imacrabpinch
-
Posts
28 -
Joined
-
Last visited
Content Type
Forums
Blogs
Knowledge Base
Posts posted by imacrabpinch
-
-
I've noticed an issue that has been ongoing for quite some time. I've validated that the issue isn't only me - as it happens to other creators as well.
After a great deal of time rigging clothing on the beta grid, I later find out that in main grid the rigs are just a little off - causing clipping in places that were not clipping in beta grid - same shapes, same mesh, seemingly same weights.
Is anyone else aware of this issue? Is it being worked on? Do certain sims on the beta grid do a more accurate job?
To continuously have to fix rigs costs in both linden and time. Any workarounds?
Thank you.
-
How would one go about creating a BOM Head and / or body so that system skins could be used? My initial route is to simply sculpt and reshape the default SL avatar (that contains original UV maps), while trying my best not to crazily distort the UV. I am assuming the topology of the new mesh would have to match the original? Or is there another way I'm missing.
-
To get this effect, once the product is finalized you scale it up (the fattening/inflate scale) ever so slightly and set the transparency of the 2nd one to something like 50 or 60, or whatever looks good. You can do this in world. With that in mind, since you are doubling the mesh, try to have the base mesh be reasonable in tris count. There are some designers using fibermesh from zbrush, but the polycounts get way out of hand really fast - like a 600k poly sweater that people have been wearing lately.
-
On 9/1/2019 at 6:23 AM, OptimoMaximo said:
Arnold bake works using a simulated camera to snapshot from each surface normal by default. It worked in Arnold4, but i haven't tried with Arnold5 yet. Your best bet in this regard for now is to try baking its albedo with Turtle, using the camera setting as any other setting except surface front, which is the default, and then postprocess that over the arnold bake result. I don't have access to Maya until tomorrow, so i can't do experiments for now.
Thanks for your help! Took some time, but I think I figured out something I can work with. I was able to bake the diffuse using the transfer maps feature (under texturing), just had to bring in my low poly model. I couldn't figure out turtle, as the tutorials and documentation all pointed towards a texture baking editor which I can't seem to find in 2019.
-
On 3/21/2018 at 4:12 PM, OptimoMaximo said:
i guess you mean the baked iridescence effect on the color too.
If this is the case, that is done in Maya using the Hypershade. Among the Utility nodes, there is one called Sampler Info. The following pictures show the node set up
Connect the facing ratio from the Sampler Info node to the U coordinate of a Ramp Texture node, as shown here
Make sure that the ramp node is set to U ramp, choose the interpolation you like, then choose some colors for the ramp. Each color will appear on faces that are angled away from the camera consistently to their position on the ramp
Final effect on a randomly squashed sphere. This effect can be baked for any attribute you will connect the ramp node to. In this case it's connected to the color node to better show it on screen instead of having to render it. Notice that the right-most color on the ramp is the one that shows on the surface when it's perpendicular to the camera, the left-most is the color applied to the most facing-away areas
Hope this helps
I can probably put in some more hours to figure out the answer to this, but I thought I'd ask here. You sir, have helped me a lot.
The node setup works great in the render view-port, but I'm having a problem getting the effect to show up as a baked 2-D texture. What am I missing?
Here is a super quick snap of the 3D viewport using Arnold:
Here is what I'm getting when using Arnold's bake to texture feature (the front piece):
It's only showing the right most color, not the full ramp effect.
Mind you, I haven't refined it to as I want and they are very quick noisy renders, just trying to get the concept down. Also, I've only been in Maya for about a week, so I'm sure I'm missing something key. I'm using the same node setup as you had posted, with the exception of adding aistandard in place of the lambert, but I was getting the same result with lambert.
Any tips or clues? What am I missing? Thanks!
-
On 8/8/2019 at 4:46 PM, SeusEverling said:
Thank you everyone for your input. I have a desktop at home, but unfortunately, I can't really fit it in my suitcase. I've been traveling with a Macbook Pro for 3 years, but it's time to replace it, and I thought I would bite the bullet and switch back to Windows for a better SL experience. Am I the only person who needs a good laptop for SL?
Yes, a laptop costs more and may not last as long or be as up-gradable as a desktop, but you can find some deals and the difference in performance doesn't have to be that bad here in 2019. A lot of people stick their nose up to this, but I'm a big fan of Dell outlet for deals on laptops so I'd recommend to check their stuff out. Anything with a GTX of 2060 or 2070 should run SL fairly well. Even cheap laptops aren't that bad anymore but may not handle advanced lighting settings.
Check out this website: http://forum.notebookreview.com/ as theres all kinds of information about laptop reviews and such. Hope that helps!
-
On 1/4/2018 at 5:57 PM, OptimoMaximo said:
Maya ships with SolidAngle's Arnold from version 2017. Previous versions came along with MentalRay. I personally never liked MentalRay, therefore my first renderer was VRay. Basically, they do pretty much the same: physically based rendering. Therefore, it's hard to say this image was made using this or that renderer. The images you posted here, for example, who can really tell what renderer was used? In capable hands, a renderer gives the same realistic results.
Usually, the best way to begin the lighting stage of a scene setup is to have a nice HDRI image for some initial proper lighting. This simulates a fake environment from a picture, casting light accordingly. There are tons of free HDRI images online, many of which are studio lighting. All these images look like they were lit in a photo studio, don't they? Hence my guess of the use of such HDRI.
Moreover, a render can be split into different "passes", each pass delivering a specific visual component of the final image. An example could be the occlusion pass, where the same object is being rendered using only ambient occlusion, which is later used for compositing.
Now that Maya ships with Arnold, i'm all for it. But it all boils down to personal preferences.
Not to dig this back up, but maybe you could help. I have been trying to evolve my bakes a bit with a combination of things. The problem I have with most HDRIs for clothing is that there is uneven light on say, the back or sides compared to the front. Any recommendations on where to find an HDRI where the lighting would be even around, while still giving off shadows similar to the picture the OP posted?
Note: I typically use cycles and I don't use Arnold (so its been effort to experiment without full knowledge of it), but I've been tinkering with Maya and the thought of switching completely over. I know there may be those large pre-built in lighting domes and I wonder if those give the same effect? Thanks
-
On 7/11/2019 at 4:48 AM, BeckyNoir said:
I would so apreciate if you share the info. I'm driving crazy with the sticky pants
TY
Sure. I forget exactly what I was doing with data transfer, but there was a way to change the distance of which it copied the weights, so that the "sticky pants" didn't occur. However, an even better/easier method for me is to use the "convert to bind pose" function in Avastar and then locking the bones in place. I've had pretty good results like this.
Here is the video tutorial that I took information from. It doesn't specifically mention our sticky-pants problem, but I was able to extract the information to apply it to our problem. Hope that helps!
-
I've been reading through this thread and im kind of confused about some of these responses.
On 4/21/2019 at 6:59 AM, Motoko Oanomochi said:Hello,
There are "secret weapons" specific softwares, but unfortunately, it is impossible to speak about it here, on this forum.
For example, do you know there is a piece of software, somewhere, to rig automatically any 3D objects for Second Life skeleton ?
Do you know there is, somewhere a piece of software to manage specifically the LOD of animesh specifically for Second life for any 3D objects ?
In fact the best softwares, even and mainly, if they are dedicated to Second life / Sansar, they can't be mentioned on this forum.
Um. What?
I know you can autorig from MD for the sansar avatars, but what else is there? Avastar helps but I wouldn't call it an autorig.
Oh, and to the guy who can't walk in the house ... Did you change the physics type from convex hull to prim? I know it's a little like asking, is it plugged in? But this is often the solution.
Overall though, to build houses just learn your 3d program of choice. Use additional programs to help fix problems. Learn how to line up vertices to be precise, or snap tool. If you want, you can use prims in world and then convert it to mesh. Nothing fancy. There are very nice houses built this way.
The hard part is puzzling out the uv map and figuring out how many textures you're okay with using and or baking.
-
On 11/27/2018 at 10:35 AM, arton Rotaru said:
In the first issue of the Vertex magazin is a brief breakdown of texturing a jeans.
https://gumroad.com/l/GAAOsAnd this is a basic jeans tutorial from a SL creator.
https://indidesigns.wordpress.com/photoshop-basicstutorials/jeans-texture-seams-tutorial/Those might give some hints, or ideas of how to proceed.
Thank you for this. Helps. Thank you kind stranger. That magazine/book - golden. I'll read it all. The seam texturing - very useful.
Anyone else care to share their (general) texturing workflows or leave a clue?
-
Can anyone else contribute to this? It is a second life texturing forum, and I asked a pretty basic question on texturing. I understand many of you are very into optimization, but I didn't ask about optimization. I know sometimes they go hand in hand, but that would be another post.
If you are reading this, what do you make, and what is your texturing workflow? Everyone is different, and that is fine, but I'm just wondering what an average workflow would be from someone who creates decent looking items. For example, one thing I learned that has helped me a lot, is that I can just take screenshots of substance painter instead of baking it all out. - what a time saver. Any other tips, or experience, or anything? Please.
Thank you!
-
On 11/18/2018 at 1:24 AM, Penny Patton said:
I've been meaning to knock out a basic texturing tutorial, mainly demonstrating good habits in texturing for SL and why they're important.
To put it simply;
- Keep texture use as low as possible while achieving the desired effect. In other words, don't use five 1024x1024 textures when using a single 512x512 will look pretty much the same unless you zoom in closer than anyone will. Also, don't use multiple slightly different textures when using a single texture will look more or less the same.
- Don't use wasteful textures that are mostly memory hogging empty space. Set up your UV maps so as much of the texture as possible is used. If you have a dozen textures for your model, each using a small amount of the overall canvas, combine them into a single texture, using all of that empty space.
- Remember, from a practical standpoint there is very little difference between SL and a videogame when it comes to how rendering content works. And in SL you don't have the benefit of a single art team working together to keep resource use in line, so it's good to compare SL content to art assets from games like Skyrim and Fallout 4 (the base games, not the framerate killing 4k texture mods) in terms of how you should be texturing.
Maybe not exactly what you're looking for, but these tips are important because every computer has a limited amount of memory for textures. That same memory is shared with other features, like shadow maps. What's more, SL won't even use all of your texture memory properly. Excessive texture use has become one of SL's more prominent performance issues.
Thanks for the reply :D
I was looking more so for actual techniques and methods, as I'm already aware of optimization. However, I suppose digging up more info about Skyrim and Fallout couldn't hurt. Since my original post, I've been increasing my knowledge and changing up my workflow considerably - with just a couple little things making a visual difference. Only so many hours in a day, and so much to learn - I just wish there was more content available that was specific to the sl workflow.
-
Hi everyone. I came to this section of the forum to find as many texturing resources as possible, in hopes to step up my game. I'm looking to learn and improve on texturing techniques.
My current texturing methods involve using blender cycles rendering, substance painter, and photoshop. I use zbrush as well for high to low poly workflows.
Most of all, I find myself struggling with texturing realistic denim. I feel like if I could learn how to texture perfect seams like some of these stores, I would be on a good path. I can get the mesh looking good, but I'm just lacking that IT factor that I find from some of the big stores. The use of sculpting stitches doesn't seem to do it and substance painter also seems s little off. Maybe I'm missing some photoshop technique? Of course, when I attempt to reach out to these other creators they are either tight lipped, silent, or simply don't know because of outsourcing.
Any recommend resources I could read or watch up on? Don't be shy, as I would literally read an entire stack of books if it means improvement. Thank you and its very much appreciated.
-
On 11/8/2018 at 10:43 AM, Kyrah Abattoir said:
I'm not sure what you want. You're not gonna find a perfect solution, transfering weight gets you 50% of the way, the other 50% are up to you.
As the title states, data transfer weights is what I wanted. I figured it out since the original post. Thanks anyway for anyone with good intentions of helping.
-
Can anyone point me to a place or tutorial where I can learn how to rig via data transfer in Blender? When I use avastar for pants, copy weights, and transfer weights from mesh body to garmet using the tools bar, anything around the thigh has to be manually separated...aka "stickypants". It happens with female dev kits usually. I see it happen to other people too, so its not just me. Since I prefer to weight each body (unlike some people out there), it leaves me having to go through this correction process for quite some time.
I heard data transfer is a way to avoid this but no one ever elaborates. Any tips? Just trying to mainstream my process.Thank you
-
Last night I spent a considerable amount of time trying to fix weighting issues in blender on a piece I made, and it still frustratingly isn't perfect, so it leads me to a general question for the crowd here. Does maya offer any advantages in rigging over blender? I have a student copy of maya, and I'm not sure if its worth learning.
Just for some further details, I usually don't have an issue fixing any problems I run into with blender, but I do tend to always run into "this or that" unexpectedly. Main examples include - the model distorting slightly when in world compared to the look in blender and the sticky pants issue (when transferring weights for anything on the thighs for most models, the pants stick together where I have to manually weight paint to fix - takes about 10 minutes extra per body). The sticky pants issue makes me dread getting to the rigging stage of any pants. I heard there was a data transfer way to do this, but I have read up and still haven't figured that one out.
Other issues I commonly work around - One of the popular dev kits I use always shifts upwards when I rig in blender version 2.79/avastar 2- so its common practice for me now to switch to an older version specifically for that one dev kit, then switch back. Also, manual weight smoothing (blur tool) doesn't work correctly in older versions of blender so I use the latest for smoothing. Then another issue is if I rig in an older version, I need to switch to the new version to export in A pose, things like this. What I'm trying to say is, I feel like I'm always switching back and forth versions, etc. I've heard of other people doing the same thing in blender, but I haven't heard of that for people using maya.
Someone also told me that in maya, you don't need to worry about the posing of the avatar before rigging. In blender, its a simple convert to bind pose button - but this is still one extra step.
A little rambling but I hope this makes sense. I'm 100% self taught so its always been a guess and test kind of mode I'm in.
So, is there any advantage to learn maya specifically for rigging? Like I said, I've always been able to figure out any issues, but if maya offers any advantage, I'd learn it.
Thank you!
-
8 minutes ago, AyelaNewLife said:
That's actually the reason I made the swap, even though I slightly prefer the Slink shape. I just got so tired of going to shopping events or searching for something specific on the MP and running into the wall of Maitreya exclusives. Everyday clothes (and more smarter/sexy ones) are easy enough to find, but trying to find anything remotely niche is a lottery.
It's almost a creator-enforced monopoly, and it honestly sucks. But I have no idea how to fix that problem (or even if it deserves to be fixed).
Just out of curiosity - what kind of niche clothing do you feel is needed the most?
-
This is probably something unpopular to say, but I like how sl is non-optimized and laggy. If it wasn't, it wouldn't be the experiment they call sl.
-
1
-
1
-
1
-
-
Sometimes I get that problem too, and I have to open the file up in blender 2.72 for it to work right. Also, play around with the "convert to bind pose" feature of avastar.
-
23 hours ago, OptimoMaximo said:
Although what Fluffy says is true and you can attach/detach Lattices after import (using node editor wiring and piping as there's no menu call for attach/detach, at least in version 2017), the problem sits in the specific mesh you're trying to fit, so a previously working Lattice won't necessarily work as expected on any other model you may want to refit.
As i was saying in my previous post, it's just a matter of doing it in steps, using a low res Lattice at first for rough reshaping, applying the deformation, again with another Lattice with higher resolution and keep reshaping this way until you get a close-to-perfect match, and finally fixing the smaller discrepancies using SoftSelection (Maya)/Proportional Editing (Blender).
There is another system though, when two devkits bodies are very similar, using the shape sliders, but this method usually ends up with stretched geometry/textures and, therefore, it is not always optimal. The wrap deformer (Maya) provides a similar and better solution to use the shape sliders approach, although it also provides pretty headaches to avoid name clashing that Maya is so picky about when setting up the resizing scene, requiring imports of static versions of the target mesh body before rigging, resulting in some time being wasted in comparison to the Lattice method, which is anyway more accurate in the end.
Thank you for your advice. I've been following it, but use of the lattice modifier seems to take a long time to perfect, especially since the dev kits are T pose and A pose, leaving any sleeves to be rotated. Its been hard to tweak perfectly. I'm not completely satisfied with the results or method.
I'm still a bit unsettled to this solution, because it may actually be faster to re-UV and rebake for every body at this point. - just wish I knew for sure exactly what the other 50+ creators who use an MD to Zbrush to 3D program workflow are doing.
-
34 minutes ago, Fluffy Sharkfin said:
@imacrabpinch As Optimo points out above, using the Lattice Deformer tool is probably the easiest way to refit your meshes to different bodies.
One of the benefits of the Lattice Deformer (in Maya at least) is that you can detach and reattach different lattices to your model. So once you create a set of lattice deformers that adjust the overall shape of a model to fit the bodies you're working with you can export them and then reload and apply them to other models.
In the majority of cases you'll still need to make additional adjustments afterwards since each model will have different dimensions and topology so its very rare that the same lattice will work perfectly for two different models, but it does cut down on the amount of time you need to spend adjusting each model compared to having to creating a new lattice for each article of clothing you make.
Thanks. I believe each lattice could be exported/imported in blender as well. I suppose I'll just need to work on using that tool more efficiently.
What frustrates me is I think there is another way that is easier, but I haven't yet worked it out - probably because I'm a noob. I figured out I can triangulate a garmet - save and Import it into MD - shape it exactly how it should be ,100% accurate within 30 seconds or less- save and import back into blender - and then data transfer the UV map from one to the other (and because it was triangulated, they match). The new problem I have is the UV map, although looks like it should, is still not translating in the correct vertex order - and I feel like there is something else I could do to fix this but my knowledge limits me- frustrating. I should probably be on the blender forums to ask about these.
For now, I'll continue to work on my lattice shaping skills.
Also curious : Do you know if anyone on MP sell the "starter" lattice for each body?
Thanks all for the help - just trying to learn as much as possible.
-
I wanted to bump this due to conflicting information. When I ask around, I see many other non-experienced people having the same questions. I've been doing my homework (because I rarely ask lazy questions), but I really want to be able to make clothes for all bodies as efficiently and accurately as possible. Note, I have all the official dev kits I need.
First of all, fitting to multiple mesh bodies seems to be a very common practice. I realize every workflow is different and there should be more than one way to do things, but I am still not crystal clear on the best practices of creating clothes for multiple mesh bodies.
Here is what I know will work, but is not optimal: Simply create the garmet in MD and drape on each body as a first step - ensuring you are satisfied with your UV map. - then you can use the same UV throughout. The problem I have with this, is that it limits your UV map. I've seen at least a couple creators use this approach - but I like to use zbrush and remap into polygroups, etc so I can play with it later on more.
What I thought I could do (but can't seem to) : Use the OBJ to garmet import feature in MD as a last step - drape it on the body I need, and export. - The problem I have with this is that the OBJ file seems to lose its UV map (no matter what settings I've tried). Also turns it to triangles initially, which can't seem to be quaded once back in MD - I can quad this back in blender, but it doesn't end up matching 100% (therefore can't transfer/link the UV!) - I could technically work this way and re-uv and rebake each one, but that seems very un-necessary and time consuming.
The last resort - I find this an imperfect method - would be to simply grab (using sculpt tools) to fit the body kit in blender/maya. People have also said to use the lattice modifier, but I also have found this to be somewhat imperfect results - or maybe I need practice.
I just really want to know how EVERYONE else is doing this, seemingly common practice thing. I know a couple years ago before mesh bodies, there was a mayastar plug-in that changed the size of the mesh garmet to fit xsmall, small, large, etc. - Is there a trick having to do with using the sliders?
What am I missing?! Please help. Thank you!
-
Hi all,
I'm a relatively new creator, but I'm learning a ton and advancing in the best way that I can. A problem area I'm having right now is in learning how to make realistic looking denim. I know its been a practice since the beginning of sl times, but I can't seem to find many resources (especially specific to sl).
With my current workflow, I've been able to make decent looking denim, but it has more of a cartoon look to it once its uploaded. I mainly use a combination of MD to Z Brush to Blender, and bake a high poly model onto a low poly model using cycles. I use a few programs in-between for various reasons, but that's the gist. For the denim texture look itself, I was applying it as noise in zbrush. I do use PS, but usually very minimally to overlay or fix problem areas of the bake. For example, I was adding stitches directly in zbrush instead of PS (as the high poly), but I'm thinking it may be better to simply use PS for that?
Overall, I want to be able to make similar denim to those couple of popular stores that start with a V or a B (not to name specific stores, but I admire the detail in the mesh/textures they sell). Should I be spending more time in photoshop? Should I be projection painting? Should I be using more realistic resources instead of focusing on cg? If it is a PS thing, are there any resources or tutorials that would help? I'm willing to learn as much as possible. I feel like I'm missing a vital piece of the puzzle.
Any tutorials/forums/resources/advice (even paid), would be great help. Thank you!
-
As a newer creator I'm a bit interested in this question as well. I know if your workflow only uses blender and MD, and you have the body dev kit, it would simply be a matter of importing that body into MD, draping it on the body (may take some tweaking), and then going back into blender or maya to rig it. You can use the demo bodies to test it in beta grid if you don't have them all.
However, I'm a bit confused of what to do when you use zbrush or other sculpting in the workflow, if you wanted the in-game model to have the z brushed shape/detail. If I were to drape the z brushed model onto a body, it would lose its shape. Not sure how one would do this easily, short of sculpting each model and just trying to make them identical. I realize some of the finer details would just be a high poly baked on texture, but I'm not sure how to keep that overall wrinkle shape except for re-sculpting each one. Can anyone help provide insight on that?
Are people just kinda grabbing and reshaping to the other body?
Hopefully that made some sense.
Facebook
Instagram
Twitter
YouTube
Flickr
Beta Grid vs Main Grid Rigging Differences?
in Mesh
Posted
so are you saying to upload closest to ground level instead of a skybox?