|
I've got a vertices problem. My no-budget workflow has been this: 1. Model in AutoCAD 123D, export to STL 2. STL to 3DS in meshlab 3. Render and arrange in Art of Illusion The problem is when I render I get this weird "puffy" effect due to where the vertices are. Like this:  Is there any way to add more vertices on the flat surfaces to fix this in Meshlab? I tried using uniform mesh resampling, which either created a badly messed up model, or a model that looked a little better but the polygon count went absolutely through the roof. Three-Phase fucked around with this message at 15:43 on Jun 8, 2013 |
#
?
Jun 8, 2013 15:30
|
|
|
|
| # ? May 16, 2024 16:44 |
|
|
At first glance I would have said that was Reyes rendering's polygon smoothing but it doesn't look like Art of Illusion does that. Are you using a displacement effect or something? Taking a guess that you should probably bevel/chamfer (depending on which it's called in Autocad) the outside edge(s) of the object. Typically that's what you'd have to do for Reyes rendering and this looks like a similar problem, whatever's actually causing it. This will add a fair number of polygons but ... What's your purpose for this particular model? Is poly count a really big priority?
|
|
#
?
Jun 8, 2013 18:40
|
|
|
Flesh Forge posted:At first glance I would have said that was Reyes rendering's polygon smoothing but it doesn't look like Art of Illusion does that. Are you using a displacement effect or something? I was going to update this. You are basically absolutely right. When I went into the mesh editor in AOI, it had different options for mesh smoothing. Selecting "none" fixed that problem. This is a modular motherboard mount.
|
|
#
?
Jun 8, 2013 20:52
|
|
|
Oh, I guess Catmull-Clark subdivision could do something like that too but generally it does not look like a bunch of anuses. What is probably causing all that ugly rippling is topology that is not suited to subdivision or for Reyes poly smoothing. Sometimes you will get "rippling" surface like that from Phong shading (sometimes called "smooth shading" in various apps) but generally you will see that right away in preview rather than only in final rendering. If you're OK with how it works with your mesh smoothing option off, great. If you want to know the basics of why topology is important to being able to predict how your meshes will behave under subdivision, watch this thing (actually watch all of them from this author if you have not): http://vimeo.com/2158706 essentially your model is a big gob of triangles, you want it to be all quads wherever possible when working with subdivision.
|
|
#
?
Jun 9, 2013 00:53
|
|
|
And now somebody can pick apart my own character model, fair is fair. Still need to redo the feet and the topology in the extremities. I spent a lot of time on the head and face, adjusted the leg pose to are more traditional "attention" pose (which exposed a bunch of proportion problems, don't model in X pose unless you really have good references) and redid the hands. The face looks a little derpy in wireframe but I think eyelashes and texture will be a lot more appealing. Poly count is currently 5712, wireframes are subdivided 1x, renders are subdivided 2x. http://imgur.com/a/Jofem/all            
|
|
#
?
Jun 9, 2013 08:49
|
|
|
Got a question for the Max+vray guys: I have a (shittily) organised 3d library of various types of foliage/trees at work, and I want to be able to place these as vrayproxies with materials attached. A bit like the Lumion library. Do you guys have any tips in speedily blocking out the greens (or any type of object really) in my max scene? I'm thinking in the direction of having a .max file filled with materialised proxies, and just sort of cherrypicking them, and instantly painting with random size/rotation, but how? No idea. Any library script I tried so far (project manager, HCG) do not seem to support vrayproxies. Another way could be: Have my library organised without proxies, and using a project manager script to put them in the scene, then painting with those, afterwards converting them all into instanced proxies. I'm trying to avoid any scatter plugins, but if they are the most efficient choice time/money wise, I'll pitch it to the cigars here at work. International Log fucked around with this message at 14:01 on Jun 12, 2013 |
|
#
?
Jun 12, 2013 13:12
|
|
|
I do 3D Arch viz, the place I'm at wants to put together a real time environment of the space we've designed to show to a client who happens to be very micro-manage inclined. The theory is, if we can walk them through the space and show them every angle, they can be happier with the results than our usual video presentation would leave them. Anyone have any insight into what would be the best software solution for this? Our intern is selling us on UDK, but it seems like a huge learning curve for a four week delivery. We animate in 3DS Max - so something that integrates with that easily is necessary.
|
|
#
?
Jun 12, 2013 15:40
|
|
|
Frankenfinger posted:I do 3D Arch viz, the place I'm at wants to put together a real time environment of the space we've designed to show to a client who happens to be very micro-manage inclined. The theory is, if we can walk them through the space and show them every angle, they can be happier with the results than our usual video presentation would leave them. Anyone have any insight into what would be the best software solution for this? Our intern is selling us on UDK, but it seems like a huge learning curve for a four week delivery. We animate in 3DS Max - so something that integrates with that easily is necessary. While I haven't used UDK (so I can't compare it), something like this would be pretty easy to assemble in Unity once you have all the pieces modeled and textured.
|
|
#
?
Jun 12, 2013 15:48
|
|
|
Anyone here have experience with 3D printers and the like? Our EP would like to do a test print of a model that I did, but we're having... trouble preparing the model properly. Does the model have to be watertight, no matter what? Is there an exporter out there that can properly interpret T-junctions and intersecting elements? What's the best resolution (polygon-wise) to work with? Is it better to simply print off a complex model in chunks and then assemble it after printing? We're all sort of in the dark when it comes to best practices on this one.
|
|
#
?
Jun 12, 2013 17:30
|
|
|
Hazed_blue posted:Anyone here have experience with 3D printers and the like? Our EP would like to do a test print of a model that I did, but we're having... trouble preparing the model properly. Try this thread. http://forums.somethingawful.com/showthread.php?threadid=3365193
|
|
#
?
Jun 12, 2013 18:01
|
|
|
International Log posted:I'm trying to avoid any scatter plugins, but if they are the most efficient choice time/money wise, I'll pitch it to the cigars here at work. Forest pro's library is amazing and the best one i've used by a distance. We have every evermotion plant/tree library and a few others loaded into it - making thumbnails and proxies is a couple days work - but now I can open the library, select 10+ plants, select the surface(s) I want them on and paint directly onto the mesh. You can pick individual models to paint, set groups within one forest system, use 30+ in one scene, then there's all the random controls for scale/rotation/colour tint. You will save so much more time than you didn't even realize you were wasting. Floating licenses too. Frankenfinger - back out now. real time is not a quick solution. it's a custom built thing that happens after every decision is made and takes twice as long as working to views/rendering out sequences. The quality will be significantly lower and you're about to give yourself so much work, just to please a client who wants to look into the corners. A firmer hand in meetings selling them on focused views will go a lot further. Lower poly count (or god forbid, unwrapping & creating normal maps), no GI, ugly reflections, no glossy effects, having to cover every angle instead of the ones you see. translation issues, re-creating materials. the list goes on. By all means experiment with it and see if you can make it work for you - but do not agree to do something in real time for a live job that has a deadline when you've never done it before. Wasn't sure if it was gauche to post up a photo of my business card so I didnt - but the point I'd have been trying to make with it was that I got made art director last week, and I'm quite pleased about it... cubicle gangster fucked around with this message at 18:33 on Jun 12, 2013 |
|
#
?
Jun 12, 2013 18:20
|
|
|
While real-time engines are certainly becoming capable of doing most of the things cubicle gangster says you'd lose, it would take you at least a year to test the different tech, choose what works, build a suitable and workable pipeline for it, customize it to your needs, and then set up the education and training needed to get everyone up to speed on developing for real-time. In short, forget it.
|
|
#
?
Jun 12, 2013 18:50
|
|
|
"People" "Training" - these are not words I understand. Seriously I'm one of four guys in our entire department, one of the two who actually does 3D viz in said department, plus our intern who won't be here forever. Word has come from our creative director that this is what we "need" to provide. I have to find a solution that I can pick up and implement in four weeks. I don't get the option to say no. The good news is the space is remarkably small, think convenient store size. I just need the application that gives me the smallest learning curve. I haven't touched Unity, but UDK is a nightmare to work with coming from straight 3DS Max.
|
|
#
?
Jun 12, 2013 18:58
|
|
|
It actually isn't a nightmare if you had time and training materials. I would recommend Unity if you need "quick and dirty". Just so you know, you are being set up to either fail or very painfully stumble across the finish line. Since you seem to be implying that your operation is a little smaller in size, I hope that your jobs have lower expectations, because you will not be producing an experience with the same graphical fidelity as arch vis renders. mutata fucked around with this message at 19:12 on Jun 12, 2013 |
|
#
?
Jun 12, 2013 19:07
|
|
|
Frankenfinger posted:I do 3D Arch viz, the place I'm at wants to put together a real time environment of the space we've designed to show to a client who happens to be very micro-manage inclined. The theory is, if we can walk them through the space and show them every angle, they can be happier with the results than our usual video presentation would leave them. Anyone have any insight into what would be the best software solution for this? Our intern is selling us on UDK, but it seems like a huge learning curve for a four week delivery. We animate in 3DS Max - so something that integrates with that easily is necessary. I'll second GFBeach's recommendation of Unity. I've been doing a lot of side research into real-time visualization since late February, and can tell you that it actually is very fast and easy to use once you learn the software and get comfortable with the workflow, even if the quality isn't as good. Unity can import FBX files from Max, and updating the geometry and textures is as simple as overriding the FBX file in the Unity project (since it imports anything in there automatically). Here's an example of what I got when I took a Revit file and put it into Unity (free) via Max. It's my own hosting and the file is relatively big, so please be careful: http://www.davidwernerarch.com/MonasteryVW.html The whole conversion process took me only a few hours, and updating the model after a revision can take as few as five minutes. I managed to amaze a couple of my classmates a few months ago by doing the whole process with their models while they looked over my shoulder. If you want to continue with Unity and have any questions, send me a PM or an e-mail to the address on my profile page. You're right that UDK has a bit of a learning cliff, and all of the problems that Cubicle Gangster and mutata brought up are very present in it. Don't listen to the intern. Edit: Hmmm... I should probably fix some of the mistakes in that model. Musical_Daredevil fucked around with this message at 19:30 on Jun 12, 2013 |
|
#
?
Jun 12, 2013 19:10
|
|
|
UDK is great, and I work in it all the time, but I've been learning it for a year now and I'm finally getting comfortable.
|
|
#
?
Jun 12, 2013 19:14
|
|
|
cubicle gangster posted:Forest pro's library is amazing and the best one i've used by a distance. We have every evermotion plant/tree library and a few others loaded into it - making thumbnails and proxies is a couple days work - but now I can open the library, select 10+ plants, select the surface(s) I want them on and paint directly onto the mesh. Thanks for the advice, and congrats on the promotion! You certainly deserve it. The only thing you have to do now is hover behind artists all day. 
|
|
#
?
Jun 12, 2013 19:37
|
|
|
Thanks  I was talking to my boss about it after... I thought it was quite funny that just as you become totally comfortable handling any kind of image, you stop making them. big laughs, lets play golf etc.
|
|
#
?
Jun 12, 2013 20:54
|
|
|
Frankenfinger posted:"People" "Training" - these are not words I understand. Seriously I'm one of four guys in our entire department, one of the two who actually does 3D viz in said department, plus our intern who won't be here forever. Word has come from our creative director that this is what we "need" to provide. I have to find a solution that I can pick up and implement in four weeks. I don't get the option to say no. Would Lumion suffice?.
|
|
#
?
Jun 12, 2013 21:50
|
|
|
Edit: Nvm, turns out I had to select "affect shadows". I also failed to mention this was in vray.
Ervin K fucked around with this message at 04:53 on Jun 14, 2013 |
|
#
?
Jun 14, 2013 00:32
|
|
|
Just because glass renders as transparent, doesnt necessary means the lighting thinks its transparent. Depends on the specifics of your lighting, what software you're using, how it's being rendered. Best bet without knowing any of these is to set your glass to ignore lighting information.
|
|
#
?
Jun 14, 2013 02:35
|
|
|
drat. I'm trying to remember the name of a tool in 3DSMax and I can't for the life of me recall it. It's an operation that lets you overlap two different meshes, and then using a paintbrush, you can alter one of the meshes by adapting its topology to fit the exact shape of the other mesh. Anyone know what I'm talking about? edit: Conform Modeling, in Graphite. Found it. Hazed_blue fucked around with this message at 18:07 on Jun 14, 2013 |
|
#
?
Jun 14, 2013 18:03
|
|
|
Is the consensus still zbrush > mudbox? Just got a surface pro and keen to try out. Softimage user.
|
|
#
?
Jun 15, 2013 12:06
|
|
|
Zbrush has an amazingly hostile user interface. I've had it since Zbrush 2 (several years now) and still have not gotten much use out of it beyond the most basic painting of displacement/bump maps. I never use it for creating mesh at all, I stick to Cinema 4D for that (also for texturing really).
|
|
#
?
Jun 16, 2013 03:19
|
|
|
OTOH, I create everything inside Zbrush now from meshes to displacements and textures. Its so dreamy. 
|
|
#
?
Jun 16, 2013 03:38
|
|
|
Game industry character artists pretty much universally work in zbrush as their primary package these days. Zbrush has the higher learning curve, so I'd learn that one. You can pick up Mudbox in a weekend if you had to.
|
|
#
?
Jun 16, 2013 04:57
|
|
|
Yeah Mudbox is really simple especially if you have experience with Maya. The newest version has retopology tools though which is cool, and makes it more useful. ZBrush is the industry standard though. Here's an overview of how Zbrush was used in the production of Wreck It Ralph: http://www.zbrushcentral.com/showthread.php?178524-Wreck-it-ralph!!!
|
|
#
?
Jun 17, 2013 02:07
|
|
|
I'd say MudBox is more user friendly, it is alot easier to get into than ZBrush. The UI, hotkeys and general workflow will and should be very familiar if you have used 3ds Max or Maya in the past. It is, after all, an AutoDesk application. As for Zbrush being an industry standard..? Could be, idk. You can get pretty much the same results from both... I'd try them both and see which I feel more comfortable with. cptsuicide fucked around with this message at 18:44 on Jun 19, 2013 |
|
#
?
Jun 19, 2013 18:40
|
|
|
The Zbrush controls and initial intuitivitytvity isn't as good as mudbox if you're a max user but I found that once you're over that hurdle it's just such a more fleshed-out and dare I say intuitive package.
|
|
#
?
Jun 19, 2013 18:51
|
|
|
Can someone help me out here? I'm working in Maya 2013. I'm trying to connect the edges of the cylinder with the face closest to them on the boxy thing. What would be the best way of going about this? I've tried the Make Hole tool, and that sorta works, but it just makes an ngon around the edges of the cylinder and the square, which I don't really want because this will most likely end up in a game. soapydishwater fucked around with this message at 20:53 on Jun 19, 2013 |
|
#
?
Jun 19, 2013 20:44
|
|
|
Does it have to be watertight? I'd just intersect the cylinder into the boxy thing and call it done, most likely. Add some bevely ironwork detail if you like/budget allows/that even suits the style of the lamp post. If you really need it welded together though, I'd line up the end of the cylinder with the nearest quad on the boxy thing(what even is that?), delete the quad and the cylinder cap if there is one, then rebuild it all with Append Polygon. You'd probably want to add another loop or two on the boxy thing to avoid a billion fussy triangles where they connect but... yeah. That's one way to do it I guess. That looks like a lot of geometry on the cylinder compared to the boxy thing, by the way. I don't know what this lamp post is supposed to look like in the end but the difference in density is a bit strange.
|
|
#
?
Jun 19, 2013 20:54
|
|
|
Whelp, DD Venice beach is shutting down. Most work will be done in Vancouver or whatever city they can get the best tax credits in. http://www.occupyvfx.org/?p=252 Looks like my next gig is going to be in not-LA, question is will it still be in the US.
|
|
#
?
Jun 19, 2013 21:00
|
|
|
floofyscorp posted:Does it have to be watertight? I'd just intersect the cylinder into the boxy thing and call it done, most likely. Add some bevely ironwork detail if you like/budget allows/that even suits the style of the lamp post. Yeah, I know it looks pretty weird right now  I plan on making the boxy thing the actual lamp part of the lamp post, it's just really simplified into the general shape right now so I can get positioning right.I'll just put the box onto the cylinder, I suppose it doesn't have to be watertight. As for the density part, I think I may just redo the post... I was kinda dumb when I made the base cylinder and it ended up a little bit too high poly, if you consider that there will most likely be a lot of these lamp posts in one scene. soapydishwater fucked around with this message at 21:52 on Jun 19, 2013 |
|
#
?
Jun 19, 2013 21:47
|
|
|
I made some Metal Gear Rising Fan-Art, thought i would cross-post it here as well. It's rendered in real-time in Marmoset Toolbag. Enjoy!    and here's a shot with the wireframe: 
|
|
#
?
Jun 19, 2013 22:48
|
|
|
soapydishwater posted:Can someone help me out here? I'm working in Maya 2013. Just incase you want the watertight version.
|
|
#
?
Jun 21, 2013 18:39
|
|
|
Just an FYI too, that cylinder has way too many edges for what will be an asset in game, unless it's being used as a high res to bake down. In which case you'd make the cylinder more like a elongated cube with the same amount of edges as the boxy thing you want to attach it to. And when you subdivide it (hit 3 on your keyboard), it should give you a round tube. If you need to keep it having that many edges but keep the boxy thing that low, I'd just intersect it.
|
|
|
#
?
Jun 21, 2013 20:33
|
|
|
Crosspostin' The new york times wrote an article on the 432 film. http://www.nytimes.com/2013/06/23/realestate/selling-park-avenue-condos-at-250000-a-minute.html?_r=0 There were 3 of us producing cg, and one person managing it and filming (keith, mentioned in the article). Shots were split 50/50 between me and another guy and then we had one other person to fight over as help. CG team was small, but because filming spread the whole production out over 6 months we could get away with it and get everything done ourselves. The film wont be public for a long time yet - probably after most of the units are sold/the building is completed, but I'll be posting it as soon as it's available. We filmed a making of during the whole process too, so they'll likely come out together. edit: just realized, in that top shot with philip petit in the greenscreen studio, you can see me in the background. just to the left of the ladder, white t-shirt. to the left of me is matthew, then gabor (the other cg guy), keith & harry macklowe in the blue. cubicle gangster fucked around with this message at 05:11 on Jun 23, 2013 |
|
#
?
Jun 23, 2013 05:07
|
|
|
I'm doing my first 3D job for actual money and since I started lighting it, it's been like this: 1) make light 2) preview 3) too bright 4) emit decay 5) too dark 6) bump up power 7) render 8) still looks bad; sure am glad a crap-grade preview render took fifty hours 9) tweak 10) render again 11) welcome to hell. I basically just came here to say gently caress Maya. 
|
|
#
?
Jun 23, 2013 15:26
|
|
|
Then you learn the value of rendering everything into seperate passes and comping in photoshop/aftereffects
|
|
#
?
Jun 23, 2013 15:28
|
|
|
|
| # ? May 16, 2024 16:44 |
|
|
SynthOrange posted:Then you learn the value of rendering everything into seperate passes and comping in photoshop/aftereffects I tried doing that, but it was still taking forever (as opposed to jesus loving christ god drat this is taking forever, which was before I tried that). Unless I'm doing something else that's bad and unnecessary and causing slowdown, which is completely possible, I have accepted that I'll have to figure out how to queue renders and just leave it alone for a while or something.
|
|
#
?
Jun 23, 2013 16:10
|
|