Btw Ged.. with a 8800GT and 512Mb just forget all what I told about VRAM.
Try to reinstall xNormal... perhaps the ui.png texture was corrupted for some reason and that's why you cannot see the UI well. Also sure you have the latest NVIDIA drivers... the 8800GT is relatively new card so I bet the drivers aren't very mature yet.
That banding usually appears because probably you need more samples per pixel. By default uses 16... try 64 or more. Also sure you disable the "adaptive AO" setting it to the same than the # of rays per pixel ( 64 in this case ... so use 64 rays per sample AND 64 also in the "Adaptive interval" )
Also increase a bit the AO bias... perhaps is a problem with self occlusion.
Also try to enable(or disable) the "weighting"... it can cause strange artifacts sometimes.
Caution with the 3dsmax OBJ exporter because messes the vertex normals. Better export it as SBM, 3DS, FBX, ASE, etc.
Have you tried to scale the mesh up and see if that helps? Perhaps is a floating point precision error due to small object radius.
Btw, if you can send me that mesh by email better... I'm curious how that result was produced hehe!
Recently I've been trying a new approach to occlusion rendering:
- Render the occlusion to high-poly vertex color.
- Transferring that to my low-poly texture.
Anyone using that tech? Pros / cons , anyone?
Is fine. Usually is very fast and gives good results to compute the per-vertex-AO with the Simple GPU AO tool... and then uncheck the "Ignore per-vertex-AO" so the xnormal software renderer could use the per-vertex-AO computed from the Simple GPU AO tool.
Btw, i'm working on the 3.14.6 with some UI improvements lilke Drag'n'drop, better remember folder system, multiple mesh erase, automatic save meshes reassignment and also experimenting with a new thing called "harden normals". Also gonna buy a graphic tablet to improve input support.
How do you alter the bias and rays in the baking options? I select 'Ambient occlusion' in the drop down list and below are the antialiasing options, but I can't see anything about rays or bias. The only place I've seen those are in the simple AO tool, which I'm not using (couldn't handle the mesh I was trying to put in).
Fraid the meshes are very big and would probably take ages to send, I'll see if I can manage but I'm fairly rushed already. Cheers.
How do you alter the bias and rays in the baking options? I select 'Ambient occlusion' in the drop down list and below are the antialiasing options, but I can't see anything about rays or bias.
Well, assuming you are using any 3.14.0 or above:
Select "Ambient occlusion" in the "Maps to render" group box. Then press the "..." button to display the AO options... Then change the "Bias" from the default 0,000010 to, for example, 0,01. Also disable the "weighting", play with the attenuation, increase # of rays, disable adaptive ao, play with spread angle, etc...
On the other hand the adaptive AA can generate some artifacts in the AO too... So try to set it to
min AA=1, max AA=4, thresholds all to 0.05. Use the diagnostics window to see how the samples are taken.
If nothing of this solves it try to send me a subdiv 3 or 4 highpoly model ( it will be compressed into 10Mb aprox ) + the lowpoly model so I can debug it pls.
I tried settings like you said, but now get worse results (grey texture with white outline of some parts and a red grid). I've tried changing everything with no luck, also tried putting results back to the way they were initially but that didn't seem to help. No idea what's going on. I sent you the high and low poly models I'm trying, hope you have better luck than me
thanks, that was my first thought so I flipped my low poly normals oever and over, but it turns out its actually my high poly normals that must have been flipped by the exporter. Still not sure if I can get this working for my character though. Also I cant edit the cage or see it in the viewer unless I re-export my model at a size about 50 times bigger than it is.
so Ive learned 2 things
1. make meshes very very big
2. check the normal direction
jogshy : Just wanted to say thanks for xNormal. It saved my bacon last night. XSI was having some difficulty with my maps and the AO for Ultimapper was really slow and doing a poor job after fiddling with the settings. But with some tweaking xNormal got the job done and way faster. You saved my bacon last night, that's for sure.
hi jogshy, I was wondering is the preview window size set? can it be made any bigger? I was considering taking dom war shots from it but cant really get the size I would like
edit: also, just checking is it limited to one light?
hi jogshy, I was wondering is the preview window size set? can it be made any bigger? I was considering taking dom war shots from it but cant really get the size I would like
edit: also, just checking is it limited to one light?
I'm making it a bit bigger for the 3.14.6... like 512x512 maximum. I don't wanna make it larger because the .NET's bitmaps are kinda slow.
About to take the screenshots from them.... better open the output file with paint.net or photoshop and then resize it using bicubic or bilinear filters.
And yep, the xNormal's 3d viewer is currently limited to one light due to shading instruction limits... I plan to change that for xn4 using deferred shading.
I'm keep being amazed how awesome this program is. I completly stopped using max for generating nm maps. Thanks a lot, it helped me in dom war ;-)
But I have one tiny problem.
Is there some link where I can download sbm export plugin? I have currently installed two versions of max, and instaler always install plugin for the other version than I want.
I need plugin for max9.
Thanks
Is there some link where I can download sbm export plugin? I have currently installed two versions of max, and installer always install plugin for the other version than I want.
I need plugin for max9.
Thanks
I can send it to you by email.
So is there a bug? I have tried with max 8, max9 and max2k8(x86/x64) and seems to detect them ok. You should have a file called xNormal_max78_sbm_exporter.dle or xNormal_max8_2k8_sbm_exporter.dle in the 3dsmax's stdplug folder.
So is there a bug? I have tried with max 8, max9 and max2k8(x86/x64) and seems to detect them ok. You should have a file called xNormal_max78_sbm_exporter.dle or xNormal_max8_2k8_sbm_exporter.dle in the 3dsmax's stdplug folder.
I have Vista, first I have installed Max9 and then Max7 (So I guess there is some register which has beeen overwritten). Plugin is properly installed in Max7 but not in Max9. I tried using xNormal_max78_sbm_exporter.dle from Max7 in Max9, but it doesn't work.
If you send me xNormal_max8_2k8_sbm_exporter.dle to ....... I will be grateful.
Thanks a lot.
Sorry if the bug has already been reported, but when I launch the viewer in fullscreen mode, 1280x1024, models are stretched but not the UI (I have 2 19") So I must deactivate the second monitor, and after it's work fine
Hey Jogshy, i have vista64 and have a 8800GTS card but i dont think i can use the DX10 options in xnormal like SSAO . i tried loading one of the samples to test it out but it doesnt work. it says im using dx9. where do i change the settings? btw. i installed xnormal x64.
Hey Jogshy, i have vista64 and have a 8800GTS card but i dont think i can use the DX10 options in xnormal like SSAO . i tried loading one of the samples to test it out but it doesnt work. it says im using dx9. where do i change the settings? btw. i installed xnormal x64.
Probably you need to change the default graphics driver from DX9 to DX10.
Go to the plug-in manager ( the small plug icon button near the close button ) and change it:
Ofc, you'll need Vista or Windows 7... DX10 does not work with WinXP/2k
Cheers, I was about to ask the same thing. Just downloaded the newest version and I'd just like to say big props to jogshy for keeping it free (and consequently, real). Really deep and useful prog I cant live without.
( yep, for some strange reason I like Vista .. ok, no, I hate it )
- Some UI improvements like drag and drop, multi selection, etc
- Bigger preview
- Multimonitor enhacements
- Cage autoassignment message-box after pressing the "save meshes"
- Added two buttons to clear the lowpoly/highpoly meshes
- Several bugs fixed
Btw, speaking about Vista... I had to turn the damm UAC off to make the drag and drop to work... so be warned if you use that condemned $%&"!(/ operating system...
Btw, I'm preparing a SBM mesh importer for 3dsmax so you could import cage, ambient occlusion as vertex-colors, etc... Also adding 3dsmax 2009 support... btw, the new gw:OBJ exporter included in max2009 is much better than the old max2obj !
Jogshy! I have a few new suggestions. (Do I need to say that XN is getting better and better, still? )
I would love to have a very simple extra baking option : wireframe! This doesn't require to fetch any info from the highpoly ; just a simple wireframe render of the lowpoly. Maybe you could have color settings and antialias, but that's not necessary. The reason why I am asking this is, when I bake all my maps in XN I find it annoying to have *almost* everything I need in order to start texturing ... except the wireframe, which requires me to go back to whatever 3d program!
Also, it seems like the more exotic baking options you have (proximity, convexity etc) don't produce really convincing results from the artist's point of view. They might be mathematically correct (can't tell!) but I know that XSI provides these maps in a more useable state. I don't use XSI myself but I saw its bake results for those maps and it looked like great starter maps for cavity painting, weathering aso. Maybe someone can provide some examples? ATM I take my XNormal-generated normalmap to Crazybump to create such starter maps, but If I could avoid such step that'd be awesome.
I would love to have a very simple extra baking option : wireframe! This doesn't require to fetch any info from the highpoly ; just a simple wireframe render of the lowpoly. Maybe you could have color settings and antialias, but that's not necessary. The reason why I am asking this is, when I bake all my maps in XN I find it annoying to have *almost* everything I need in order to start texturing ... except the wireframe, which requires me to go back to whatever 3d program!
Just to add to that, if it was also possible to export them as vectors in a PSD file, that'd utterly rock.
Also, how is progress on xNormal 4.0 coming along?
I would also add an edge padding to the simple AO generator, reason being is the black from the textures that I use from the AO mask (so my diffuse looks pretty to show) can bleed when the mipmaps are displayed at a distance.
I would also add an edge padding to the simple AO generator, reason being is the black from the textures that I use from the AO mask (so my diffuse looks pretty to show) can bleed when the mipmaps are displayed at a distance.
There is edge padding for the simple AO tool. Look a little harder
What would be nice is adding edge padding to the object->tangent space normalmap converter, as it ignores everything outside the uv borders currently.
it seems like the more exotic baking options you have (proximity, convexity etc) don't produce really convincing results from the artist's point of view. They might be mathematically correct (can't tell!) but I know that XSI provides these maps in a more useable state.
I coded the convexity, proximity and thickness based on the indications of some Valve artists... but i'm not completely sure what are they for or if they are useful really.
What would be nice is adding edge padding to the object->tangent space normalmap converter, as it ignores everything outside the uv borders currently.
ok, consider it done.
Also, how is progress on xNormal 4.0 coming along?
Well, I really need to finish a 3D engine before ( I calculate 3 months ). Then I could start the xn4's development by the end of this year... but i'm waiting some things to happen... like OpenGL 3 ( which has a considerable delay ).
I've done some tests though... i'm gonna use C++, wxWidgets for the look'n'feel UI, LUA/python scripting for offline rendering shaders and glFX/HLSL for the realtime shaders(which could be done in a 3rd party editor like FX Composer, Rendermonkey, ShaderFX, etc... )...
I was thinking also about converting xn4 into a GI renderer with sculping features... but perhaps that would be too complicated.
What's sure is that will be ported to Ubuntu 8.04, Solaris and Windows. The Mac version was cancelled because they are too expensive for my budget and Apple is ignoring all my dev support petitions
Well, I really need to finish a 3D engine before ( I calculate 3 months ). Then I could start the xn4's development by the end of this year... but i'm waiting some things to happen... like OpenGL 3 ( which has a considerable delay ).
I've done some tests though... i'm gonna use C++, wxWidgets for the look'n'feel UI, LUA/python scripting for offline rendering shaders and glFX/HLSL for the realtime shaders(which could be done in a 3rd party editor like FX Composer, Rendermonkey, ShaderFX, etc... )...
I was thinking also about converting xn4 into a GI renderer with sculping features... but perhaps that would be too complicated.
What's sure is that will be ported to Ubuntu 8.04, Solaris and Windows. The Mac version was cancelled because they are too expensive for my budget and Apple is ignoring all my dev support petitions
Well, i'm gonna release the 3.15 soon
It certainly sounds promising. I think you're right in thinking that the GI rendering and sculpting probably wouldn't be a good idea. The amount of time required to implement a toolset well-featured enough to make it a worthy alternative to other programs would just be too great. Better to retain the program's focus and continue to improve the workflow, I think.
I would offer to help out, but I'm quite the newb programmer. I find it all very interesting regardless though. :P
By the way, what's the 3D engine for. Is it xNormal-related, or a completely different project?
Is it xNormal-related, or a completely different project?
A different project. I started it some years ago... then created xNormal to generate the normal maps for a column model... now it's time to come back and to finish the engine hehe!
- Added three new map render types: wireframe, cavity and ray fails.
- Added indirect illumination and radiosity to the realtime ray tracing graphics driver.
- Added 3dsmax 2009 support.
- Added the SBM mesh importer for 3dsmax ( including ambient occlusion data ).
- Added "edge padding" option to the tangent/object normal map converter.
- Improved the floating point accuracy of the internal renderer ( less ray fails, less mesh scaling problems )
- Reduced the memory required to store the mesh normals ( that means greater polycount! )
- Solved deveral bugs.
ok i tried it out and im only disapointed with the cavity thingy heheh because even if i tweak the radius etc it always gives bad result, if i go with low radius it is plain white, if i go middle etc is full of holes , there doesnt seem to be happening anything .
there is another thing tho, everytime i exit xnormal it automatiocally clears the list of models, i think it would be cool that you could load the program and the settings be the same as the previous launch , then if we want to clear we use the Clear button , no ?
also the program reports less memory available than his predecessor ( ?? )
kickass tho , ambient oclusion is really clean now.
ok i tried it out and im only disapointed with the cavity thingy heheh because even if i tweak the radius etc it always gives bad result, if i go with low radius it is plain white, if i go middle etc is full of holes , there doesnt seem to be happening anything .
Don't set the radius near zero(left)... try to keep it always a bit more to the right. If the result it's too white you can play with the constrast slider a bit. Btw, the search radius is proportional to the object's radius ( 1/50 times exactly ) so perhaps you'll need to bake the hands sepparately from the body to get better results.
Well, I tested it with the smiley example and got a reasonable image:
(radius 70%, contrast 70%, 64 samples, jitter=on, notice the fine detail's cavity wrinkles )
but I must admit it still needs some tweaks. I'm gonna try to improve it with the time! Atm is very experimental. If you can pls send me any example(with wrinkles, pores, etc) I can try to adjust it better.
there is another thing tho, everytime i exit xnormal it automatiocally clears the list of models, i think it would be cool that you could load the program and the settings be the same as the previous launch , then if we want to clear we use the Clear button , no ?
Yep, it should save the last settings automatically. If it does not then it's a bug...But I cannot reproduce that behavior, I need more info.
Hey Joghsy, would it be possible to exclude the interface and mouse-pointer from screenshots? I know I can turn the interface off, but that still leaves the mouse-pointer, which I routinely forget to place near a corner or something, heh. This is basically asking for a way to counter my own stupidity, but that's what programmers are for, right?
Also, I've ran into a problem a few times where (quite randomly, it seems) turning off the interface makes everything go a lot slower.
Hey jogshy. Thanks a lot for making such a tool available for free. I use it daily, and the whole team I was with at Ubi have been converted into using it because we were 2 users of it and it was so fast and produced so clean results...
Some notes about what I'd like to see in future versions:
check boxes to bake Diffuse / normal / AO in one go, maybe even the possibility to add custom slots where you could add you own map to render by default. Maybe there is a way to do it right now but I don't see any right now (I liked it in previous versions and kinda miss this).
Just let me put a suffix next to the checkboxes this way all my objects will have _N or _D at the end of the map.
Also I'd like to see groups baking of some sort. I like to do groups of objects for baking, like for example, separate my high res / low res model in different groups that doesn't intersect with each others (often the highs have a lot of small instanced objects), to avoid baking part by part and avoid rays penetrating in other objects, and at the end I compute the final normal/AO/Diffuse in PS. But an option to set up groups and compute the final map for me would be ace. Not sure if i m clear but I could make a quick example...
Even better, tag objects (by giving then specific names, like a prefix) to identify groups and then import each groups in xnormal and let him split each group by itself. This way it would calculate the AO on all objects at once and render it separatly but with accurate result (right now if I have an object at the very top it's not getting a lot of AO as the other objects are on another group and the AO isn't calculated on everything). Again maybe there is a way to do it but no one I know was able to do it...Would work for vertex AO as I use sorrely this tool with GREAT results. Maybe an option to subdivide the high before calculating AO and saving the SBM on top of that to have really clean Vertex AO results?
Also, a way for Xnormal to bake materials that you set up in your 3D proggy...I know it sound tricky, but as I do right now it's a long and tiring process...I set my materials in Max on my high to have a base diffuse, and then Render textures the high mesh after auto unwrap (Sometimes it's a pain, because of max always wanting to do everything multi subs, so when you have mutliple objects with multi subs...Anyway...) Then Export my highs with UVs and bake the diffuse texture on the low mesh. Why not try to export from max (or maya) maybe just converting to vertex color...this way I'd just not have to bother about having to map quicly map and then bake the diffuse of my highres...I'd just toss some materials on parts I want them and Export my Highs without any mapping coordinates but with materials that Xnormal could see when baking the low meshes...
hey jogshy i tried what you told me, but i think it is still too bugyy , because i have to find the perfect value and the perfect contrast , would be cool if it could have numerical input, but still it is very accurate, ill try it with something else tonightand post images for you to see. Thanks a bunch !
hey jogshy i tried what you told me, but i think it is still too bugyy , because i have to find the perfect value and the perfect contrast , would be cool if it could have numerical input,
Yep. i'm changing it for the 3.15.1 and I'm also trying to refine the algorithm.
ok santi , heres what i got from the cavity tool, the amboc tool seems to be improved ( that or i finnaly found the right settings eheheh ) but the cavity i see that the image has exactly what i want but the cavitys itself are way to thick making it unusable , i tried playing with the radius and contrast but still too thick it would be awesome you made this as clean as your amboc tool , i would be using this the whole time !
ok santi , heres what i got from the cavity tool, the amboc tool seems to be improved ( that or i finnaly found the right settings eheheh ) but the cavity i see that the image has exactly what i want but the cavitys itself are way to thick making it unusable , i tried playing with the radius and contrast but still too thick
The right image seems to have too much contrast... but don't worry, the new options I'm preparing for the 3.15.1 will solve it.
Btw, don't use the 3.15.0.... I realized the normal compression introduced makes some small(but noticeable) errors in the tangent space normal maps. The 3.15.1 beta will be available soon.
Btw... I was trying to reproduce that mesh clean error you said but I cannot find it... do you know when happens, pls?
Replies
Try to reinstall xNormal... perhaps the ui.png texture was corrupted for some reason and that's why you cannot see the UI well. Also sure you have the latest NVIDIA drivers... the 8800GT is relatively new card so I bet the drivers aren't very mature yet.
Nice effect lol!
That banding usually appears because probably you need more samples per pixel. By default uses 16... try 64 or more. Also sure you disable the "adaptive AO" setting it to the same than the # of rays per pixel ( 64 in this case ... so use 64 rays per sample AND 64 also in the "Adaptive interval" )
Also increase a bit the AO bias... perhaps is a problem with self occlusion.
Also try to enable(or disable) the "weighting"... it can cause strange artifacts sometimes.
Caution with the 3dsmax OBJ exporter because messes the vertex normals. Better export it as SBM, 3DS, FBX, ASE, etc.
Have you tried to scale the mesh up and see if that helps? Perhaps is a floating point precision error due to small object radius.
Btw, if you can send me that mesh by email better... I'm curious how that result was produced hehe!
Is fine. Usually is very fast and gives good results to compute the per-vertex-AO with the Simple GPU AO tool... and then uncheck the "Ignore per-vertex-AO" so the xnormal software renderer could use the per-vertex-AO computed from the Simple GPU AO tool.
Btw, i'm working on the 3.14.6 with some UI improvements lilke Drag'n'drop, better remember folder system, multiple mesh erase, automatic save meshes reassignment and also experimenting with a new thing called "harden normals". Also gonna buy a graphic tablet to improve input support.
Fraid the meshes are very big and would probably take ages to send, I'll see if I can manage but I'm fairly rushed already. Cheers.
Select "Ambient occlusion" in the "Maps to render" group box. Then press the "..." button to display the AO options... Then change the "Bias" from the default 0,000010 to, for example, 0,01. Also disable the "weighting", play with the attenuation, increase # of rays, disable adaptive ao, play with spread angle, etc...
On the other hand the adaptive AA can generate some artifacts in the AO too... So try to set it to
min AA=1, max AA=4, thresholds all to 0.05. Use the diagnostics window to see how the samples are taken.
If nothing of this solves it try to send me a subdiv 3 or 4 highpoly model ( it will be compressed into 10Mb aprox ) + the lowpoly model so I can debug it pls.
I tried settings like you said, but now get worse results (grey texture with white outline of some parts and a red grid). I've tried changing everything with no luck, also tried putting results back to the way they were initially but that didn't seem to help. No idea what's going on. I sent you the high and low poly models I'm trying, hope you have better luck than me
its wierd cause Ive used xnormal before and didnt have such problems.
Im using 2 different pc's for these works and both are as updated in drivers as I can get them.
thanks, that was my first thought so I flipped my low poly normals oever and over, but it turns out its actually my high poly normals that must have been flipped by the exporter. Still not sure if I can get this working for my character though. Also I cant edit the cage or see it in the viewer unless I re-export my model at a size about 50 times bigger than it is.
so Ive learned 2 things
1. make meshes very very big
2. check the normal direction
My settings:
I imported your OBJ. Entered the 3D viewer. Adjusted the cage and saved it as a SBM. Reassigned the meshes and rendered with those settings.
So I bet the problem was just the adaptive AA, the bias and the attenuation ( notice I set the linear atten to zero ).
Hope it helps.
edit: also, just checking is it limited to one light?
I used to think trying to get it as close to the low poly as possible was the goal but my map didnt even bake when my cage was too close.
About to take the screenshots from them.... better open the output file with paint.net or photoshop and then resize it using bicubic or bilinear filters.
And yep, the xNormal's 3d viewer is currently limited to one light due to shading instruction limits... I plan to change that for xn4 using deferred shading.
But I have one tiny problem.
Is there some link where I can download sbm export plugin? I have currently installed two versions of max, and instaler always install plugin for the other version than I want.
I need plugin for max9.
Thanks
So is there a bug? I have tried with max 8, max9 and max2k8(x86/x64) and seems to detect them ok. You should have a file called xNormal_max78_sbm_exporter.dle or xNormal_max8_2k8_sbm_exporter.dle in the 3dsmax's stdplug folder.
I have Vista, first I have installed Max9 and then Max7 (So I guess there is some register which has beeen overwritten). Plugin is properly installed in Max7 but not in Max9. I tried using xNormal_max78_sbm_exporter.dle from Max7 in Max9, but it doesn't work.
If you send me xNormal_max8_2k8_sbm_exporter.dle to ....... I will be grateful.
Thanks a lot.
Sorry if the bug has already been reported, but when I launch the viewer in fullscreen mode, 1280x1024, models are stretched but not the UI (I have 2 19") So I must deactivate the second monitor, and after it's work fine
xp32bit, xnormal v.3.14.3.8989
@zOffty. Hehe nice model! Yep. it's a bug that I must correct.
Go to the plug-in manager ( the small plug icon button near the close button ) and change it:
Ofc, you'll need Vista or Windows 7... DX10 does not work with WinXP/2k
ok
i was trying to search for the xnormal options but coulndt find it. ok
now i know
( yep, for some strange reason I like Vista .. ok, no, I hate it
- Some UI improvements like drag and drop, multi selection, etc
- Bigger preview
- Multimonitor enhacements
- Cage autoassignment message-box after pressing the "save meshes"
- Added two buttons to clear the lowpoly/highpoly meshes
- Several bugs fixed
Btw, speaking about Vista... I had to turn the damm UAC off to make the drag and drop to work... so be warned if you use that condemned $%&"!(/ operating system...
Works perfect with two monitors now!
I bought a new laptop with Vista Premium, and just noticed my tablet works in the 3d Viewer 3.14.6, thanks!
I would love to have a very simple extra baking option : wireframe! This doesn't require to fetch any info from the highpoly ; just a simple wireframe render of the lowpoly. Maybe you could have color settings and antialias, but that's not necessary. The reason why I am asking this is, when I bake all my maps in XN I find it annoying to have *almost* everything I need in order to start texturing ... except the wireframe, which requires me to go back to whatever 3d program!
Also, it seems like the more exotic baking options you have (proximity, convexity etc) don't produce really convincing results from the artist's point of view. They might be mathematically correct (can't tell!) but I know that XSI provides these maps in a more useable state. I don't use XSI myself but I saw its bake results for those maps and it looked like great starter maps for cavity painting, weathering aso. Maybe someone can provide some examples? ATM I take my XNormal-generated normalmap to Crazybump to create such starter maps, but If I could avoid such step that'd be awesome.
Thanks!
Just to add to that, if it was also possible to export them as vectors in a PSD file, that'd utterly rock.
Also, how is progress on xNormal 4.0 coming along?
There is edge padding for the simple AO tool. Look a little harder
What would be nice is adding edge padding to the object->tangent space normalmap converter, as it ignores everything outside the uv borders currently.
Sheesh.
ok, consider it done.
Well, I really need to finish a 3D engine before ( I calculate 3 months ). Then I could start the xn4's development by the end of this year... but i'm waiting some things to happen... like OpenGL 3 ( which has a considerable delay ).
I've done some tests though... i'm gonna use C++, wxWidgets for the look'n'feel UI, LUA/python scripting for offline rendering shaders and glFX/HLSL for the realtime shaders(which could be done in a 3rd party editor like FX Composer, Rendermonkey, ShaderFX, etc... )...
I was thinking also about converting xn4 into a GI renderer with sculping features... but perhaps that would be too complicated.
What's sure is that will be ported to Ubuntu 8.04, Solaris and Windows. The Mac version was cancelled because they are too expensive for my budget and Apple is ignoring all my dev support petitions
Well, i'm gonna release the 3.15 soon
It certainly sounds promising. I think you're right in thinking that the GI rendering and sculpting probably wouldn't be a good idea. The amount of time required to implement a toolset well-featured enough to make it a worthy alternative to other programs would just be too great. Better to retain the program's focus and continue to improve the workflow, I think.
I would offer to help out, but I'm quite the newb programmer. I find it all very interesting regardless though. :P
By the way, what's the 3D engine for. Is it xNormal-related, or a completely different project?
- Added three new map render types: wireframe, cavity and ray fails.
- Added indirect illumination and radiosity to the realtime ray tracing graphics driver.
- Added 3dsmax 2009 support.
- Added the SBM mesh importer for 3dsmax ( including ambient occlusion data ).
- Added "edge padding" option to the tangent/object normal map converter.
- Improved the floating point accuracy of the internal renderer ( less ray fails, less mesh scaling problems )
- Reduced the memory required to store the mesh normals ( that means greater polycount! )
- Solved deveral bugs.
there is another thing tho, everytime i exit xnormal it automatiocally clears the list of models, i think it would be cool that you could load the program and the settings be the same as the previous launch , then if we want to clear we use the Clear button , no
also the program reports less memory available than his predecessor ( ?? )
kickass tho , ambient oclusion is really clean now.
Well, I tested it with the smiley example and got a reasonable image:
(radius 70%, contrast 70%, 64 samples, jitter=on, notice the fine detail's cavity wrinkles )
but I must admit it still needs some tweaks. I'm gonna try to improve it with the time! Atm is very experimental. If you can pls send me any example(with wrinkles, pores, etc) I can try to adjust it better.
Yep, it should save the last settings automatically. If it does not then it's a bug...But I cannot reproduce that behavior, I need more info.
Also, I've ran into a problem a few times where (quite randomly, it seems) turning off the interface makes everything go a lot slower.
And yep.. i'm gonna change a bit the cavity options to get more control over it.
Some notes about what I'd like to see in future versions:
check boxes to bake Diffuse / normal / AO in one go, maybe even the possibility to add custom slots where you could add you own map to render by default. Maybe there is a way to do it right now but I don't see any right now (I liked it in previous versions and kinda miss this).
Just let me put a suffix next to the checkboxes this way all my objects will have _N or _D at the end of the map.
Also I'd like to see groups baking of some sort. I like to do groups of objects for baking, like for example, separate my high res / low res model in different groups that doesn't intersect with each others (often the highs have a lot of small instanced objects), to avoid baking part by part and avoid rays penetrating in other objects, and at the end I compute the final normal/AO/Diffuse in PS. But an option to set up groups and compute the final map for me would be ace. Not sure if i m clear but I could make a quick example...
Even better, tag objects (by giving then specific names, like a prefix) to identify groups and then import each groups in xnormal and let him split each group by itself. This way it would calculate the AO on all objects at once and render it separatly but with accurate result (right now if I have an object at the very top it's not getting a lot of AO as the other objects are on another group and the AO isn't calculated on everything). Again maybe there is a way to do it but no one I know was able to do it...Would work for vertex AO as I use sorrely this tool with GREAT results. Maybe an option to subdivide the high before calculating AO and saving the SBM on top of that to have really clean Vertex AO results?
Also, a way for Xnormal to bake materials that you set up in your 3D proggy...I know it sound tricky, but as I do right now it's a long and tiring process...I set my materials in Max on my high to have a base diffuse, and then Render textures the high mesh after auto unwrap (Sometimes it's a pain, because of max always wanting to do everything multi subs, so when you have mutliple objects with multi subs...Anyway...) Then Export my highs with UVs and bake the diffuse texture on the low mesh. Why not try to export from max (or maya) maybe just converting to vertex color...this way I'd just not have to bother about having to map quicly map and then bake the diffuse of my highres...I'd just toss some materials on parts I want them and Export my Highs without any mapping coordinates but with materials that Xnormal could see when baking the low meshes...
Cheers!
Gonna make all that for xn4.
Btw, don't use the 3.15.0.... I realized the normal compression introduced makes some small(but noticeable) errors in the tangent space normal maps. The 3.15.1 beta will be available soon.
Btw... I was trying to reproduce that mesh clean error you said but I cannot find it... do you know when happens, pls?