So this looks pretty clean here, have you noticed any areas that will need more geo, skewed details or anything like that?
It would be cool to see a variety of angles and larger images. FPV angle specifically too.
Yes there were several areas that were having major issues where corrective geometry was needed (e.g. an occasional ngone or two I missed and caused cage intersections). Skewed details, there was a bolt on the body that was sort of skewing into an oval shape instead of a circle, fixed that however. I also want to go back into the high poly and modify some of the details with more bevels for a stronger normal map.
Here are some more shots where I am pointing out some of the minor problems and a major problem area I am having trouble diagnosing.
Major Problem
FPV. Using 3point shader here obviously, but I must be using it incorrectly. I put down an omni light, refresh 3point shader, and select the omni light to use. But no matter where I drop it, the multiplier, or high low/ far away/close the omni light is this is the lighting it always results in. Its not updating in real time either, and I am fairly confident its suppose to.
@Perna, the major problem was fixed with adding SG there thanks. I did not recognize this as a SG issue initially because it was so extreme. Regarding the 3point shader problem, I resolved this as well. It turns out a hidden skylight on another layer was the culprit, I removed this and all is well.
Anyways another test bake fixing the minor and major problems from before with a few new small issues that need to be fixed.
edit: also SG use
Edit: Fixed the above bake related issues now, moving onto step 3.
Just had to mention to everyone involved what a great thread this is. Also, hey Perna/EQ! Your advice throughout this forum has been amazing. Thanks for taking the time and being so patient.
I corrected all the problem areas found in my test bakes, and then scaled up/down the correct areas. Therefore I am working through step 5 at the moment which is UV packing. Here is what I have so far. I visited online portfolios and looked at flats to get an idea of how much padding I need. How is this looking? Hope I didn't jump the gun again, I went through and double checked I had taken the right steps before beginning packing.
Thanks again
Here are the steps again for reference, bold indicates a completed step 1. Make UV islands uniform in density 2. Before laying out final UVs quick auto-pack and test bake with no SG use, find problem areas, add SG where needed. Add corrective geometry if needed. 3. Go in and scale up detailed areas 125/150% and scale down occluded areas (e.g. the bottom area of the mag holder no one will ever see) 4. Make Sure UVs are final at this point
5. Proceed to final uv packing ASSUMING your temporary quick packing + test bakes are error free
6. Do another test bake
7. Another round of corrective geometry or smoothing group tweaks - I'll often find one or two little things after my final uv pack is set, that I didnt notice with the quick bakes
8. Prepare for final bakes
- Explode high/low
- Make sure highpoly meshes have RGB, or RGBCMYW colored materials assinged so you can easily bake material masks
- Make sure any floaters are set up properly, ie
Edit: read your reply Perna, thank you. Somehow I missed the circle shape being rotated I was keeping an eye out for that too. Going back to fix that and border padding, thanks again.
Right underneath the main body island there's a circle shape in the UVs which is slightly rotated to the side. Any reason for not having those edges straight?
User error. I was keeping an eye out for rotated pieces such as this, but clearly I passed over it (I blame lack of coffee this early in the morning ). Thank you for bringing this to my attention. Is the edge padding enough/too little?
Here is an update removing the rotation from the circle and showing density for the gun, FPV shots and adding uv border padding.
Unfortunately I am using max 2011, but I mimicked the settings you described with materials and saved them to my own maxstart for future use. You may notice the body itself, especially at higher tiling, is not the exact same density as the other 125% close parts. This is because the gun body was too long to fit into the UV box without scaling it down a little bit. Density acceptable?
wait, left of the 128x128 image there's a big box, then a medium, then a small box.. and the edge of the small box seems to have very stretched UVs.
!
Corrected thanks. I think I am ready to proceed with final bakes, except for this one error that I am having trouble correcting.
Lighting issue. I have attempted adding smoothing groups there (which is what I suspected the problem to be), turning hidden edges, and redoing the topology in that area but it keeps producing something similar to what you see below. The mesh is planar there as well, I checked with flat shading. 3ps quality mode enabled.
Im pretty convinced its a smoothing group issue and since this thats not working at the moment I think I will reset max and import to a new scene and see what happens.
Thanks again perna. I ended up fixing the issue I described above after resetting max and starting a new scene and adding a SG to that area. Here is a test bake, max 2011 is wonky. I see no problems so hopefully I can proceed to the final bake here shortly.
Here are some shots of the test bake, zoom out + edges, and the smoothing groups.
For the hidden edges, I was not sure if I needed to turn them or not since there are no errors with quality mode enabled. If I do need to start turning them I have read to generally keep them flowing in the same direction. Also do I need to rebake if I turn edges on the LP?
The geo or SG setup around the hole through the bottom of the stock looks a bit awkward, hard to tell at that size. But if shading there is perfect it's nothing to worry about
Yeah shading looks fine there, so I will leave it as is. Anyways I think I am ready for final bakes e.g. hammersley with 1.0 quality.
You will notice I split off the mirror part of the handle where the text will be (inside that capsule shape). The text is not there yet I was simply going to use xnormal height to normals, because in my tests that looked slightly better than the subD text I had modeled.
Also thinking about changing the edge padding on the bakes from 4 to 3 since its spreads out more and leaves less padding than I had planned on.
There's a very ugly and obvious center seam showing in the bottom pic, near the sight.
looks like there's a center seam in the grip area too (same bottom pic)
Here is the fix for anyone reading through this thread. I normally make 2 save files before doing bakes, one file where the mirrored UVS are overlapping, and another where they are moved over 1 unit in the UV square. I was still working from the overlapping one during the bakes by mistake resulting in the seams. To fix this follow the workflow in the polycount normals wiki
Typical Mirroring Workflow
Delete the mesh half that will be mirrored.
Arrange the UVs for the remaining model, filling the UV square.
Mirror the model to create a "whole" mesh, welding the mesh vertices along the seam.
Move the mirrored UVs exactly 1 unit (or any whole number) out of the 0-1 UV square.
Bake the normal map.
Here is the update, no more seams on any of the mirrored pieces. I think I will now proceed to final bakes with hammersly 1.0 quality. In regards to the maximum edge padding, I am curious why its needed if you have no visible seams at default value of 2 or at an increased value of 3 or 4. I am not questioning you at all, rather I am trying to understand the logic behind setting it at max.
Looks like a bug, my UV's are all within the 0-1 range. And I have encountered strange problems in max 2011 hand over fist, I can only assume this may be another to add to the list. I plan on uninstalling it before starting my next project.
0-1 Space
I moved mirrored pieces over exactly one unit everytime
If you are using RTTAssist and having it combine maps, then padding at the highest will be an issue since it is taking a seperate image for each object and then combining it on top of each other, thus giving you the effect you posted of your normal map. If you were baking this directly from max without RTTAssist the padding would take into account each island and would terminate. That is my first thought, I could be wrong.
Who wants to watch me struggle through my first attempt to texture now?
Anyways I blocked out some simple colors in the diffuse. And I made a game plan for what I think the spec/gloss levels for various materials might be, although I may very well be wrong. I plan on keeping the diffuse pretty plain, and putting the material properties mostly in the specular. My goal for this gun is to have it look brand new right off the factory line, no scuffs or scratches.
Material Properties game plan, using the noise from actual materials rather than artificially generating in PS
Diffuse need to change a few parts still but the block out is getting there
@chaosquack, thanks for the suggestions. I actually was using RTT assist to combine the maps. However I rebaked and combined them manually and still produced the same effect. I think this is a max 2011 issue, and I plan on ditching it. I was able to get the padding higher than the lower values I was using before, but still no where near the max of 64. I will just have to cut my losses and get this finished and make sure my next model is made in 2009 or another version that has max padding work.
for your diffuse, pick from shadows. for specular, pick from highlights. for gloss, pick from what you see.
i'll post this again.
Thanks for the feedback. Am I doing it correctly now? Please note the only objects modified since my original image is the body, handle and silencer. In addition
Replies
It would be cool to see a variety of angles and larger images. FPV angle specifically too.
Yes there were several areas that were having major issues where corrective geometry was needed (e.g. an occasional ngone or two I missed and caused cage intersections). Skewed details, there was a bolt on the body that was sort of skewing into an oval shape instead of a circle, fixed that however. I also want to go back into the high poly and modify some of the details with more bevels for a stronger normal map.
Here are some more shots where I am pointing out some of the minor problems and a major problem area I am having trouble diagnosing.
Major Problem
FPV. Using 3point shader here obviously, but I must be using it incorrectly. I put down an omni light, refresh 3point shader, and select the omni light to use. But no matter where I drop it, the multiplier, or high low/ far away/close the omni light is this is the lighting it always results in. Its not updating in real time either, and I am fairly confident its suppose to.
Anyways another test bake fixing the minor and major problems from before with a few new small issues that need to be fixed.
edit: also SG use
Edit: Fixed the above bake related
Thanks again
Here are the steps again for reference, bold indicates a completed step
1. Make UV islands uniform in density
2. Before laying out final UVs quick auto-pack and test bake with no SG use, find problem areas, add SG where needed. Add corrective geometry if needed.
3. Go in and scale up detailed areas 125/150% and scale down occluded areas (e.g. the bottom area of the mag holder no one will ever see)
4. Make Sure UVs are final at this point
5. Proceed to final uv packing ASSUMING your temporary quick packing + test bakes are error free
6. Do another test bake
7. Another round of corrective geometry or smoothing group tweaks - I'll often find one or two little things after my final uv pack is set, that I didnt notice with the quick bakes
8. Prepare for final bakes
- Explode high/low
- Make sure highpoly meshes have RGB, or RGBCMYW colored materials assinged so you can easily bake material masks
- Make sure any floaters are set up properly, ie
Edit: read your reply Perna, thank you. Somehow I missed the circle shape being rotated I was keeping an eye out for that too. Going back to fix that and border padding, thanks again.
User error. I was keeping an eye out for rotated pieces such as this, but clearly I passed over it (I blame lack of coffee this early in the morning
Here is an update removing the rotation from the circle and showing density for the gun, FPV shots and adding uv border padding.
UV
Density
32x32
64x64
128x128
Corrected thanks. I think I am ready to proceed with final bakes, except for this one error that I am having trouble correcting.
Lighting issue. I have attempted adding smoothing groups there (which is what I suspected the problem to be), turning hidden edges, and redoing the topology in that area but it keeps producing something similar to what you see below. The mesh is planar there as well, I checked with flat shading. 3ps quality mode enabled.
Im pretty convinced its a smoothing group issue and since this thats not working at the moment I think I will reset max and import to a new scene and see what happens.
Here are some shots of the test bake, zoom out + edges, and the smoothing groups.
For the hidden edges, I was not sure if I needed to turn them or not since there are no errors with quality mode enabled. If I do need to start turning them I have read to generally keep them flowing in the same direction. Also do I need to rebake if I turn edges on the LP?
Yeah shading looks fine there, so I will leave it as is. Anyways I think I am ready for final bakes e.g. hammersley with 1.0 quality.
You will notice I split off the mirror part of the handle where the text will be (inside that capsule shape). The text is not there yet I was simply going to use xnormal height to normals, because in my tests that looked slightly better than the subD text I had modeled.
Also thinking about changing the edge padding on the bakes from 4 to 3 since its spreads out more and leaves less padding than I had planned on.
Some more shots of normal + ao
Normal
AO
Here is the fix for anyone reading through this thread. I normally make 2 save files before doing bakes, one file where the mirrored UVS are overlapping, and another where they are moved over 1 unit in the UV square. I was still working from the overlapping one during the bakes by mistake resulting in the seams. To fix this follow the workflow in the polycount normals wiki
Typical Mirroring Workflow
Delete the mesh half that will be mirrored.
Arrange the UVs for the remaining model, filling the UV square.
Mirror the model to create a "whole" mesh, welding the mesh vertices along the seam.
Move the mirrored UVs exactly 1 unit (or any whole number) out of the 0-1 UV square.
Bake the normal map.
If you forgot to do the above and already spent the time rendering out finals you can try to fix it using the steps established in this thread
http://www.polycount.com/forum/showthread.php?t=51088
0-1 Space
I moved mirrored pieces over exactly one unit everytime
Really an accomplished work.
Nice thread!
Anyways I blocked out some simple colors in the diffuse. And I made a game plan for what I think the spec/gloss levels for various materials might be, although I may very well be wrong. I plan on keeping the diffuse pretty plain, and putting the material properties mostly in the specular. My goal for this gun is to have it look brand new right off the factory line, no scuffs or scratches.
Material Properties game plan, using the noise from actual materials rather than artificially generating in PS
Diffuse need to change a few parts still but the block out is getting there
diffuse blocked out on model
@barrie allen, thanks appreciate it.
@chaosquack, thanks for the suggestions. I actually was using RTT assist to combine the maps. However I rebaked and combined them manually and still produced the same effect. I think this is a max 2011 issue, and I plan on ditching it. I was able to get the padding higher than the lower values I was using before, but still no where near the max of 64. I will just have to cut my losses and get this finished and make sure my next model is made in 2009 or another version that has max padding work.
Right. The shots you are responding too was just a simple diffuse color block in no spec or gloss. Refer to updated shots in this post.
i'll post this again.
Thanks for the feedback. Am I doing it correctly now? Please note the only objects modified since my original image is the body, handle and silencer. In addition
Sampled Tones