Several people have requested this, how come? Is a sequence of images easier to work with than raw video somehow? The recording performance would not improve as TGA frames would be uncompressed bloat-mongers just like our AVI frames are now.
Also if it can save out images as slow as it wants there wouldn't be studdering (or as taxing on the hardware, it would just take some time for each frame to save out before moving onto the next one). As if it were taking time for each frame to save out and THEN move onto the next frame instead of trying to do them all at once. It doesn't matter the compression when we would be re-compressing it in a package like after effects to get a smooth turntable.
Kind of like rendering an animation in any 3rd party package. I feel like this might be easily done if you just take a part of the save image process that's already in Marmoset and allow you to output a range of frames with a range of mesh/sky/camera rotation.
Then again I'm not a programmer but I'd probably be able to whip up a quick mouse click + keyboard input macro to:
1) Set mesh rotation rate (or "Mesh Turntable" value)
2) Pause every .25 seconds or so (T Hotkey)
3) F12 to save out screenshot at desired resolution
4) Allow the screenshot save (F12) to complete it's operation before moving onto the next Pause/Unpause
5) Let it sit until all 300-800 frames or however long you wanted the turntable(s) to be are done saving out.
It'd take time for it to go through but the final output would be way better than trying to record it in real time like marmoset/fraps/camtasia do right now.
Several people have requested this, how come? Is a sequence of images easier to work with than raw video somehow? The recording performance would not improve as TGA frames would be uncompressed bloat-mongers just like our AVI frames are now.
I think performance wouldn't be an issue as long as you only wrote 1 image at a time and then proceeded onto the next only when the previous image had been created and saved. It wouldnt matter if it took 30 mins to render 30 frames as they could be stitched together in a sequence with a video editor at the correct speed later. It wouldnt (shouldnt?) be a real time capture.
Because each frame would be rendered out individually, it would be only be a matter of how long 500 or so frames would take to render, rather than a limitation of the speed of your hardware. Everyone would be guaranteed the same quality
The spinning monkey intro in my Blender MatCaps tutorial was rendered out using this method, as realtime just wasn't cutting the mustard (for the reasons you mentioned above.)
Having an option for setting the amount of FPS would be perfect (for example, a 5 second video length@30fps would need 150 images to be rendered or 75 images for 15fps ), as it would allow higher quality than realtime video capture allows and ultra slow motion stuff too. (though having a value that is tied to the turntable speed as well, would also be awesome)
Alpha in video is a great point, didn't think of that.
Fixed time-step recording would indeed be awesome. It'd be simple for turn-tables, just rotate by speed/30 every frame for 30hz. Arbitrary movement and mouse input would be trickier, the app would get laggy and unresponsive when recording but maybe that's ok?
Alpha in video is a great point, didn't think of that.
Fixed time-step recording would indeed be awesome. It'd be simple for turn-tables, just rotate by speed/30 every frame for 30hz. Arbitrary movement and mouse input would be trickier, the app would get laggy and unresponsive when recording but maybe that's ok?
Sounds great to me!
A pop up upon completion and a progress bar would be nice too, if its feasible.
Alpha in video is a great point, didn't think of that.
Fixed time-step recording would indeed be awesome. It'd be simple for turn-tables, just rotate by speed/30 every frame for 30hz. Arbitrary movement and mouse input would be trickier, the app would get laggy and unresponsive when recording but maybe that's ok?
Yes please! As long as I know that it's saving out the range of frames it'd be an awesome addition. Gotta finish my reel in a month
i found a bug in marmoset, i didn't try the latest so maybe its known and fixed already.
We are right now using marmoset for internal presentation to a client, before their engine is running and the tools are on a level that we can use them externally, however their pipeline demands to use PSD files für the textures, due to a few technical rules. Once i (and others in our team) load a psd in the diffuse slot, it seems to tile the diffusemap instead of using 0-1 UVspace it uses more. It seems like this behaviour is only happening for the diffuse tho, normal and spec works.
Hi guys, Marmoset is great, Ive actually rendered my entire reel (all the beauty shots) in it!
I agree rendering out image sequences (TGA or TIF) w/alpha channels in them would be an INCREDIBLE feature.
Secondly could you guys please make a check box to show NON-TRIANGULATED wire frames? Is that super hard? It would be awesome to show wire frame/edge flow without the triangulation.
Secondly could you guys please make a check box to show NON-TRIANGULATED wire frames? Is that super hard? It would be awesome to show wire frame/edge flow without the triangulation.
I'm not sure if they can do this considering the mesh is triangulated when you export it from a major package anyways. You could always just use a UV snapshot of your model and overlay it in your diffuse if you need to show your quads that bad.
Once i (and others in our team) load a psd in the diffuse slot, it seems to tile the diffusemap instead of using 0-1 UVspace it uses more. It seems like this behaviour is only happening for the diffuse tho, normal and spec works.
Neox, .psd files work for me in the diffuse slot. Just tried with 1.03 and it works there too.
I don't know if it makes a difference, but I'm using CS3.
Alpha in video is a great point, didn't think of that.
Fixed time-step recording would indeed be awesome. It'd be simple for turn-tables, just rotate by speed/30 every frame for 30hz. Arbitrary movement and mouse input would be trickier, the app would get laggy and unresponsive when recording but maybe that's ok?
just don't forget about outputting z depth as well, please. much easier to control that in a compositor when rendering out a reel than dealing with pre-blurred alpha masks from marmoset.
Loving this tool but there is something that is bothering me to do with mesh importing. I have to set up all my materials again if I import my FBX again from max. As I had just spent a bit of time setting up my materials and saved as a .mesh file it's all good, but I wanted to update the mesh again and it put the default material back on all my chunks.
It would be better if I could import a changed mesh without having to re-assign my material back to all my chunks again. Don't really want to collapse my object to a single mesh in max.
It would be good if it read in the diffuse map automatically too.
Loving this tool but there is something that is bothering me to do with mesh importing. I have to set up all my materials again if I import my FBX again from max. As I had just spent a bit of time setting up my materials and saved as a .mesh file it's all good, but I wanted to update the mesh again and it put the default material back on all my chunks.
It would be better if I could import a changed mesh without having to re-assign my material back to all my chunks again. Don't really want to collapse my object to a single mesh in max.
It would be good if it read in the diffuse map automatically too.
Be sure you are clicking the 'Save Mesh & Materials' button located on the chunks/materials menu palette. This is the only button that will save your material bindings for the model that is loaded. The 'Save Mesh' button on the main File menu palette only saves the Mesh import scale settings that you may set on the slider below.
In regards to the Diffuse Map concern, could you clarify what you are asking for? If you have enabled auto-update for textures in Toolbag, any changes to your diffuse texture map that is detected on disk should update immediately in your viewport.
Hi, I'm testing this software out for my studio and have a scene with 10 characters loaded in it. I noticed once you pass 17 chunks, there is way to select more chunks, making it impossible to add materials to more than 17 chunks. Is there a workaround I can use to this? I am finding the same limitation to the material list.I found that you can click and expand the list of chunks, but it still won't load all of them.
Thank you so much for your time!
just don't forget about outputting z depth as well, please. much easier to control that in a compositor when rendering out a reel than dealing with pre-blurred alpha masks from marmoset.
Has this new release fixed the issue of when lights are added to the scene it screws alphas? ie it highlights the entire polygon, transparts parts included. It's pretty poor that this happens since alphas area completely essiential techinque.
I don't see the issue with this tool costing 50 bucks. Should Crazy Bump be free too? I feel like the same people that are complaining about this costing 50 dollars are the same people complaining they got ripped off when they spent 99 cents on an iphone game and didn't like it. In the grand scheme of things if you use this software to display your work and it saves you time and makes your work look better, and you get a job from your portfolio, is 50 dollars really that big a deal? It's 50 bucks, that's less than the cost of one game. As mentioned previously if you are really opposed to paying for this than you can use a free alternatives, but why would you it's 50 bucks. Plus why wouldn't you want to support the people that make this software, maybe you think games should be free too.
All that being said I recently interviewed a character artist for a job we are offering and all his work was displayed in Marmoset, really great looking presentation. I think this is a great tool.
Be sure you are clicking the 'Save Mesh & Materials' button located on the chunks/materials menu palette. This is the only button that will save your material bindings for the model that is loaded. The 'Save Mesh' button on the main File menu palette only saves the Mesh import scale settings that you may set on the slider below.
In regards to the Diffuse Map concern, could you clarify what you are asking for? If you have enabled auto-update for textures in Toolbag, any changes to your diffuse texture map that is detected on disk should update immediately in your viewport.
But doesn't this then save the current model as a .mesh file? Meaning if I update the FBX from max, it doesn't get read in? Unless of course I delete the mesh and load in the FBX, but then that requires me to assign the materials again. Doh!
What I mean by auto loading in diffuse map is that when I open and new mesh, there is a default grey diffuse map assigned to it. Would it be possible for the correct diffuse map to be loaded in when you open a new mesh?
I know it's not much work for 1 texture, but if you have 100, that's quite alot of time consuming work
Not sure if this is the right place to post this but I give it a go.
I am trying out animation in Marmoset and so far it looks great. However, my character has a cape and for animation of the cape I am trying out APEX Clothing plugin. I am right now working in Max and exporting everything as FBX to Maya. I have managed to bake the APEX Clothing animation as a vertex animation to the main timeline. Can the stooge exporter export that kind of animation or are my only option to keyframe the cape with bones?
Not sure if this is the right place to post this but I give it a go.
I am trying out animation in Marmoset and so far it looks great. However, my character has a cape and for animation of the cape I am trying out APEX Clothing plugin. I am right now working in Max and exporting everything as FBX to Maya. I have managed to bake the APEX Clothing animation as a vertex animation to the main timeline. Can the stooge exporter export that kind of animation or are my only option to keyframe the cape with bones?
Unfortunately we only support skeletal animation right now.
Hi, I'm testing this software out for my studio and have a scene with 10 characters loaded in it. I noticed once you pass 17 chunks, there is way to select more chunks, making it impossible to add materials to more than 17 chunks. Is there a workaround I can use to this? I am finding the same limitation to the material list.I found that you can click and expand the list of chunks, but it still won't load all of them.
Thank you so much for your time!
Not that this excuses our ghetto, non-scrolling list boxes, but you can use the arrow keys to cycle through mesh chunks and materials for now. Click the list you want to cycle through to give it keyboard focus and try the up/down arrows.
Has this new release fixed the issue of when lights are added to the scene it screws alphas? ie it highlights the entire polygon, transparts parts included. It's pretty poor that this happens since alphas area completely essiential techinque.
We were not aware of this problem but it does seem like the direct lighting pass is ignoring texture alphas, thanks for the report. Alpha testing and alpha-to-coverage still work properly under dynamic lights, which may provide you with a suitable work-around for now.
Secondly could you guys please make a check box to show NON-TRIANGULATED wire frames? Is that super hard? It would be awesome to show wire frame/edge flow without the triangulation.
This would be awesome for presentation purposes as having a triangulated wireframe doesn't look pretty, however, I believe marmoset (like most game engines) automatically triangulates meshes when imported. Correct me if I'm wrong though...
Love the tool bag. I bought it a few months ago, it is awesome. Thanks guys
I had a question though, which maybe really obvious. Is there a way to render out of marmoset with alpha?
For example, if I wanted to render a character and not have a background attached. I know I can change the background color and all that good stuff but for formatting's sake it would make it easier for me to do it in Photoshop.
Love the tool bag. I bought it a few months ago, it is awesome. Thanks guys
I had a question though, which maybe really obvious. Is there a way to render out of marmoset with alpha?
For example, if I wanted to render a character and not have a background attached. I know I can change the background color and all that good stuff but for formatting's sake it would make it easier for me to do it in Photoshop.
If you save a screenshot as a .TGA, they are saved with an alpha channel. Just open it up in Photoshop and view all of the channels.
There appears to be a bug with the ambient occlusion. If I run the toolbag in a maximised window I get a noticeable darkening of the AO and allot of artifacts
(running on a Geforce GT 330m 512mb). First pic is maximised, second pic is a non maximised window.
Just wondering if Marmoset supports a second material id.
So for instance, In Max if I had a building that was one mesh but had several tiling materials (say 3 material IDs in a multi-material), could I replicate this set up in Marmoset?
Just wondering if Marmoset supports a second material id.
So for instance, In Max if I had a building that was one mesh but had several tiling materials (say 3 material IDs in a multi-material), could I replicate this set up in Marmoset?
Hopefully that makes sense!
If you break the mesh into chunks (1 mesh with multiple, separate parts) then you can have as many materials as you have chunks.
Thanks. So in Max speak, if I have one object that is comprised out of several elements, I can assign different materials to it in Marmoset? Is that right?
That sounds like a good work around, but for the sake of vertex count, is it possible to use seperate materials IDs in Marmoset?
I guess it doesn't really matter, as you use Marmoset for presentation rather than importing it full engine, where vertex count would matter. Still I'd be intersted to know, so it'll save me a small amount of work!
Neox, .psd files work for me in the diffuse slot. Just tried with 1.03 and it works there too.
I don't know if it makes a difference, but I'm using CS3.
i have to test more whats the difference between my psd files, basicle the head gets strangely tiled while the body and all armor/hair/addon parts work like a charm
i have to test more whats the difference between my psd files, basicle the head gets strangely tiled while the body and all armor/hair/addon parts work like a charm
You have found a new PSD bug, congrats! Some PSDs without alpha channels are loaded incorrectly and do a weird grayscale tiling of what looks like separated color components. The work-around for now is to always add an alpha channel to the PSD.
EDIT: This was actually a byte-alignment problem with all our non-power-of-2 textures and has been addressed for Toolbag 1.04. Current work-around: change your image dimensions slightly (or add an alpha channel, fixing it in a round-about way).
When using marmoset to view models with textures with normals/spec, is It best to add lights to the scene or use the starting skylight to make all your texture intensity from, then add lights?
Help is appreciated If anyone understands me, sry if it is already mentioned, time constraint.
When using marmoset to view models with textures with normals/spec, is It best to add lights to the scene or use the starting skylight to make all your texture intensity from, then add lights?
Help is appreciated If anyone understands me, sry if it is already mentioned, time constraint.
I usually use one of the skylights and then add a couple of spotlights later, but its all down to taste really
Sorry if someone has asked about this before, but when i open the color selector (for example light color), I am not able to change it in the swatch. When i click, it opens up the 8monkeylabs.com home page. Any clue?
Hey guys, I'm having a weird issue with shadows. Didn't find an answer searching through either thread but I might have missed it.
I'm getting weird polygon shading when shadows are turned on, its only on certain edges and I'm not sure whats causing it (happens whether normal maps are applied or not). Scaled down the mesh to human size, tried switching FOV, Tried different OBJ export options in max, but it doesn't seem to be doing anything. The shading shifts with the camera angle, so sometimes its not visible, sometimes it is. Any clues as to whats causing this or how to fix it?
toolbag-developers: just wondering if you have a public bug/featuretracker for your app somewhere? for example something along the lines of http://code.google.com/p/movist/issues/list
if not, wouldn't it be worth setting one up on your site? much easier to follow up on than a forum thread surely.
I'm having issues with my normal map showing up in marmoset and one of the arms of my character looks off, inverted or something even with culling turned off.
multiple maps were made in xNormal (3), combined in photoshop. everything looks fine in maya 2011 but when I bring it into marmoset i have issue. When I exported from my mesh from maya i made sure that normals were tuned on when exported. I've tried inverting the green channel in marmoset and inverting it within photoshop as well and still no go.
I'm not sure of the coordinance to be honest, whatever the default is in xNormal
here's an image of the normals, let me know if you want me to put up the actual .psd file somewhere if need be. The only thing I can think of is maybe my alpha channel is screwed up, I disabled the alpha at one point trying to see fit hat was it and i got this weird looking wood/speckle texture on it. anyway here's the normals
yup, geometry looks inverted, possibly among other things. have you run the maya commands for unifying/fixing normals and checked that they are really all pointing outward - have you frozen your mesh afterwards?
also, you might try triangulating the mesh you export. that has fixed a great many issues with vertex normals for me (coming from max though).
lastly, try exporting via FBX if you haven't done so already and see what that does. if you're using some old maya version, make sure you install halfway recent FBX plugins (2010 ones do the job for me).
thanks thomasp, I'll try your suggestions, i figured the normals ont he arm might be flipped it's just a dupe of the other arm that doesn't have issues. I did freeze my mesh as soon as I started the uv process. if i triangulate the mesh will it effect my uvs at all? I'll back up the file and give that a try too if the other things don't help.
Replies
Also if it can save out images as slow as it wants there wouldn't be studdering (or as taxing on the hardware, it would just take some time for each frame to save out before moving onto the next one). As if it were taking time for each frame to save out and THEN move onto the next frame instead of trying to do them all at once. It doesn't matter the compression when we would be re-compressing it in a package like after effects to get a smooth turntable.
Kind of like rendering an animation in any 3rd party package. I feel like this might be easily done if you just take a part of the save image process that's already in Marmoset and allow you to output a range of frames with a range of mesh/sky/camera rotation.
Then again I'm not a programmer but I'd probably be able to whip up a quick mouse click + keyboard input macro to:
1) Set mesh rotation rate (or "Mesh Turntable" value)
2) Pause every .25 seconds or so (T Hotkey)
3) F12 to save out screenshot at desired resolution
4) Allow the screenshot save (F12) to complete it's operation before moving onto the next Pause/Unpause
5) Let it sit until all 300-800 frames or however long you wanted the turntable(s) to be are done saving out.
It'd take time for it to go through but the final output would be way better than trying to record it in real time like marmoset/fraps/camtasia do right now.
I think performance wouldn't be an issue as long as you only wrote 1 image at a time and then proceeded onto the next only when the previous image had been created and saved. It wouldnt matter if it took 30 mins to render 30 frames as they could be stitched together in a sequence with a video editor at the correct speed later. It wouldnt (shouldnt?) be a real time capture.
Because each frame would be rendered out individually, it would be only be a matter of how long 500 or so frames would take to render, rather than a limitation of the speed of your hardware. Everyone would be guaranteed the same quality
The spinning monkey intro in my Blender MatCaps tutorial was rendered out using this method, as realtime just wasn't cutting the mustard (for the reasons you mentioned above.)
Having an option for setting the amount of FPS would be perfect (for example, a 5 second video length@30fps would need 150 images to be rendered or 75 images for 15fps ), as it would allow higher quality than realtime video capture allows and ultra slow motion stuff too. (though having a value that is tied to the turntable speed as well, would also be awesome)
Fixed time-step recording would indeed be awesome. It'd be simple for turn-tables, just rotate by speed/30 every frame for 30hz. Arbitrary movement and mouse input would be trickier, the app would get laggy and unresponsive when recording but maybe that's ok?
Sounds great to me!
A pop up upon completion and a progress bar would be nice too, if its feasible.
Thanks
Yes please! As long as I know that it's saving out the range of frames it'd be an awesome addition. Gotta finish my reel in a month
We are right now using marmoset for internal presentation to a client, before their engine is running and the tools are on a level that we can use them externally, however their pipeline demands to use PSD files für the textures, due to a few technical rules. Once i (and others in our team) load a psd in the diffuse slot, it seems to tile the diffusemap instead of using 0-1 UVspace it uses more. It seems like this behaviour is only happening for the diffuse tho, normal and spec works.
I agree rendering out image sequences (TGA or TIF) w/alpha channels in them would be an INCREDIBLE feature.
Secondly could you guys please make a check box to show NON-TRIANGULATED wire frames? Is that super hard? It would be awesome to show wire frame/edge flow without the triangulation.
I'm not sure if they can do this considering the mesh is triangulated when you export it from a major package anyways. You could always just use a UV snapshot of your model and overlay it in your diffuse if you need to show your quads that bad.
Neox, .psd files work for me in the diffuse slot. Just tried with 1.03 and it works there too.
I don't know if it makes a difference, but I'm using CS3.
just don't forget about outputting z depth as well, please. much easier to control that in a compositor when rendering out a reel than dealing with pre-blurred alpha masks from marmoset.
It would be better if I could import a changed mesh without having to re-assign my material back to all my chunks again. Don't really want to collapse my object to a single mesh in max.
It would be good if it read in the diffuse map automatically too.
Be sure you are clicking the 'Save Mesh & Materials' button located on the chunks/materials menu palette. This is the only button that will save your material bindings for the model that is loaded. The 'Save Mesh' button on the main File menu palette only saves the Mesh import scale settings that you may set on the slider below.
In regards to the Diffuse Map concern, could you clarify what you are asking for? If you have enabled auto-update for textures in Toolbag, any changes to your diffuse texture map that is detected on disk should update immediately in your viewport.
Thank you so much for your time!
Z Depth would be great
All that being said I recently interviewed a character artist for a job we are offering and all his work was displayed in Marmoset, really great looking presentation. I think this is a great tool.
But doesn't this then save the current model as a .mesh file? Meaning if I update the FBX from max, it doesn't get read in? Unless of course I delete the mesh and load in the FBX, but then that requires me to assign the materials again. Doh!
What I mean by auto loading in diffuse map is that when I open and new mesh, there is a default grey diffuse map assigned to it. Would it be possible for the correct diffuse map to be loaded in when you open a new mesh?
I know it's not much work for 1 texture, but if you have 100, that's quite alot of time consuming work
15% off this week
I am trying out animation in Marmoset and so far it looks great. However, my character has a cape and for animation of the cape I am trying out APEX Clothing plugin. I am right now working in Max and exporting everything as FBX to Maya. I have managed to bake the APEX Clothing animation as a vertex animation to the main timeline. Can the stooge exporter export that kind of animation or are my only option to keyframe the cape with bones?
Unfortunately we only support skeletal animation right now.
Not that this excuses our ghetto, non-scrolling list boxes, but you can use the arrow keys to cycle through mesh chunks and materials for now. Click the list you want to cycle through to give it keyboard focus and try the up/down arrows.
We were not aware of this problem but it does seem like the direct lighting pass is ignoring texture alphas, thanks for the report. Alpha testing and alpha-to-coverage still work properly under dynamic lights, which may provide you with a suitable work-around for now.
This would be awesome for presentation purposes as having a triangulated wireframe doesn't look pretty, however, I believe marmoset (like most game engines) automatically triangulates meshes when imported. Correct me if I'm wrong though...
I had a question though, which maybe really obvious. Is there a way to render out of marmoset with alpha?
For example, if I wanted to render a character and not have a background attached. I know I can change the background color and all that good stuff but for formatting's sake it would make it easier for me to do it in Photoshop.
If you save a screenshot as a .TGA, they are saved with an alpha channel. Just open it up in Photoshop and view all of the channels.
I went back and checked the last batch of renders I did. Some of the angles have alpha channels and some don't.
(running on a Geforce GT 330m 512mb). First pic is maximised, second pic is a non maximised window.
So for instance, In Max if I had a building that was one mesh but had several tiling materials (say 3 material IDs in a multi-material), could I replicate this set up in Marmoset?
Hopefully that makes sense!
That sounds like a good work around, but for the sake of vertex count, is it possible to use seperate materials IDs in Marmoset?
I guess it doesn't really matter, as you use Marmoset for presentation rather than importing it full engine, where vertex count would matter. Still I'd be intersted to know, so it'll save me a small amount of work!
I dont think its possible any other way though.
BTW, I love the work on the fountain, I was fallowing your progress over at Eat3D!
Thanks for the kind words
I plan on rebaking and retexturing it as soon as i get time
i have to test more whats the difference between my psd files, basicle the head gets strangely tiled while the body and all armor/hair/addon parts work like a charm
You have found a new PSD bug, congrats! Some PSDs without alpha channels are loaded incorrectly and do a weird grayscale tiling of what looks like separated color components. The work-around for now is to always add an alpha channel to the PSD.
EDIT: This was actually a byte-alignment problem with all our non-power-of-2 textures and has been addressed for Toolbag 1.04. Current work-around: change your image dimensions slightly (or add an alpha channel, fixing it in a round-about way).
Help is appreciated If anyone understands me, sry if it is already mentioned, time constraint.
I usually use one of the skylights and then add a couple of spotlights later, but its all down to taste really
I'm getting weird polygon shading when shadows are turned on, its only on certain edges and I'm not sure whats causing it (happens whether normal maps are applied or not). Scaled down the mesh to human size, tried switching FOV, Tried different OBJ export options in max, but it doesn't seem to be doing anything. The shading shifts with the camera angle, so sometimes its not visible, sometimes it is. Any clues as to whats causing this or how to fix it?
if not, wouldn't it be worth setting one up on your site? much easier to follow up on than a forum thread surely.
multiple maps were made in xNormal (3), combined in photoshop. everything looks fine in maya 2011 but when I bring it into marmoset i have issue. When I exported from my mesh from maya i made sure that normals were tuned on when exported. I've tried inverting the green channel in marmoset and inverting it within photoshop as well and still no go.
examples.
when I first import the mesh:
once the normal map is added (.psd)
color map:
Can you post the normal map please?
here's an image of the normals, let me know if you want me to put up the actual .psd file somewhere if need be. The only thing I can think of is maybe my alpha channel is screwed up, I disabled the alpha at one point trying to see fit hat was it and i got this weird looking wood/speckle texture on it. anyway here's the normals
yup, geometry looks inverted, possibly among other things. have you run the maya commands for unifying/fixing normals and checked that they are really all pointing outward - have you frozen your mesh afterwards?
also, you might try triangulating the mesh you export. that has fixed a great many issues with vertex normals for me (coming from max though).
lastly, try exporting via FBX if you haven't done so already and see what that does. if you're using some old maya version, make sure you install halfway recent FBX plugins (2010 ones do the job for me).
so in maya i should set all values to 1? right now only x is in 1 the others are at 0