Today was just one of those work days where everything was going awfully, fucking pulled it back though.
Made a triplanar material and fixed the blending issues most tuts/resources I found on the subject seemed to produce, next step is to edit the blend so it isn't just a gradient.
I dunno, just had a good day in the end!
I'm making materials on monday so I'll post them and the results of my triplanar stuff then.
I don't know if its anything I could be doing wrong but tri planar projecting almost never works for me and just produces a really low res grain across everything.
https://www.youtube.com/watch?v=sFSTfAOOZ9M
As everyone has mentioned above, just thought I would drop this in here as well for some further explanations etc.
N-gons ain't bad.
Yeah, Arrimus3D (another 3ds max channel) also said that.
It depends... Is it a high poly to low poly baking workflow? Or will the model be used in an animation? Maybe games?
Well, I can safely say that in many cases it will totally be acceptable.
Would not generally recommend ChamferZone or Arrimus honestly. Both teach a lot of bad practice.
Who would you recommend?
I've been doing amateur models for a while now and every time I end up making low poly. Been trying to get to one of those "medium" poly assets that still do look great but with not many faces. How do you guys go about that? Am I missing something?
For example a gun I made:
https://sketchfab.com/models/7fafccf409004024939ca454687679d0
And then the same gun but "medium" poly:
https://sketchfab.com/models/2e1841c7f3e042c98e201f6e7947c763
If I go to model inspector and turn on wireframe I can see that it doesn't even consist of so many poly's, yet the model looks amazing and well done.
High-poly, bake normals, decimate/retopo.
Baking normals and retopologizing aren't things I have done before. Will look into, sounds interesting.
https://upload.wikimedia.org/wikipedia/commons/thumb/a/a8/Sega-Saturn-Console-Set-Mk1.jpg/1200px-Sega-Saturn-Console-Set-Mk1.jpg
Yeah about that
Not any one person. Instead I'd recommend watching various videos of really really good artists at work and focus on what they're doing with the objects. Ignore the software and its tools entirely. Watch some of the video and written documentation on how to manipulate the software you're using, and toy around with it. Do note that this isn't what XxxYoMommaxxX has to say on Substance Designer, go to the Allegorithmic youtube channel and get it straight from the source. Everybody has videos these days if that's your gig, but read a lot too.
Watch through at least one newish tutorial of the entire pipeline of making an asset (but dont be surprised if all of it turns out to be bad lazy information, you just want to get an idea of the workflow). You'll pick up some terminology here, some specific to your software and some not.
At the same time as you're learning your software, do deep dives into the Polycount wiki and any other places you can dig up online for more specific info on the "hows" and "whys" of the stuff you're doing. The wiki has a ton of videos and tutorials embedded/referenced on most pages for practically any topic you'd need a primer on. However, a lot of the content that you'll be seeing on there is way out of date. And that's the key thing that holds true almost everywhere. Tutorials like anything else in the world are out of date the second they come out. Not only is the quality bar always going up, but the techniques to make art are constantly evolving. Arrimus, CZ, Blender Guru, etc aren't hitting the modern information and certainly not imparting the theory that enables you to problem solve for yourself.
Instead you should be reading and listening to what people who are big and well respected in the industry are doing. There's a lot of them, and everyone has different ways of doing things. Some are out of date on some topics but produce stellar quality, some are on the bleeding edge of almost everything and always experimenting. Often they strongly disagree on things. Take it all in, and you'll be picking up more of the terminology and way of thinking than any youtube channel out there. Most pros are sharing their stuff on Artstation and Polycount. Don't bother with paid tutorials.
And of course, pick an object to make and make it. Don't follow a tutorial and don't cheat by looking at other people who might've made the same thing. Do it from real life, using tons of references, to the highest quality you can. Post it here and/or on Polycount and get feedback.
So altogether: Watch people making great art (your eye and taste will improve over time), learn the ins and outs of the tools, see the basic pipelines and workflows in action, pick up terminology, keep up on what the pros are doing, always be making something and pushing yourself. Be self sufficient, cause nobody out there making crappy tutorials can lift you up into being a good artist.
Each time you do any of those things you'll get better at filtering out what you don't need and internalizing what's important. Be sure to participate in discussions too. Join our Discord if you haven't, and talk about what you're doing and share progress. Ask questions if you need, but don't be a pleb that can't use Google. The best answer to most question's you'll have is a few keywords away, so always attempt that unless you're genuinely stumped.
Did a thing
https://files.facepunch.com/forum/upload/1213/8d235612-289a-402a-8a4f-7351168cffd0/pooltoy.mp4
Material improvement
http://puu.sh/AGMwv.png
Looks great! Few tips to closer emulate the look of 3D printed objects based on observations I've made from my own prints:
The side walls on that drawer look paper thin. Most 3D printers have a nozzle size of around 0.3 to 0.5mm, which is noticeably much thicker. As a result, not as much light would get through even a single shell wall. Furthermore, single shell walls don't print very well (they start deforming and even just outright failing after only a few vertical layers), so they're generally avoided; I'd make the walls at least two layers thick like you did with the front of the drawers.
For the top layer of the front of the drawer, it looks like the two layers are just sandwiched together; by default every layer on a 3D print is a closed loop, so top layers generally look like cocentric circles until they get large enough where a crosshatching pattern is used instead, like so:
https://files.facepunch.com/forum/upload/217538/834b3287-04a3-4fd6-a9e8-194e1bbe60ac/Screen Shot 2018-06-17 at 5.33.30 PM.png
https://files.facepunch.com/forum/upload/217538/b0f7331c-f827-463c-bbc9-ae93f3f42071/Screen Shot 2018-06-17 at 5.32.41 PM.png
https://files.facepunch.com/forum/upload/217538/970e21db-80fb-4b8e-8b86-417be7922ca0/Screen Shot 2018-06-17 at 5.35.20 PM.png
https://files.facepunch.com/forum/upload/217538/623aa0c7-791a-4530-9845-eed596e78a09/Screen Shot 2018-06-17 at 5.36.42 PM.png
I think i should say that I'm pretty experienced with both the processes of 3D printing, and designing for 3D print.
The walls are 0.8mm thick (2 shells) and the actual printed model in real life transmits even more light through it than what the render shows.
Trnsmittance depends a lot on the filament, and this blue filament does let quite a bit of light through. (Most filaments do when printed this thin)
I don't know how to do top and bottom fill patterns procedurally. It would be very complex compared to what i'm doing, I imagine - so I won't do it for this simple 'product shot'.
https://files.facepunch.com/forum/upload/57926/d25dc8c8-1e44-4ef1-a20d-f65771bddc04/Output_v002.jpg
accidentally posted this in 3d prints lmao
big wip!
rubble, barricades, and the train off in the back are from turbosquid, commieball is @Stev , lookdev is about 50% there, gonna be texturing a bunch this week
Ooh, weird, okay well I'm back working on it today so I guess I'll have to see how it goes.
Any idea why it creates the grain?
the saturn could only draw quads due to its architecture being based on transforming sprites.
Does anyone know how I can restrict tessellation (not displacement) in ue4 to a masked area, as in, restrict the created geo to only the areas that its needed? Currently every part of the mesh is tessellating which isnt great, since its unneeded geo (visually) but also means im left with a lot of stretching and jagged edges.
In my case it would be, the parts of this rocky floor that extrude (white areas) and not the rest of the surface?
https://files.facepunch.com/forum/upload/157900/e11a1613-abbe-4d4d-b334-779831bf4b74/image.png
quads are tris because we describe them that way. the saturn is like an advanced mode7 from the snes. it takes in a texture as a quad and does linear transformations to get it into the right perspective.
https://i.imgur.com/q6CjAVg.gifv
just found this absolutely gorgeous gem which should explain it perfectly
That's insane
That is fucking insane, amazing the crazy solutions and things programmers do to achieve things.
Footage of Laura Croft entering the nth dimension.
"quads are tris for us only because we describe them that way" preach it!
That's awesome though, its such an imaginative work around, the history of the saturn is something I've never really looked into, going to check it out after seeing that.
A funny consequence of Saturn only supporting quads is that multiplat titles would often add an arbitrary vertice to triangles to make them into psuedo-quads so they wouldn't have to remake their models for the Saturn
That has a sort of cute, programmers/artists cross dev cooperation vibe, I like it.
Sonic R has some ridiculous programming for the Saturn, such as environmental mapping being done via creating a tri-rendering engine, with what the developer described as a "software version of the Playstation's hardware rendering".
Plus, the Saturn can't UV map. The Saturn's architecture as a whole sounds like an absolute nightmare to work with.
you also had to work in assembly to fully utilize the hardware.
Parts of the codebase being in assembly wasn't out of the ordinary for games pre-sixth generation though.
Sorry, you need to Log In to post a reply to this thread.
