Show us that cool new project you got in the works anon-kun!Previous Thread: >>1014090
Crow, baby crow, egg, nest
I'm pretty happy with how this one turned out.
>>1017692Sick
>>1017692Very interesting. What's it about?
>>1017712Just a Hell scene. Meant to be the more primal inner layer where all the souls that couldn't win slave jobs with the demon lords stay. Whether the slaves are luckier than these guys is debatable.
Made a jacket, gloves and glasses>>1017713NTA but that's brutal, these are souls not even demons wanna slave over so they just get nailed on a cross for all of eternity? I wonder what kind of people they'd have to be, perhaps people so bland in life that demons would rather not have around, incapable of amusing their twisted sense of humor
>>1017714Yeah, that's the idea. Kinda like the filth from Ultrakill, just so dull and bland that they aren't even worthy to be maltreated slaves. Although, considering the job positions (holding up the throne of a 2000 pound demon, being dinner, being a public toilet, etc etc), getting your charred half-dead corpse nailed to a cross on a lava lake doesn't sound so bad by comparison.
Started to block out an anime-style scene.
>>1017721Certainly looking authentic so far. Keep it up!
>>1017698here is the basic concept for it, instead of the tweak bones facing out, switch their direction so the tail its in the middle and then using 1 non-deform bone between the bicep and forearm that stretches and compresses i can get a compression scale for how close the forearm and the bicep are to each other, and then i copy that scale at various tiny amounts to the "tweak" bones to pull back the bones in the biceps/upper forearms, then using the same reference you can do x2 copy scales and isolate each axis to get compression on the X, and expansion on the Z to get the impression of the arm muscles working for example. then you need a second row of bones just slightly offset to counteract what the original arm bones are doing in some places. so as the original might expand on the X to bulge out the triceps, you might want to counter on the X on the inside to and compress for the deformation effect. then in the pit of the elbow i have x3 bones at the point of the arm/forearm, 1 parented to the arm, 2nd to the forearm, 3rd parented to either one and then copies the transforms of the other to get an average again, and the "counter" bones mentioned earlier that are closest to the center will track to the 3rd bone. This 3rd bone will also get a copy scale from the middle compressing to pull in the counter bones, play with the copy transforms/scale influence to create the curve you want when the arm compresses, at 0.5 it will create a straight line. For this part I also split the 3rd bone in 2 parts, the bottom one will reverse the "pull" effect but will have a scale limit to stop it halfway until the top part retracts and folds in again, this is to create the effect of skinny arms where the pit will stretch out before collapsing.the shoulder is simple for now, just a def bone pointing backwards to the collarbone/shoulder starting 1/4 of the way down the topmost arm bone, trying to make a mechanical clavicle nowI hope some of that made sense
>>1017767the backwards bone on the shoulder, doesnt have to be backwards it can be doing that in either direction, its a simple solution to shoulders
>>1017767oh and the elbow just copies the rotation of the 3rd bone but has a "flat" bone at the top because elbows are not pointy but create a flat "top" when stretched all the way, thats all that does by again copying the scale of the compression bone and isolated on the Y, bits on the end just point back to the either ends of the arm for smoothness
>>1017767and the same system works for the legs of course
>>1017767Wouldn't this be easier done with some corrective blendshapes rather than weightpainting all those extra bones?Genuine question
>>1017778there is no weightpainting, just obsessed with getting it correct with automatic weights
>>1017627Ok OP, but remember, you asked to see it
>>1017792I regret making this thread
Boolets.Finding the actual dimensions for some of these was surprisingly difficult and at times I seriously questioned whether I had nailed the right ones and wasn't giving some of these cartridge types completely wrong proportions. If any /k/ommando sees anything off feel free to give me a shout, but I'm fairly certain they got the right sizes.Also gadayum modeling the .50 BMG makes you realize how ridiculously huge they are, a fucking horsecock sized chunk of lead traveling at 1000m/s. Nutty.It was also interesting to discover the 5.7 is like a mini 5.56, you know those bottleneck shaped bullets which are normally used for rifles. It's nice when a hobby makes you get further insight on the real world, it's educational
>>1017721Progress. I still am unsure about a lot of things of this scene, particularly i don't know how am i gonna shade the mid ground objects (the rest of the office floor)
>>1017627
>>1017926The Eyes Are Dead And It Haunts Me (really needs specular highlights)
Working on my first gooner model lads. It took many attempts but it looks like I'm getting closer to a result I'm happy with
>>1017714I'd like to make a jacket as good as that. There's no arms on the base model other than the jacket, is there?
Got a boo boo? This'll patch you up [currently working on medical themed props, pretty satisfied with the results]>>1018014>a jacket as good as thatWhy thanks!>There's no arms on the base model other than the jacket, is there?You mean on the dude's actual body right, like under the jacket? That's correct, there are no arms under it (except for the forearms, there's still some of those left hidden in the sleeves, full hands too, so that I can remove gloves easily if I want to), I was inspired by Bethesda Games' way of handling clothing for their characters
Lava Bridge
might rig and make the B2 weapons, might not
>>1018050Looking good man
>>1017926Update
Made some ballistic tests for the shotties we're working on for our gaym, wanted to test shotgun projectile spread and the results were pretty interesting, anything above 5° rotation per individual projectile and the gun won't hit the broad side of a barn
>>10181640.5° rotation
My submission to a pwnisher style contest between me and my friends :)
>>1018233I would swap the order of the cuts so it fits the composition better
Tried making a gooner model
>>1017901Progress. Not sure if i'm gonna add more stuff or just detail what it is now.
>>1018239I think it would be good even if you stopped now, depending on what you're aiming for I would add a little bit of motion in the far background, maybe reduce the number of red lights and have them pulse, maybe cloud movement
>>1018247Thanks! There's supposed to be some flickering on the windows of the buildings, but it's not showing up, probably need to check the values on the textures.
>>1018239Awesome stuff man, looks great
>>1017636Cool
>>1018238If her hair was smooth I could goon to itAlso the wall and the floor seems to merge together, I would charge materials in the floor
>>1018050I like it, but the quick animation makes it difficult to see in detail
>>1018233Sick, gigachad fat fuck demon, lookin' good, he's in the same universe as >>1017692 and is Hell's cook>>1018237Shut the fuck up retard that's the most idiotic piece of "criticism" I've ever heard
I think I just accidentally added goatse to this texture I'm making.
>>1018239Looks amazing already. It has the quality of older anime, when the industry still cared and put in effort.
>>1018279>>1018262Thanks guys, i'll detail it a little bit more and post the finished version on wednesday probably.
>>1018266consider pausing?anyways I´ll post more later
>>1018294>consider pausingNah, I don't care that much
>>1018239Outstanding job
crunching down the textures for the retrofags
>>1018315>pixelating without palettizingWell on your way to being an indie griftware merchant I see.
>Mirror object>Mirrored part shading is all fucked up like the normals are flipped>The normals aren't flippedWhat gives bros?
>>1018319How did you mirror, by setting the scale to negative one?Post a pic?
>>1018273>Shut the fuck up retard that's the most idiotic piece of "criticism" I've ever heardWith the camerawork, it's unclear what he's even doing until the animation is over. It's bad direction
>>1018335I did it that way originally and flipped the normals and it still looked fucked. Then I did it with the modifier and it still looked fucked.I know the other part is fucked too where I haven't put a seam or anything, but that's not a big deal.
>>1018319>>1018344Check for internal faces
>>1018233The overhead fan does a lot for the lighting in the scene. Really I think everything it looks like you cared about is done well, simple assets could be more interesting like with the way this guy chops his cutting board should be beat up. But like I said that doesn't sound like your priority.I think you could add some light under the stockpot, either flames or an emissivie red coil and improve it with that easy enough though
>>1018360Totally should have done that, the stove in general looks really flat.>>1018273Glad you liked it :)>>1018237Originally that was the case, but then the knife flip into stabbing the board motion looked too jerky and non-organic if that makes sense? Probably just a skill issue but I couldnt make it look right with cuts going the other way.
>>1018359Nope it's not that. There's a really clear join even when the verts are merged and it's in solid view. I just can't work it out.
>>1018363Working on the character for the next challenge right now, hard surface modeling for the first time in like 600 hours of blender, shit is impossible. The armor is supposed to be inspired by insect chitin plates. Looks too thin for my liking but really not sure how to make it look bulkier, would love any suggestions or advice if you guys got any.Also tried making a Mignola shader (just a simple color ramp) but also combined it with the bsdf to preserve some of the metallic look of the armor plates, doesnt look the way I want it to. Is there a way to only mix certain properties of the BSDF through mix shader? I want just the roughness and metallic properties.
Look, it's furry Jesus!
>>1018239It's doneLossless version https://files.catbox.moe/wribv6.mp4
>>1018415Looks very nice!
>>1018415Love this
Does anyone know how this guy manages to adapt texture from drawings to character? Or at least where to start? It's very immaculate and seamless, I'm amazed. https://www.youtube.com/watch?v=6xqNecJGIac
>>1018419If I have to guess, he makes the model, then when it's time to texture he stencil-textures the drawing onto the model, after which he makes adjustments, touching it up
>>1018415really good. the phone is comically small though
>>1018445also, ikea drawers lel
>>1018415Looks ok but everything it's out of scale, some things are tiny and others are too big
Thanks guys>>1018445>>1018461ProbablyI started measuring everything correctly, like the computer loosely based on a Macintosh Quadra 700, but then at the end i just eyeballed everything.
>>1018462nice. clock and the smaller sticky notes still look a bit off, but not crazy.u got any resources/tuts on doing this sort of style?
>>1018471I watched several tutorials but all of them were kinda crap and didn't helped to what i wanted to achieve so i just experimented.I think the key points that make at least this render work is the noise modifiers applied to the freestyle pass so it's not completely uniform, the bloom/glare of course, the tonemapping on the compositor, and being more deliberate with the colors of the highlights and shadows rather than letting the environment/light decide. All the materials is just a basic toon shader (diffuse > shader to rgb > color ramp) with some having some slight adjustements inbetween and chosen colors for the color ramp.
>>1018315>>1018316On the bright side, if you aren't going to palettize, why not just put them as mipmaps in a DDS file, and have a setting in your game for the highest resolution mipmap allowed? They all look good, btw.
>>1018379Is the one in the background (presumably, the backside of the foreground one), meant to evoke the nuclear symbol? If so, that's a cool secondary image. Although the pitch black color is a bit distracting. I might lighten it slightly. Over all reminds me of C&C Generals.
six different designs, and four different types of rigs to end up where we're at here.
>>1018534Slightly different in that her bridge piercing is higher up where it should be.
>>1017627turret of something loosely based on a T-80B
>>1018534The sharp, square shoulders clash with the character design and presumable fabric choice, makes for unpleasant silhouette so I'd soften those but still keep the overall "square" shape you're going for.>>1018535This is more style critique but imo the head proportions feel a little too squashed and flat on top. It might need a small vertical scale upwards and/or rounding the top of the head in the center like a real head. Quick and dirty resize, I think the happy medium is somewhere closer to yours but slightly taller.From drawing anatomy, I start with the eyes or the brow ridge roughly on the midpoint between the chin and top of the head for realistic/semi realism. Different styles play with this proportion but it felt a bit too short on the top side in the closeup. If you analyze faces by making a triangle between the bottom of the lips and outer eye bounds, that triangle is a small portion of the overall frontal facial plane view. The cheek planes are rather large which I think could be balanced out by more mass in the brain case or lowering the eyes/nose a little and scaling down. As they are now, it indicates to our human senses a much larger skull so it looks a bit uncanny although it does indicate the character's presumed personality.
>>1018524it is for generalsthe backside looks ok in RTS far camera viewim still not sure why baking would make it that dark, probably cuz the lights were setup at the front of the vehicle
>>1018069New clothes.
>>1018579>it is for generalsThat's pretty cool. GLA?>probably cuz the lights were setup at the front of the vehicleThat's almost certainly the reason. You could try to set up a very low power light behind the vehicle. That should lighten those areas slightly without having too much or any effect on the other sides of the vehicle. But truthfully, it's quite a minor issue, and I don't think players of your mod are likely to find it jarring. >>1018541Very aesthetically pleasing. Is that a toilet plunger in the bustle? You might consider adding another smoke launcher on the other side of the turret. I would also set the sides of the tubes to smooth shading. Over all very nice.
>>1018541now for the hull>>1018586yes it is, it's an asset for a little tank game i'm working on. (thanks)
>>1018567Genuinely didn't expect that level of analyses. Thanks anon, I'll do another pass over. I'm finding with myself after a while I seem to stop seeing the smaller details, it gets kinda drowned out by the vision, I suppose. So it's good to see this from an analytical stand point.
St. Michael
>>1018582Noice
>>1018016Suggestion: add gauze tape, a couple of dark green prescription bottles. The green ones are much less common than the orange, but are common enough that they my warrant inclusion.But yeah, I can see why you're satisfied. They all look really, really good.
>>1018586>You could try to set up a very low power light behind the vehiclei already have skydome light (arnold) which I presume should've been enough + distant light to add shading that is not just constant lighting
>>1018365Update on my man-spider knight, if anyone got any suggestions or critiuqe im all for it. To do: rivets and straps, body texture, a couple khopeshes, hair has to be re-done and then Its rigging time
>>1018535tats feel out of place, the character is very minimal with flat colors while tattos are super detailed and have a lot of shading variation, its very obviously a stencil someone put on top, and doesnt feel like its actually in her skin. great job otherwise
>>1018364If you're still having this problem, maybe try this
>>1018582Mewtwo waifu now in progress
>>1018787Thanks for the suggestion bro but it didn't work. It's okay anyway it's only in the background, but it's the kind of thing I like to know how to solve in case it ever comes up again and it's in the foreground.
I see zie briefcase is safe?>>1018858Sure I understand, I did what I could
Trying to git gudder at making depth-based textures entirely in GIMP. It's gotten a lot easier since the 3.0 update since you can make an edit and have a stack of filters update in real time.The background pattern is completely procedural, being a bayer matrix and maze put through a kaleidoscope, median blur, and bevel. I change the seed on the maze and get a completely different pattern.I don't think I can go back to nodes or baking autism now.
>>1018809
>>1017627a table
>>1018929That doesn't look like a table to me but rather some kind of vintage file drawerRegardless, that's a nice ass model nigga
>>1018887Impressive. It's tickling something in my brain that appreciates procedural work and patterns.>>1018929Looks professional. Get a job.
>>1018928
>>1018962I've seen worse, unfortunately
wip
>>1018962Head shorter and wider
>>1018962
>>1018987Help me out. Tell me about big fat tots in zbrush and trying to merge all the geo together. I got so frustrated with mine that I ended up hand stitching everything together in maya.
>>1019027That's one hell of a bush
>>1018887A bit of progress on this. Getting more comfortable layering normal map elements and also developing this alien-ish style at the same time.Not happy with how the pipes look... The more in-depth distortion filters in GIMP aren't persistent layer effects so I've avoided them. Might just have to bite the bullet on that part though.>>1018939>It's tickling something in my brain that appreciates procedural work and patterns.A lot of the math involved in procedural generation of patterns and stuff corresponds to physical or chemical phenomena in real life, so there's probably some subconscious tingling in recognition of that.In any case, it's fun to play with such things.
I opened up blender for the first time in FOREVER and I made SOMETHING!I did it all on one side and then applied a mirror modifier.
>>1019056Very nice, taking a first step! As long as you keep doing it again you'll be gold
>Friend's daughter (she's 11) gets herself really upset when she was doing something on her computer for hours and forgot to save it>Being the nice guy I am I offered to try and help her get it back>Against all of the odds it was blender she had been using>Obviously recovered it and she was elated>What she had made was actually really decent>Asked her how long she had been using it>It was her first time>She wasn't even watching a tutorial she was reading the (mostly outdated) manual>She was obviously making mistakes but no more than anyone else when they first startI should take her under my wing and give her a crash course so she becomes elite shouldn't I? It seems the right thing to do.
>>1019028Im using blender, but the same apply on zbrush. I sculpt on a separate mesh the torso and the tits, then I join them and remesh with a hi-density settings to maintain the shape, then keep smoothing and sculping
>>1019079Yeah but your boobs merged together. That's my issue.
>>1019079she appears to have fused boob syndrome, or medically symmastia.
>>1019080make it more dense or separate the boobs, and take in note this is for figi production so its going to have some merge but its ok
ok, I think fund a solutionI separated the boobs enough to have a canal in between. Then join and remeshed, and made the boobs poligroups/faceset to control each individually sorry if is rough but I did it in a hurry
Shout out to all mah bidet using niggaz out there, REPRESENTYou DO use a bidet right you're not one of those weirdos who goes around with their asshole encrusted with shit, I hope
>>1019132first time I see one of these, here are like picrel
>>1018593armored rustbucket on standby
>>1019136I'd penetrate that
>>1019136That's some good shit nigga, are you that anon who did those furry military character(s?) too?
>>1019132The last time I saw a bidet like that was at my grandparent's houseNowadays it's more common to have a small shower handle on the side of the toilet, more hygienic too
>>1019143aye
>>1019149haha that's awesome, good show
>>1019076Yeah it'll be fun
wew, managed to set up 3d weaves from pattern imagespretty easy t bhtrying to recreate most of this:https://www.youtube.com/watch?v=ARobs3F28Hcthe rest shouldn't be too hard
mushroom girl from fobs
WIP
its rough restarting, i feel like ive forgotten everything and now it feels overwhelming to relearn and try to learn all the stuff ill need to know too
>>1019159I think that's going to be my plan. It's a shame I do hard surface and she only seems to like soft body though. I'm going to run into some issues there for sure.
>>1019076>>1019196just make sure you talk to your friend regularly about what you're doingi'm not implying anything, but people get weird about their kids
>>1019197Oh yeah absolutely. I mean we're really good friends, have been for like 20+ years now and she's my goddaughter. All of us are really close. Besides that it's much more likely we would be doing it in the kitchen or somewhere anyway because it's a laptop. I helped him build her bed once and just that felt uncomfortable being in her bedroom so even I would insist to do it downstairs.I appreciate you looking out for me though anon.
>>1019173added a parametric knitting pattern alternative to the weaving and set up the uv deformer. one minor issue i haven't worked out with the uv deformer, but other than that things moving along as expected
>>1019194You remember things over time. You'll get back up to speed faster.Also I made a flashbang instead of doing anything I'm supposed to be doing. I think it's pretty cute.
I have a bit of a strange question if someone wants to lend a hand.>Put object in wireframe>Front view is fine>Go into the side view>It's like the object is solid>Put it into the other sideview>It's wireframeIt's not just the view that's fucked. If I'm in front view and select two edges (with one in front of the view so it selects both with one selection) then it selects both. In the fake solid view it only selects one, but in the alt sideview where it looks wireframe it also only selects one.I have reference images and I thought they were somehow blocking something, but I use refs all of the time and this never happens.Anyone encountered this before?
My first model
>>1019228Your different views are in different modes
>>1019240Congrats
>>1019240Looking good for a first model, keep up the good work
>>1019027
>>1019242I didn't even know that was possible. Even still when I change them to wireframe independently they're still fucked.
>>1019219Done. I guess I have to do real work now.
What do you guys think about this cris workflow?I’m planning to mix a bit of weight painting, for areas like the traps, with the old-school approach from a game like Mario 64, using separate meshes, then tie it all together with a cell shader.Can’t be bothered to sculpt, retopologize, UV map, or bake normal maps, so I’ll just add more geometry and rely on vertex colors instead.If the proportions look totally wack, feel free to call it out.
>>1019255Well we'd have to compare it to the concept art. And I'd like to see fingers and feet first. I've made a lot of models upt to this point and made proportions judgement calls. Then I adds the hands and feet and my perception of the whole figure changes.
>>1019248As a suggestion
>>1019253Very nice!
>>1019259>>1019022There's a reason for that
>>1019293Sure. Even then. The eyes are too tall and narrow. It's your model. You can do whatever.
>>1017627working on a dagger atm
This and 1 more post for the flashbangs.
>>1019319I wanted to learn how to get the tab swinging in realtime in UE so I tested it out there.
Look at my cat.Today was the day of baby's first bone constraint!>>1019300I saw those in a histroy museum in one of arab countries.
>>1019495>The gas chamber memory eraserThe dual citizens won't like that!
>>1019498I'm resharing my character
>>1019497Youre on your way
>>1019136now for the interiors... driver's position
Repair kits
>>1019545>lowpoly >no detail>neon icons>contrasts from floor>high resolution>1 white dirt paintlooks like shit nigger. embarrassing for your 15 years of being here
added support for holes and quickly swept the curveshad to rework some stuff from earlier to support animation and multiple patternsperformance is reasonable, 5 million points, 31k curves here. take a minute to apply the pattern onto the mesh, but after that it's near/realtime to transfer animations to.next: general pattern noise, overall fuzz/frizz, width and holes from textures and maybe wrap it up with pattern blending
>>1019545Slum shack WIPMost likely, interiors will be modular while exteriors will be not, probably gonna make a bunch of unique shack models instead
>>1019248done, for now
>>1019631Gave her mother hair
>>1019697Dress test
New skin for Halloween incoming
>>1019201>>1019558>>1019173This looks interesting. I wonder if I can do something like this in Blender's geometry nodes. hmm...
>>1019808you can definitely do the core effect i.e. copying patterns on to the geometry based on uv space:https://www.youtube.com/watch?v=RXmS9R-DFTMonly thing i'm not sure about is constructing patterns from images. here's the weaving patterns book if you want to give it a go:https://www2.cs.arizona.edu/patterns/weaving/webdocs/pb_serrure.pdf
>>1018344I may be a little late but have you tried resetting normal vectors? Edit mode -> select everything -> Alt+N -> Reset Vectors
>>1019830awww cute!
>>1019756Physic sim
mostly done. just a couple of things to fix, but i think that's the end of my dev blog.
i lied. i wanted to show off a bit. i'm pleased with this set up.
>>1019943That's awesome
>>1018674final render, characters are mismatched, everythings detail level is random, its animated like shit, directed like shit, the corridor looks like something from an UE asset flip but its all because I ran out of motivation to finish it like a week into a four week challenge, still like the kngiht though and proud of myself for gritting my teeth and dragging it across the finish line instead of dropping out
this entire thread looks liks shit kek, nothing worthwhile as always
>>1019945thanggs>>1019950gj, anon. lotta work and better to finish than not.
I'm still a newbie but at least I can make decent beds
>>1019960I don't know how sincere this post is, but it's the reason a lot of people deride asset monkeys. Don't get me wrong, you have skill, but that model isn't very interesting. A lot of you act like you're Michelangelo just because you're good at modeling props.
>>1019970>but that model isn't very interesting.you just have absolutely zero taste
>>1019976Dude, that thing looks like it was superglued together. Have whatever preferences you like, just don’t be disappointed when you post that on social media and no one pays attention to it.
>>1019979>that thing looks like it was superglued together. >no one pays attention to it.you reek of jealousy there. buddy. have 7 years of experience as an artist.
>>1019970>>1019979
>>1019960you should actually add super glue gobs to all the gaps and straps of tape.
Knife from CAD to game-ready. I'm still working on it.
tactical perfume dispenserModeling is almost done.
>>1020003is this freecad?
>>1020017blender
>>1020003Groovy!
>>1020019It's not>>1020017Plasticity. I would have probably committed sudoku trying to make that in freecad.
>>1020027lel. i need to spend some time with plasticity. i did a tutorial once and enjoyed it, but i barely model so never went back.freecad did have a bit of a overhaul recently right?
>>1019990Very cool anon. Good job. Which cad program do you use?
>>1020028Yeah, but I don't expect it to have improved in free form modeling. One of the worst things of working with it is creating all the datum and reference lines, etc, which are mandatory. Plasticity is just a million times faster.
>>1020029Plasticity
Do you see any glaring anatomy issues with my dirty wh*re, my cockgobbling sl*t, my ravenous h*le, my sleazy *oid?But seriously, I'm intending to give this female basemesh an athletic build so the fact that muscles are pretty defined is by design. It is idealized especially around the waist area but other than that I'm aiming to keep this body relatively sober, not quite realistic but not extremely stylized either.I'm noticing the outer thigh is a bit too thin in proximity of the knee so I'm adding some meat there now
>>1020128
>>1020129Quick 'n dirty skinning test
>>1020131it looks like shit
>>1020128Disgusting. Furries should be banned from /3/.
>>1020128Needs a tad more pelvic bon protusion
>>1019936
>>1020128>>1020129>>1020131Seems okay, thus far, though I'm concern about the front ribs location.
snippet of an animation i'm working on, been very unorthodox as it's my first blender thing and am working with preexisting models as opposed to going from the ground up but it's fun. gonna learn how shaders work eventually so i can do more stylized stuff, although applying it to the models i'm working with will require a bit of reverse engineering
>>1020168
Just started the flipped normals 9 course meal.Hopefully when I'm done I'll be able to make any character I want and not have it look like shit.I think most of my issue is attention to detail and basic anatomy, we'll see though.
>>1020220I think if you go through it and do some homework you could become pretty good Tell me if it worked when you're done with it
this is the floridian ass muncherhe muncheson your ass, jerks off to you while your ass bleeds out from flesh wounds and then flees to his home in tampahe exists for the sole reason of me practicing how to make humanoids and rigging them on blender cause im actually kinda shit at modelling and rigging, might post more pics of him, this one depicts him completely unrigged
>>1020260das bussin unc fr, deadass
>>1017767Highly autistic. But also very impressive. You could publish this. This is real science. Kudos.
Someone posted a video and paper about creating hair using a basic mesh on /v/. I saw it and decided to try and recreate it. I'm not clever enough to code. But I'm moderately competent at geometry nodes. So I tried my hand at it. Here's a little video.I managed to get it to function just barely. But I'm excited to get it this far. Because it looks to be possible. I just need to find solutions to a few quirks that are causing gaps to appear. For now, the mesh requires that I mark the bottom and the tops in vector groups. And I have to manually fill in the in between layers with big Ngons. So editing the mesh is a little less than seamless. Still, it's worth those few extra steps in order to get these results. Each strand interpolates inside of the volume of the mesh nearly perfectly.(the Ngons create a slight inaccuracy in position)And so far, it's very light weight. Creating the volumetric UV only takes a few milliseconds. It takes about 3ms to create 1,000 hair curves. The most expensive part, is calculation to make the points stick to the mesh. That doubles the time. Totaling about 12ms per 1,000 curves. Then, giving the points geometry costs more. About 20ms per 1,000 for the node group I've created. Which captures a bunch of info about the curve. I'm on a shitty PC right now. So it's probably a lot faster for those using a decent graphics card and a newer version of Blender.(I'm using 4.1)Anyway, I hope I can iron out the remaining problems, because this will make crafting hair a lot easier and faster than even Blender's system of hair nodes. It's so much more intuitive to just move some boxes around, and it automatically interpolates exactly how it should.
>>1019802I can show it now. This was pretty quick, I was surprised. >>1019839Thanks bro
>>102038610/10, saved
>>1020363Alright. I've ironed out a couple things.-Cleaned up the UV calculation, so it's more reliable, and there's no gaps. I don't know how well it will fair when the base mesh is too complex. Since it's performed by using a couple of shortest path nodes, and juggling numbers. If the path is too complex, then it might trip up and return the wrong numbers and break. But for now, it's reliably works with basic cube shaped topology.-I figured out how to automate the creation of internet geometry. Now that I don't have to do that manually, it makes editing the base mesh much simpler.Adding the automatic internal mesh part, increased the time by another 6ms per 1000 curves. Taking it from 12ms to 18.Then, expanding the test mesh into a full bob cut added a lot more to the time. Resulting in a total of 30ms per 1000 curves.This isn't bad, because it means it's fast enough to edit in real time without a noticeable slow down.In the video, I pumped it up to 15,000 curves. That's only 250ms. Still not bad, because again, I'm on my shitty PC. It would be a lot faster on a real machine.Adding a couple of shape nodes to give the hair texture, and the node to give it geometry, it totals about 420ms. Just under half a second.
>>1020460Another hair video. You can see my PC really struggle at the end, trying to display 20,000 curves.I realized the calculation used to bind points was overkill. Removing 3 unnecessary sample surface nodes, increased the speed up by a lot. Reducing the cost from 420 to 250.The joy of discovering that optimization was short lived after I made this new test hair. It's longer, and curlier. Thus requiring a lot more points. So the cost is a lot higher anyway. Plus, it required clumping, which adds even more time.The shaping of the base mesh takes a little more work then I initially realized. But it's still overall very simple. This time, I started by forming a quad mesh around the whole scalp. Then extruded out, and added 2 loop cuts. Then used sculpting tools to shape the mesh. Using a mask on the scalp faces to prevent them from moving. the base mesh could be better sculpted, but I didn't want to fuss with it too much just for a test. When the curves are enabled, it covers up some of the wonkiness of the base mesh.I felt the need to push up to 20,000 curves, because 15,000 wasn't enough to cover the scalp.
>>1020530is this actually more performant than the existing hair nodes? production grooms are in the 100k curve range. also is this based on on the yuskel papers? i thought the performance gains for those ones came from building curves directly on the gpu and the planes were just used as a cheap way to store groom data.
>>1020551>is this actually more performant than the existing hair nodes? Honestly? I don't know. I haven't crafted a full head of hair using the existing nodes in a long time. I've started a number of times. But get bogged down with all that's required to make those nodes work. Dunno why. They just don't suit me.For me, this is about ease of use. Doing the base mesh like this settles many issues. Interpolating between curves is perfect. Creating parts in the hair is easy. There are fewer points to manipulates in order to shape the hair, so you're not grooming a thousand parent curves. If nothing else, I'll pursue this further just for the ease of it.Since I'm on my shitty PC(my good one broke), I can't really tell how fast it is on a real PC.I have noticed that it's *relatively* faster to create points, and then convert points into curves. Rather than create curves first and instance them. So I believe that setting it up the way I did is faster than the typical instancing method most tutorials teach.>production grooms are in the 100k curve rangeWith the mesh disabled, Blender handles curves a lot better. Creating 100k curves takes about 1,300ms up front. However, the viewport has no lag when navigated(pic related), and there is only a slight amount of lag when I toggle options. Running the timeline is 1 frame per second however. So I guess that 100k curves would be light work on a real PC. It's the mesh that adds the most amount of time.>is this based on on the yuskel papers?Yeah. In the most loosest way possible. I just skimmed and got the gist. Creating the base mesh looked doable.>i thought the performance gains for those ones came from building curves directly on the gpuThat's right. I don't know how to do that part. I experimented with SDFs a while back and got interesting results. But it didn't scale as well as I wished.>>1004257>>1005159
>>1020559>For me, this is about ease of use.fair enough. tb h when i first tried out grooming i assumed it would be easier than it is, but the amount of skill necessary is not dissimilar to being a good digital sculptor. a lot of it just comes down to building an intuition for guide interpolation and that only comes with time. plus, for a really good groom the amount of guides needed is much higher than i initially assumed.>I don't know how to do that parti'm not sure you'll be able to really. i don't think blender has anyway for users to directly build geometry using gpu compute. it would probably have to be an external plugin. houdini did a similar thing with their feather tools recently where the final hi-res barbs get built on the gpu at render time. i've never looked into how they're doing that. houdini does have an opencl node that can interact with geo but i've never tried building geo with it.
>>1020560This is probably my ignorance speaking. But if there was some kind of nearest sample node for shaders, then I bet an SDF method would be viable. I don't see why we can't tell the pixels of our screen to sample the nearest point of a cloud. But if we could, then I think I could build the rest. I wish shader nodes worked more like geometry nodes with an index to manipulate. I want to believe it's possible, though I'm certain if I brought that up in front of people who know better, they would have a dozen reasons why that's impossible.
>>1020561i'm only familiar with grid based sdf representations i think the resolution to represent hair would be prohibitively high. not sure how it would be done with a pure mathematical representation. anyhow, good luck.i'd say have a bit more of a play with the existing hair nodes as well since you can convert the curves you're creating here into guides - iirc there's a node that looks up the nearest surface an adds the necessary attribs for the generation and post processing nodes to work.
>cgpersia is down with host errori'm out the loop. just temporary or down for good?
>>1020564peers is up so it'll be back. they go down from time to time.
>>1020563"grid based sdf"? Can't say I know what that means. Might look into it later. I'm too tired at the moment.But if it's how I'm imagining, then I don't think you need a resolution higher than your display. A 1920x1080 display is only 2,073,600 pixels. That shouldn't be too hard to handle. Blender already does what I *assume* is a distance calculation on points. When you look at point clouds in cycles, they're displayed as proper spheres. There must be a way to expand on this. Make the sphere's connect somehow.>iirc there's a node that looks up the nearest surface an adds the necessary attribs for the generation and post processing nodes to work.huh?
>>1020566in houdini we deal with sdf not in their pure mathematical representation, but as vdb volumes. those volumes have a grid resolution. basically you're looking up values for each voxel and the voxel where values transition from positive to negative is displayed as the surface.re: the points as spheres, i think all renderers i've used repesent loose points this way. i think it's just instancing a very efficient sphere primitive on them. drawing lines between points etc would be more work i'm guessing.>huh?https://docs.blender.org/manual/en/latest/modeling/geometry_nodes/hair/utility/attach_hair_curves_to_surface.html
>>1020565yeah but 2FA is down as well so I can't login and use the site. GRIM
>>1020568might be the aws outage
>>1020567>those volumes have a grid resolutionAh, I think I get what you mean in the general sense. I was just thinking if creating something like that was possible in geometry nodes. Or maybe if it's already possible using the volume nodes. I haven't played with volume nodes much.>drawing lines between points etc would be more work i'm guessing.Maybe.>attach_hair_curves_to_surfaceOh, that node. I might be able to use it. Though, I just tried it, and it didn't work. So I'll have to figure out what wrong later.I can probably build a light weight version myself. It would probably be more efficient to attach the points before they're converted into curves.
made a new low poly base, testing it out, not sure if i´ll finish them
>>1020690Is the left model supposed to be Emma from ZA?
>>1020690How'd you make this? Need to make low poly character bases
>>1020910>extrude faces, slice edges, move verts
>>1020363>>1020460I'm surprised this is considered new tech. This is how I always imagined hair systems should work. Every hair styling guide talks about hair like it's box-modeling, yet they decided to make it some painful sculpting process.I'm gonna learn how to use geometry nodes now.
>>1020968The approach is an old classic, doing it through geometry nodes is new. Although presumably the paper might have had some interesting contributions on how to generate the mesh, but we don't know which paper that was.
https://www.youtube.com/watch?v=OF8J9BwCC5o
>>1020972When you instance leaves, instance them according to normal to surface (and rotate geometry in leaves accordingly, to work properly when instanced), it is very noticable a lot of them are randomly floating around not attached to the breanches. Also the pink stuff looks way off. Otherwise, it looks good. Music is nice too.
Triple HThe child of constraint was acting funny with the water bottle and I had to remove the fluid since I couldn't get it to where I wanted it to go.
>>1020968The the other anon said, it's not new tech. Blender devs could have done this a long time ago, but simply didn't want to. Or maybe they didn't even think to. It's neglected tech.I want to share my node group. But at the same time, I want to make sure it's more solidly functional before letting it out in the world. But if you want to try your hand at making something similar, I would maybe suggest learning Python instead. I got into nodes because they looked easier than coding. But I kept running into roadblocks, because the things I want to create are too ambitious for nodes. If I got into Python instead, I would be able to do more, as Python has greater control over the functions of Blender.>>1020971This is the paper:https://www.cemyuksel.com/research/hairmesh_rendering/Real-Time_Hair_Rendering_with_Hair_Meshes-SIGGRAPH24.pdfhttps://www.cemyuksel.com/research/hairmesh_rendering/I pretty much just skimmed it a couple times, starred at the geometry images a while, and thought "this might work". The real magic is in the GPU work they did. Which I'm not capable of replicating. They took the information from the geometry, and then build the hair directly on the gpu from that. That's why their method is so blazing fast. I only approximated the geometry part. And not even accurately. Just good enough to look nice.
Working on learning to pose models to ref
Finally finalized a base model. I like the pure white look, but I'm still unsure on whether I wanna keep that or give em actual skin tones.
>>1021026 (Me)Also, robo drone design based off a paper airplane I made for fun in Church last week.
>>1021009I'm opening this every time I scroll and think: "hell yeah".
Barbarian that is turning out better than expected
>>1021073Death by Snu Snu
working on a cave troll for me videogame. Later on I'll apply smooth shading, but now I'm wondering if I should add more detail or move onto texturing.
>>1021085Are poles acceptable for characters/rigged models? I do neither so it's handy to know for when the time comes.
>>1021090Poles will screw you over in the long run for anything that softly deforms.
here's the big guy with the little dude>>1021090>polesI didn't even know what that is until I google it just now lol>>1021091>anything that softly deformsshould I worry about them if my model has shoulders and the upper legs separated from the main body?
>>1021092Its the big tummy you should worry about
>>1021093thanks anon. It probably shows but I'm quite a noob when it comes to this so having people point out those things to me sure is a great help. I'll keep an eye on that tummy while doing the animations.In any case, how are you supposed to tackle spherical things like a big belly on a low poly model?
>>1021085You generally want to avoid triangles and stick to quads because triangles interfere with your ability to easily add more detail or loop cuts. Stick to quads and optimize to tris as the last step. It's quite low poly already and you won't be hurting anything. That said, your topo is actually pretty good for a beginner and not a horror show like most noobs. It's generally fine provided the mesh stays rigid.>>10210925-star poles are unavoidable but they should be kept away from places that deform. A point with 6+ (quad) faces can be avoided and should be retopologized.Poles are bad for deformation because that means one point is affecting how many faces look and leads to deformation artifacts.
>>1021092Those are great! Love the retro style, those would have definitely been pestering me in a laye era dos svga fame like LBA
>>1021091Geez dude, racist much? Polish people have been pretty nice in my experience. but you're making them sound like a bunch of cutthroat fraudsters and womanizers.