Well I'm entering on GA but I love Polycount so I want to try and help this thread out. I have 6 yrs Cg rendering experience in Max from lighting studios to mental ray setting fun.
Here is a quick and dirty sample scene to get you guys started.
I'm pretty sure this was the first mention of Alchemy for DomWar thumbs. The trend is spreading like a fart in the shower. Not "Beauty Shots" but... *shrugs*
I'm pretty sure this was the first mention of Alchemy for DomWar thumbs. The trend is spreading like a fart in the shower. Not "Beauty Shots" but... *shrugs*
That program is like an auto tuner;trendy, but vapid... and quickly old.
I can foresee people having problems getting alpha maps to work. I think it would be helpful if people could post short guides or links to help people get alpha to work in real time viewers such as:
1.max/maya viewport
2.UE3
3.marmoset
4.xnormal
Im also asking this because I know I had difficulty with alpha last year and might have a hard time again this year.
Lols, yeah i noticed alchemy began to spread like wildfire. Its great for those quick thumbs but not much else past that. Just dont try and use it for more and you will be happy.
Funny thing is that i havent been using it myself... Good ol pencil and paper for me .
I can foresee people having problems getting alpha maps to work. I think it would be helpful if people could post short guides or links to help people get alpha to work in real time viewers such as:
1.max/maya viewport
2.UE3
3.marmoset
4.xnormal
Im also asking this because I know I had difficulty with alpha last year and might have a hard time again this year.
Do you mean sorting issues (transparency cutting holes throughout the whole mesh) ? There's only one way to really fix that, and that's to separate alpha-d meshes according to how to they stack on top of each other in the mesh. I can make a detailed explanation on how that all works in a bit.
Do you mean sorting issues (transparency cutting holes throughout the whole mesh) ? There's only one way to really fix that, and that's to separate alpha-d meshes according to how to they stack on top of each other in the mesh. I can make a detailed explanation on how that all works in a bit.
thanks coldkodiak Im sure that is one of the issues people will face, perhaps someone with good knowledge of Ue3 can post a link to a good tut on alphas in ue3
Im personally wondering: in marmoset can my alpha'd objects like hair share the same texture map as all the other parts of the character? if they do share the same uv map can I still have them as seperate chunks so that the alpha works properly? do they even need to be seperate chunks? I could spend some time and figure this out perhaps but Im sure there are some people with a superior knowledge of these engines around.
Im personally wondering: in marmoset can my alpha'd objects like hair share the same texture map as all the other parts of the character? if they do share the same uv map can I still have them as seperate chunks so that the alpha works properly? do they even need to be seperate chunks? I could spend some time and figure this out perhaps but Im sure there are some people with a superior knowledge of these engines around.
The bits with alpha need to be seperate chunks, and you can use the same texture but you will want to create a seperate material in the Marmoset viewer with the blending mode set to alpha.
The bits with alpha need to be seperate chunks, and you can use the same texture but you will want to create a seperate material in the Marmoset viewer with the blending mode set to alpha.
yup.
The other way to fix this without having to chunk every piece, is to group those that occupy a similar pivot position in xyz space, and make sure each chunk's pivots are aligned so that the camera can 'see' or sort the order properly. The closest always renders first unless hard coded otherwise, so chunks need to actually be sorted in a way to take into account all camera angles used.
If that makes sense. I'll illustrate this in a bit for those with a question mark above their head. .
The bits with alpha need to be seperate chunks, and you can use the same texture but you will want to create a seperate material in the Marmoset viewer with the blending mode set to alpha.
Correct.
One thing to remember tho with alpha in marmoset is that the support isnt very good. You cant have alpha over alpha, and your alpha chunk needs to be the last chunk in the list(preform some voodoo magic to get this to work).
If you read the tutorial.txt i wrote a bit about alpha there. I also think this has been covered in the marmoset thread as well.
If push comes to shove, you can always use alphatest.
One thing to remember tho with alpha in marmoset is that the support isnt very good. You cant have alpha over alpha, and your alpha chunk needs to be the last chunk in the list(preform some voodoo magic to get this to work).
I don't know if you need to bag on your own engine; I see sorting issues in every major engine all the time in finished games.
No one really has their thumb on this entirely.
For you guys assembling scenes in either UT3 or Gears, you can create some super fine images by using the following command:
tiledshot #
The # represents how many times your viewport will be "tiled" to make a super highres image (tiledshot 2 @ 800x600 = 1600x1200 image). The command will cycle through each section, and dump out a stitched together screenshot that is the # * resolution. After taking your shot, you can scale it down to the contest regulation sizes, or keep super high res ones available for print later down the road if/when posters are made.
A good way to use this would be to set up a simple matinee in a level where you light your model, and use the "play in editor" button to open a level and run the matinee. Once in there, open the console and type the command.
The images can get quite large, and if you go too high it may crash, so save often Can post a tutorial on how to set up a matinee for this if people want. The process itself is superfast, just has to cycle through the tiles first.
Hey guys, figured I'd post my questions about rendering here since this is the render/comp thread.
Presentation is and has always been my weak point and so I decided to finish the models a bit sooner then most so that I can have ample time to create a nice composition and get some good renders out. My models for DW are nearing completion and before I call it done I want to get my rendering sorted out so that I can tweak the textures to fit the rendering. Not sure if that makes sense but anyways.
I played around with Mental ray yesterday for a bit and got something I would consider decent out of it but it's nowhere near where I think the rest of the competition will be. Hence I'm here asking for any tips/trick/tutorials on how to get the most out of my models. How I should present them and suggestions on other rendering posibilities that might look good with the style my models are already showing.
My thread is posted in my signature and comments can be put there or on this thread.
Cheers!
The main thing is to experiment and learn what things do, also have a look at real life lighting situations and try it out. Though for starters try sticking your model on a ground geometry (can be just a plane, but dont forget this, its what grounds the character out and allows light to bounce off him onto the ground) then setting up 2 area omni's and turning up the size under the area light options, then stick in a sky light. Its a pretty basic setup but easy toi get looking nice, just be warned it might take a little bit to render on slower comps.
After a night of thread surfing I still haven't really come up with a concrete answer to the question "Can you mirror uvs without getting a seam in UT3?" I've used mirroed uvs to save on texture space because I have 2 characters and I had seen it done before(or so I thought). Now no matter what I try I can't seem to get rid of this seam. I've tried using a secondary UV channel for the light map. Can anyone save my model?
Helixx, there is a trick where you mirror your uvs after you cast with non mirrored uvs. Meaning, in your uv window, unwrap the one half, then copy and flip ( or use symetry ) Then take the one half ( that you will mirror later ) and shrink down and place in an area where there is no overlapping uvs. Change the W to 1 on the shrunk down uvs, make sure your mesh is welded, and then cast your normal from your high poly. After you cast, you should have one half that looks great, and the other that looks like crap...delete the bad half, and mirror the good half over and weld. Now when you bring in your object into Unreal, you should have the seam gone, since the normal was casted across both non mirrored uvs. I hope that makes sence, as it has been a bit since I have done this and if I got it wrong, please let me know and I will try and draw in better detail to what I mean. Good luck.
I think you also need to weld the UV seam, otherwise it might break the mesh at the un-welded seam when it exports causing a seam with smoothing which might effect the lighting on the normal? Haven't played with it yet, but if you still have a funky seam or lighting problems try welding the seam.
spark: i just tried this and it works great. no seam in the middle! i was panicking wondering how im gona get mirrored normal maps to work. great tip. saved me a whole lotta time
Thanks, once I welded the seams where it was creating lighting errors it pretty much fixed the problem. Now it's just trying to get all the bump detail that I'm getting in Max in Unreal.
Hey Vig, yah good catch as I forgot to state that you need to make it one continous mesh, and I usually do a smoothing group of 1 across as well. Unless of course you want hard edges and use another smoothing group to help achieve it. Glad it worked for you Warriah and Helixx!
Replies
Here is a quick and dirty sample scene to get you guys started.
Scene:
RESULTS:
Please don't laugh at my shitty photobucket
.nah, thanks dude.
Now do the right thing and join the winning team
Spark
i'm gonna need this thread - granted i get far enough
That program is like an auto tuner;trendy, but vapid... and quickly old.
1.max/maya viewport
2.UE3
3.marmoset
4.xnormal
Im also asking this because I know I had difficulty with alpha last year and might have a hard time again this year.
Funny thing is that i havent been using it myself... Good ol pencil and paper for me
I second Ged's request for some alpha help, if anyone has any to offer. Its always something that gives me trouble as well
Do you mean sorting issues (transparency cutting holes throughout the whole mesh) ? There's only one way to really fix that, and that's to separate alpha-d meshes according to how to they stack on top of each other in the mesh. I can make a detailed explanation on how that all works in a bit.
thanks coldkodiak Im sure that is one of the issues people will face, perhaps someone with good knowledge of Ue3 can post a link to a good tut on alphas in ue3
Im personally wondering: in marmoset can my alpha'd objects like hair share the same texture map as all the other parts of the character? if they do share the same uv map can I still have them as seperate chunks so that the alpha works properly? do they even need to be seperate chunks? I could spend some time and figure this out perhaps but Im sure there are some people with a superior knowledge of these engines around.
The bits with alpha need to be seperate chunks, and you can use the same texture but you will want to create a seperate material in the Marmoset viewer with the blending mode set to alpha.
yup.
The other way to fix this without having to chunk every piece, is to group those that occupy a similar pivot position in xyz space, and make sure each chunk's pivots are aligned so that the camera can 'see' or sort the order properly. The closest always renders first unless hard coded otherwise, so chunks need to actually be sorted in a way to take into account all camera angles used.
If that makes sense. I'll illustrate this in a bit for those with a question mark above their head. .
Correct.
One thing to remember tho with alpha in marmoset is that the support isnt very good. You cant have alpha over alpha, and your alpha chunk needs to be the last chunk in the list(preform some voodoo magic to get this to work).
If you read the tutorial.txt i wrote a bit about alpha there. I also think this has been covered in the marmoset thread as well.
If push comes to shove, you can always use alphatest.
I don't know if you need to bag on your own engine; I see sorting issues in every major engine all the time in finished games.
No one really has their thumb on this entirely.
tiledshot #
The # represents how many times your viewport will be "tiled" to make a super highres image (tiledshot 2 @ 800x600 = 1600x1200 image). The command will cycle through each section, and dump out a stitched together screenshot that is the # * resolution. After taking your shot, you can scale it down to the contest regulation sizes, or keep super high res ones available for print later down the road if/when posters are made.
A good way to use this would be to set up a simple matinee in a level where you light your model, and use the "play in editor" button to open a level and run the matinee. Once in there, open the console and type the command.
The images can get quite large, and if you go too high it may crash, so save often
people taking ss's of UT3: http://forums.epicgames.com/showthread.php?t=579438&highlight=tiledshot&page=142
user created tut: http://forums.epicgames.com/showpost.php?p=26092468&postcount=1397
note: he wrote this with the idea of taking ss's of bots, however when setting up a matinee you dont need to do all this jive.
Presentation is and has always been my weak point and so I decided to finish the models a bit sooner then most so that I can have ample time to create a nice composition and get some good renders out. My models for DW are nearing completion and before I call it done I want to get my rendering sorted out so that I can tweak the textures to fit the rendering. Not sure if that makes sense but anyways.
I played around with Mental ray yesterday for a bit and got something I would consider decent out of it but it's nowhere near where I think the rest of the competition will be. Hence I'm here asking for any tips/trick/tutorials on how to get the most out of my models. How I should present them and suggestions on other rendering posibilities that might look good with the style my models are already showing.
My thread is posted in my signature and comments can be put there or on this thread.
Cheers!
I had to make sure my objects are actually inheriting their pivots btw. I had to place some things under a null object to get it working properly.
Spark
I think you also need to weld the UV seam, otherwise it might break the mesh at the un-welded seam when it exports causing a seam with smoothing which might effect the lighting on the normal? Haven't played with it yet, but if you still have a funky seam or lighting problems try welding the seam.
Spark
I'm pissed that they added the smaller logo, wasn't there a week ago, would have used that one instead if I had found it earlier.
May be next time.