stooge plugin
But I am just outputting a snapshot with the regular settings. And it saves out as a tga. I dont know how to save it out differently...
How can i easily merge multiple images together? For example the 360 version. My images saves out as a TGA and dont seperate from the background. Are they supposed to? when i open my images in photoshop the images are flattened.
Also - The center of the model always seems to receive more light that the outer areas. The only way I can figure out how to fix this issue is to add extra lights alongside the model and that doesn't work perfectly.
Am I misunderstanding something here?
A. The diffuse lighting should not rotate when you rotate the camera, unless your shift key is stuck down or something (shift+drag left mouse rotates the sky lighting).
B. Specular lighting will change as you rotate the camera, as it is a dynamic effect that is based on your camera's angle in relation to the mesh and light source.
C. I'm not seeing your problem with the center of your model receiving more light, it looks even to me on the side shot, and on the angled shot, the specular reflection is bouncing off the some of the details in the center more, but that is due to the angle of the light/detail there.
When I use point lights in Marmoset, the edges of meshes within other meshes become visible. For instance, I can see the entire roundness of an eye within a head.
I have had this problem since Marmoset 1.4 (I believe that's when I purchased it). I've had two different graphics cards, both ATI cards. Right now my marmoset and ati card are up to date.
As I try to push myself, Marmoset is becoming unusable because of this. Any suggestions anyone?
----
Is there any possibility we could get alpha to coverage to scale up with the Enlargement on renders?
I've seen similar issues but I think it may be a scale problem as firestarter suggests, so I would love to have a look at the files. You can also upload them somewhere (dropbox etc) and send me a link via PM and I will take a look.
firestarter: I'm not sure I understand your request about a2c, what is the problem you're having with it exactly? Could you provide some images too as well?
So the image on the right is the 2x enlargement, it's basically scaling up the default size render for the alpha, rather than using a higher resolution. Supersampling does smooth it a touch of course. I left it off for the sake of this.
Thanks that makes sense, a quick word with one of our programmers and it looks like we're probably at the mercy of video drivers for this.
In short: AtoC is a driver hack and there's not much we can do to help it :-/.
Thanks for clearing it up =]