|
Here's a thing I did recently, mostly for texture practice, though I also got to wrestle a little with my eternal nemesis graphics design in making the logo and that picture. Textures are what I've got a question about. That's my second stab at a texture that's gone far enough to actually get finished and I'm still doing stuff in a very basic way. I've just about got a handle on what to do with the various maps in a PBR environment (I use Blender/Cycles). At this point I'm just painting my texture maps manually and not using any fancy tools except NDO for normals. That's been good and comfortable for figuring out the basics, but I'm at the point where it's starting to get annoying to go manually color stuff appropriately on three different maps when I want to add a little metal detail or whatever - especially since I know there are better ways. I've been looking at DDO (which I'm biased towards over Substance because I'm already using NDO and it doesn't cost a billion dollars) but from where I stand it looks like a sea of massively complex features. Substance looks even more imposing. Are either of those something you can gradually delve into? Right now I'm really just liking the idea of material ID maps and I don't really want to start messing with super fancy procedural materials and stuff. Can I use DDO and keep drawing my own stuff while being able to assign basic material properties (diffuse, roughness, metallicity/spec) in a nice way without having to get into all the rest all at once, or is there a better tool for this sort of thing? Am I even thinking right for what my next step should be?
|
# ¿ Jul 27, 2017 01:43 |
|
|
# ¿ May 14, 2024 12:19 |
|
Synthbuttrange posted:How is this is only useful if you have a point of reference. Like it's probably fine, not very stylized and would plop into a low poly game okay but absolutely wont cut it on a AAA game. Kanine posted:this would be right at home in homeworld or an appropriately 70's sci fi book cover <3 e: Posted while I was posting: mutata posted:To answer your question quickly, both can do what you're talking about. e2: vvv Oh... Well, the spaceship is still a pretty low poly model and the materials aren't that fancy. Elukka fucked around with this message at 04:00 on Jul 27, 2017 |
# ¿ Jul 27, 2017 03:30 |
|
I'm very ad hoc about it. Often I'll merge and split stuff several times in the process of modeling something, just by what is most convenient at a given moment. I can't really think of any hard and fast rules for doing it one way or the other. Sometimes you'll need things to be made out of continuous mesh because there'd be ugly intersections otherwise or you need it to deform or something, but for example on that knife it would be fine either way. Just do what feels good and you'll get a knack for it. e: So a week after I say I have no professional aspirations someone wants to contract me. I haven't talked to them yet, so I have no idea how serious this is, but I'm not against the idea in general. I don't know anything about how any of this works though and I need to fix that. I don't know what I need to know, I don't know how to price it, I don't know anything because I haven't given it any thought. I imagine there's probably posts about this around here and I'm just bad at finding them. Elukka fucked around with this message at 05:25 on Aug 3, 2017 |
# ¿ Aug 1, 2017 03:20 |
|
Taffer posted:Which is why it's great. Not a joke - it's very unique (what 3d software isn't, though?), and can take some time getting used to, but after a little bit you realize how quickly it allows you to get things done and how flexible it is. I've been using it for years and I'm still regularly fighting against it and getting annoyed by bad UI design choices but I'm too deep into Stockholm syndrome to use anything else. I've also got a friend who swears it's great but actually also spends time fighting it all the time but somehow just doesn't notice.
|
# ¿ Nov 11, 2017 02:18 |
|
Working on this spaceship. The more I figure textures out the more laborious my process gets. I've got a question. I'd like to do animated visual effects for these things - engine plumes, explosions, that sort of thing. For an idea of the kind of thing I'd want to accomplish, it'd be something along the lines of this https://www.youtube.com/watch?v=GIg6pWwezEU. I'm working with Blender/Cycles. My problem is I have no clue on how visual effects are actually done, so I don't know where to start - I don't know what to google. There are a lot of highly specific tutorials for flames and smoke and stuff but they don't seem directly applicable to the kind of thing I want. Maybe there's some straightforward way to do this? Otherwise, I'm willing to spend some time learning, but I don't have a starting point for that either - my usual problems with 3D tutorials is that they tend to be "here's how to do this super specific thing" rather than "here's how to gain an understanding to solve the problem" and I have a hard time applying them to anything but the exact thing they're about.
|
# ¿ May 18, 2018 18:56 |
|
cubicle gangster posted:There was a big shitfit on cgtalk back in 2013 because someone went to buy it and after they looked at his portfolio they refused to sell it to him - after a bunch of people joined his side they went through some portfolios of users and revoked their licenses. It still happens occasionally. 3dcoat posted:We would be really happy if 3DCoat could be used to create beautiful works of art to inspire and bring joy, and not provoke hatred, aggression to people, wizardry, witchcraft, addiction or carnality. We are mostly Christian team, so this question is especially sharp for us because we know that God's law treats hatred as murder and infidelity in mind as real adultery, and consequences of our sins influence the whole life. This is definitely a huge warning siren not to touch this software. Better be careful with fantasy stuff too, wouldn't want any witchcraft in your models! What on Earth.
|
# ¿ May 31, 2018 23:27 |
|
I know I'm probably late to the hype party but goddamn it's cool to actually see your model properly in real time in the viewport with Eevee. Dev version crashes constantly for me though so I guess I'll best wait for a release. Also, wip texture, still learning as I go, still ponderously slow. Been dipping my toes into fx too: https://gfycat.com/AjarSplendidDotterel I wish I could make the smoke actually properly expand like it would in a vacuum. Elukka fucked around with this message at 10:36 on Jun 25, 2018 |
# ¿ Jun 25, 2018 10:33 |
|
Nodes are kinda laborious for setting up basic stuff (but Blender now has the principled node which basically gets you your basic setup in one node) but they're a godsend because they allow you to set up any drat thing. Whenever I look at a node setup though I think "what on earth is going on and how could anyone even come up with this?", including when it's my own.
|
# ¿ Jun 28, 2018 09:19 |
|
Does anyone have experience with https://www.sheepit-renderfarm.com/? Basically, you get credit by rendering other people's stuff, and then get your stuff rendered by the farm, and the advantage there would be that you could build up credit with your idle time and then get stuff rendered faster when you have something to render. It sounds pretty nifty on paper and I'm curious if it's as good as it sounds.
|
# ¿ Sep 23, 2018 18:06 |
|
Don't expect there to be a good reason for everything, some things just are bad. Like the outliner, but I believe that's being redone in the next version. The transform orientation system (which enables you to work on arbitrary axes, kind of) is my favorite example, because it's supremely clunky and has either had the same bug for years or works as intended with some inscrutable logic. The reason you'd want to learn Blender is that you want its capabilities, which are many, but it does have its flaws. Elukka fucked around with this message at 17:26 on Oct 18, 2018 |
# ¿ Oct 18, 2018 17:12 |
|
Comfy Fleece Sweater posted:Feels great to be struggling to model some really basic version of a popular thing and then a pro artist is featured on ArtStation doing *the exact same thing* light years better with comments like “today’s doodle! Had a couple of hours to myself, modeled while watching the new Walking Dead episode”
|
# ¿ Jan 2, 2019 21:19 |
|
And here I was just looking into getting into Substance. This makes me reconsider. On the other hand, I'm currently using Quixel who themselves are doing suspect things (and also recently showed that they can accidentally and presumably deliberately disable their software remotely) and which is literally a Photoshop plugin. The stuff going on with Quixel sounds really bad, by the way. They had a website problem which somehow completely disabled Quixel Suite. They eventually fixed that, but at the same time, they removed all mention of the product or the license you own from their website and account system. Their shop still sells it though.
|
# ¿ Jan 23, 2019 21:56 |
|
Gearman posted:- The biggest concern for most people is probably just pricing model. Sebastian has said that the SaaS (monthly payment as opposed to perpetual license) is a model that works really well for them. Obviously, Adobe shares similar thoughts. Is likely that the "one price and you own it forever" offering is gone for good. A bit of a shame, but as long as the monthly cost is still affordable, I think the many positives outweigh it (monthly payments are tax deductable right away, you can cancel if money is tight, much lower up-front cost which is better for indies, students, etc.). Perhaps now would be a good time to buy a perpetual license for as long as they exist, but I'm not sure if I should trust them not to pull tricks with existing licenses.
|
# ¿ Jan 24, 2019 17:52 |
|
Sometimes Blender is more robust than I expect. I wanted to figure out realistic lighting for a space scene and had the mildly insane thought that I should just model it in actual scale and see how it looks. So I got a spaceship a few hundred units in size in the foreground, an actual sized planet out of frame, and then a moon in the backround that is 300 million units away and 9 million units across that the planet happily bounces light at. Nothing breaks. It renders fine. Maybe that's to be expected, I dunno, but I absolutely expected it to explode into a cloud of shrapnel made of floating point errors.
|
# ¿ Jun 2, 2019 03:56 |
|
A render I made. The model is a bit long in the tooth, but I felt like using that one. I actually just meant to test whether you could reasonably see spaceships at realistic separations between each other with some reasonable camera, and it turns out the answer is yes. Then I got carried away making it look neater and learned a bunch of things in the process. A 1000 mm focal length lens should be able to snap that image in reality. The foreground ship is 10 km from the camera and the next ones are 100 km, 200 km and 900 km away. The body in the background, a habitable moon, is about 300 000 km away, and is lit by the same planetshine as the ships, only fainter since it's so much further away. The lighting is about as realistic as I could make it. Besides the little lights on the ships, the only actual light source here is the sun. Here's how the scene is laid out:
|
# ¿ Jun 5, 2019 11:49 |
|
Ambient light is zero. The lighting on the bottom is sunlight reflected from the planet. It's also what's lighting the dark side of the moon, it's just much fainter there because it's 300 000 km away. The top side of the ship, the parts that aren't in direct sunlight, would be totally black if it wasn't for a lot of very intense sunlight bouncing around. The top radiator fin for example is lit by the sun bouncing off the hull. e: For reference, similar lighting in reality: https://upload.wikimedia.org/wikipedia/commons/9/9f/Earthshine.jpg Elukka fucked around with this message at 17:41 on Jun 5, 2019 |
# ¿ Jun 5, 2019 17:37 |
|
It's a pretty wide dynamic range but I tried to keep it to something a camera might sanely see. I'm not a photography expert though so some guessing is involved. The top side of the ship, directly lit by sunlight, is a few hundred times brighter than the bottom part. I based the stars on shots like this: https://i.imgur.com/coP2Z2y.jpg That's exposed for direct sunlight, but it's still got faintly visible stars in the background (gotta view at full size to see them though). I figure mine's a somewhat longer exposure since I've got a white spacecraft too and its directly sunlit part is overexposed. e: Now that I think about it, the difference between Apollo photos and ISS photos can be explained by one being shot with 60's-70's cameras and the other by modern DSLRs. Elukka fucked around with this message at 21:58 on Jun 6, 2019 |
# ¿ Jun 6, 2019 11:55 |
|
It turns out Eevee gets very high exposure shots wrong in a very fun way.
|
# ¿ May 14, 2020 02:30 |
|
Also did some renders because I wanted to see how bright the five megaton blasts from my spaceship's drive might be. It's based on a whole lot of guesses, though a real living plasma physicist was involved in making them. The glare settings are consistent between the first and second image, though in the first the actual flash is off-frame because it would completely blind the camera. All you're seeing is reflected light from the ship. The second one is from 10 000 kilometers away. In the raw render, the flash is a single pixel with a value of 500 000.
|
# ¿ May 14, 2020 06:19 |
|
Is Blender's animation system kind of a mess, or am I just using it wrong? I have things made out of multiple objects which need multiple animation cycles, and I end up with this huge mess of strips in the NLA editor where I no longer have any idea what is what. I really don't have a solid sense of what the intended workflow is. Also, tiny little rocket effects that are almost too fast to see:
|
# ¿ May 25, 2020 11:15 |
|
Further space experiments. The image and the inset are the same scene rendered with different exposure and field of view settings, with the ship 10 000 kilometers away in both. The dot of light in the larger image is actually the model you see in the inset. The stars are also still there in the inset, you just don't see them because it's exposed for sunlight. I'm surprised it just works.
|
# ¿ Jun 8, 2020 02:26 |
|
In Blender in principle you can use any size scenes, but past some point floating point errors start being problematic, as will clipping planes if you want objects across extreme distances in the same render. I've still worked with scenes hundreds of millions of units across and while they have some issues it's mostly been fine.
|
# ¿ Jun 8, 2020 03:33 |
|
BonoMan posted:These are cool but they are just far enough apart in posting that I kind of glance off of them. Any chance you could do a write up of what you're doing with all of your results aggregated? It seems interesting to me!
|
# ¿ Jun 8, 2020 04:49 |
|
I have opinions about the GPL license and the ideology behind it and they are not all favorable but goddamn I'm glad Blender is GPL. It makes it completely toxic to companies that might try to take it over and close it up, which is great because everything else in the 3D realm inevitably gets snagged up by one of the few big relentlessly customer-unfriendly companies seeking to extract a maximum amount of money out of you on a continuous basis. At one point I was going to give Substance a try, and as a bonus I would be rid of Adobe. Right then Adobe bought it and that put a considerable damper on my motivation.
|
# ¿ Jun 25, 2020 04:01 |
|
On a jury rigging note I had absolutely no idea about animation (I only had a foggy idea of what a bone was) but I managed to make this magnetic rocket nozzle fold in on itself. Some IK deal tied to empties where the struts attack to the rocket body and then a rotation and probably some other iffy things I forget. The setup isn't pretty but it folds. A little bit of complexity in making the blades on those struts fit in. Those are a thing because for physics reasons every surface has to be at a very low angle to the center of this thing where an intense nuclear reaction happens. https://cdn.discordapp.com/attachments/725631347594362897/760199310738194502/0000-0240.webm I have a slightly better idea about animating things now. There are some other mechanical things I wanna do. Fortunately none of them are too complicated.
|
# ¿ Oct 7, 2020 18:11 |
|
I'm experimenting with ways to render things like rocket plumes and space explosions (general theme: gas in vacuum) in a fast enough way in Blender. The gas sim really doesn't want to behave as if it's in vacuum - there's some inbuilt assumption of atmosphere I cannot work around, which is annoying. Currently I'm playing around with particles rendered as volumetric objects. What's that? Eevee can only render volumetrics as cubes? No matter. https://i.imgur.com/63nPLo7.mp4
|
# ¿ Oct 24, 2020 11:13 |
|
ImplicitAssembler posted:Any fluid sim is based on Navier-stokes and assumes atmospheric pressure, so yeah, you'll have to cheat it. I thought about turning particles to a volume, but couldn't find a way to do it. Blender supports creating volumes from OpenVDB datasets (and in 2.91, meshes), but I didn't find any way to turn a particle sim into one.
|
# ¿ Oct 24, 2020 22:23 |
|
I could do that, and it would work for stills, but I don't see how it would work for animations.
|
# ¿ Oct 25, 2020 00:02 |
|
ImplicitAssembler posted:You can transfer velocity and density from the particles to the geo and then to the volumes, no?. I was looking at OpenVDB and it has all kinds of incredibly useful features that just aren't currently implemented in Blender. Like turning particle systems into volumes.
|
# ¿ Oct 25, 2020 00:26 |
|
SubNat posted:(Though that's likely not relevant for you right now, I think blender just supports the optiX denoiser, but not any usage of RT hardware directly/explicitly?) e: Yeah, apparently it does since some time last year: https://code.blender.org/2019/07/accelerating-cycles-using-nvidia-rtx/ Elukka fucked around with this message at 18:48 on Dec 11, 2020 |
# ¿ Dec 11, 2020 18:45 |
|
Same, I learned like last week that OptiX was also something other than a denoiser. It is really cool how fast renderers adopted it. Now if I actually had an RTX card... I'm gonna see if used 2080 prices go down once the 3000 series is more available.
|
# ¿ Dec 11, 2020 19:56 |
|
Messing with lighting and some simple post processing a bit. This is some reasonably realistic space lighting, assuming there's a planet off frame to the left. I really need better bloom and glare than what Blender can do. Apparently Natron's got me - I'd just have to learn it. For fun, here's how Hubble should see this ship from 20 000 kilometers away. It turns out Blender's field of view has a hard lower limit way above Hubble's fov, but the same quantity expressed in focal length has no upper limit. Elukka fucked around with this message at 12:58 on Feb 26, 2021 |
# ¿ Feb 26, 2021 12:55 |
|
There's a big global chip shortage that might take a year or two to resolve. This will probably suck for a good while.
|
# ¿ Jun 4, 2021 03:03 |
|
I had the idea of scripting some animation stuff and produced some of the hackiest code possible. I'm defining some RCS thruster locations on this ship with empties so that I can use a script to actually fire them and not have to artisanally place a billion strips in the NLA editor every time I want a thruster to fire for half a second.
|
# ¿ Aug 11, 2021 23:29 |
|
Got my thruster script working. It just fires them randomly at the moment. https://i.imgur.com/DWA9mRm.mp4 Elukka fucked around with this message at 05:43 on Aug 14, 2021 |
# ¿ Aug 14, 2021 05:33 |
|
500 posted:These look really cool! Are they cone geometries with a custom shader? The shader has a gradient controlling opacity along the long axis, and additionally opacity is controlled by the dot product of the camera vector and the face normal to fade it out along the edges, so that a face is at full opacity when the camera is perpendicular to it and at zero opacity when the camera is parallel to it. There's also some moving noise. Pinching the mesh in places like I've done here introduces some more complex shape to the plume. These are dirt cheap to render with no volumetrics involved. I think they'd work really well in a game too. I haven't seen anyone use plumes like this but the shader should be easy to implement anywhere. There's one major downside, and that's that they also get less opaque if you look along the engine's axis. I tried to do some vector math projection tricks to fix this but couldn't quite figure it out. There's another effect I'm trying to figure out how to replicate in Blender, but I have no good ideas. This is a real world video of a rocket firing thrusters in short pulses: https://i.imgur.com/ffDsKZr.mp4 Essentially, a thruster emits a cone for as long as the thruster is firing. If it fired for a longer period, you would get a cone that propagates at a finite velocity and flies off when the thruster shuts down. If it fires in short pulses like there, you get these cone segments. Using particles seems impractical because there would need to be a huge density of them to get the smooth look, and Blender's particles also have a really weird and arbitrary 1 km/s speed limit, which is too slow. In Houdini, I think I could emit particles and turn those into vdb volume density, and then export the vdbs to Blender. However, then I could no longer use my scripts in Blender to control my thrusters, so I'd really rather keep it witin Blender. Incidentally, the free version of Houdini lets you run simulations and export volumes and volume animations with no restrictions. Elukka fucked around with this message at 11:49 on Aug 14, 2021 |
# ¿ Aug 14, 2021 11:38 |
|
500 posted:I was following until this part here. I don't think I know what you mean by 'engine'. Do you mean when you're looking at a single plume straight down its axis, it fades out? 500 posted:You couldn't do, like, a VDB sequence, and spawn that with your script? I actually haven't done any scripting or shaders yet in Blender, so I don't know what the limitations are.
|
# ¿ Aug 14, 2021 14:48 |
|
500 posted:
I made a slightly nicer test render with some space noises. https://i.imgur.com/neRBzJR.mp4 Elukka fucked around with this message at 04:32 on Aug 22, 2021 |
# ¿ Aug 22, 2021 04:30 |
|
I've been looking into texturing tools. As it is, I'm not using any dedicated texturing software. I'm just using Photoshop. Painting every map separately is not ideal. I make this kind of thing: The textures do all of the heavy lifting there. The bare model looks like this: I intend to start doing higher poly work, but I will still need to draw things. A significant chunk of the work that goes into this is drawing a ton of custom tiny labels and panels. I've huffed a lot of Homeworld fumes. That means I get limited mileage out of procedural and scanned textures. A little, but it's not something I want to rely on. I tried out Substance Painter (...I would like to get rid of Adobe though) but its basic image editing tools are limited. You can't make selections, transform, or copy stuff around, among other things, at least not in an easy way, and from looking at how people work in it, doing something like a simple label involves going to a separate image editor and then importing that into Substance. I can think of no reason Substance shouldn't have these basic tools - you could just as well be drawing on the texture right there and not need a separate application. Since I'm doing a lot of this, Substance feels like it'd be very clunky to me. I'm guessing the clunkiness of this is why you see models using default Substance alphas even when that's maybe not ideal. I'll still give it a proper try later. I also briefly checked out Quixel Mixer but it seems even more centered on slapping premade materials on models to the point that the only painting tools are "brush" and "eraser". So far I've found two pieces of software that seem to do what I want, i.e. combine multichannel 3D/2D painting with a solid set of basic drawing tools: Mari, and, uh, Procreate, which is a mobile app exclusive to iOS. The latter is right out. Mari has a free noncommercial version with technical limitations that don't really affect me, so it's an option, but the commercial version is unattainably expensive for me and although what I make is generally noncommercial it would bar me from ever taking commissions, which is not ideal. It's also scary and imposing, but I can probably deal with that. I'm curious if I've missed any options, workflows, or something else, or if I'm thinking about this all wrong. I figure not a lot of people end up in a place where their two options are an ultra-heavy duty VFX software suite and a mobile app. Elukka fucked around with this message at 13:57 on Dec 15, 2021 |
# ¿ Dec 15, 2021 13:52 |
|
|
# ¿ May 14, 2024 12:19 |
|
I still don't have a good sense on how much it matters for 3D printing to have a mesh that's one piece. That time I tried one out on Shapeways it didn't matter at all. Floating bits all over the place but it was fine as long as they intersected. I guess it depends on the specifics of the 3D printing software?
|
# ¿ Apr 27, 2022 09:59 |