[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/3/ - 3DCG

[Advertise on 4chan]

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • There are 113 posters in this thread.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: Wip Collage_Series 4_v2.jpg (2.76 MB, 3840x2798)
2.76 MB
2.76 MB JPG
/wip/ - Works in Progress
- Collage Edition: Series 4 -
>You, who have assembled the 7 /wip/balls, speak and name your wish...

Post your work-in-progress projects, recently finished projects, or things you'd like critiqued here.

Previous thread: >>894078


List of free resources: https://pastebin.com/cZLVnNtB
/3/ Discord for those interested: https://discord.gg/ujt5vtr4DE
>>
File: SOB.jpg (140 KB, 1024x1024)
140 KB
140 KB JPG
First in fuck yeah
Been working on this sister of battle for a few weeks after work
>>
>>896926
>No boobplate

trash
>>
File: howboutizucyadik.png (1.79 MB, 1080x1920)
1.79 MB
1.79 MB PNG
Making a bug alien reference for my players for ttrpg I'm running.
>>
>>896964
dem hips and thighs tho
>>
>>896926
Very nice, how do you achieve such clean geometry?
I think that other anon has a point though, if you're going to go out of your way to model the ladies you might as well walk the extra mile and model a nice boob shaped plate. They're titillating.
>>
File: TheGraveyardRose.webm (1.24 MB, 1440x1080)
1.24 MB
1.24 MB WEBM
>>896926
Looks good, I've almost finished Elspeth Von Draken
>>
File: Scn 15.jpg (2.08 MB, 5120x1080)
2.08 MB
2.08 MB JPG
>>896999
Checked

>>896926
Nice anon.

>>897000
Checked and nice

I'm currently on the 7th chair making some knots...
>>
>>897000
Wow, very cool, once again /3/ strikes with tripple A quality all around. Tell us more about Elspeth she looks like a cool character.

>>896964
Your bug man has lady hips. Is this intentional?
>>
>>897001
ty
>>897002
she is basically the witch with a dragon from Warhammer fantasy, Lady of Nuln and all that you can build an army arround her, I want to sculpt more models to print this summer
>>
>>897004
Ah nice. That model is pretty intense, how long did the whole thing take you to sculpt?
>>
>>897009
The first model that I sculpted last year is in the top right, took me 30 hours or so to finish it from there.
>>
>>897011
Neat. I'm guessing you know your way around zbrush. 30 hours sounds like good time.
Nice work, it looks real professional. Personally I'd make the dragons wings a tad more bow shaped on the side view for a nice profile silhouette as a small nit pick but yeah, very nice model.
>>
>>897012
I know how to use zbrush, 3ds max and substance painter, but that's done in blender
>>
>>896917
Ahh great work as usual OP, stitching the head on that body anon never finished is a great continuation of tradition.

>>896926
>>897000
Jeez, strong start to the thread.
>>
>>897000
I've had to make a few complex anatomically detailed dragon for people composed of many subtools and grueling integration and retopology work but the thought of trying to fit all that on top of it with a detailed throne to boot would leave me with headache.

Props to you on that especially the composition.
>>
>>897013
Impressive. I keep meaning to learn blender sculpt but it's just so different from zbrush that I keep putting it off.
You seem to get pretty good results with it, would you recommend it? Unless of course you poly modeled the whole thing.
>>
>>897025
>would you recommend it?
Nope, you miss lots of tools from Zbrush, and I like some procedural effects in blender but the performance is horrible, I used the ivy generator in that model and it was a pain in the ass too.
>>
>>897012
He can also get more distinct membranous wings by zmodeling a plane and utilizing remesh guide so the corners of the sharp corners on the mesh don't contract back in when upping the remesh polycount. Make the Horns, nails, and claws as separate subtools and taper and bend them in the gizmo. Create the scale on the body separately and boolean them on top of a form which from references looks like it had a standard brush run through it down the middle. The nice about this dragon's references is that it's very alpha friendly. If you want to elaborate on its form with some muscles (especially areas that have those scales resting on top of them) then this would be probably closer to a lindwurm that has been given two extra legs if you can find an anatomy diagram for that.
>>
>>897002
Other way round. Bug woman has no need for tits.
>>
>>897027
The geo node version of the ivy definitely helps. The implemented ivy gen is absolute trash in Blender.
>>
>>897001
I hope the pay is either really good or you really like chairs.
>>
>>897065
In this moment I'm hating chairs. Its not funny when you have to do a whole collection but that's the job. I'd prefer to model tiddies and dicks. The pay is ok. In the long run I like to create dependence from my clients, some of them no longer hire photographers and depend on me to create catalogs and product portofolios.
>>
>>897001
I hate to wortk with scans I use meshroom and all is slow as fuck, program?
>>
File: Be'lakorScreen.png (1.42 MB, 1418x890)
1.42 MB
1.42 MB PNG
>>897031
Yes the dragon's body is an old model and I didn't used to work that way then, this is from a recent sculpt and even if the membrane is more simple all looks better, the model has a few scales as subtools, that I should sculpt more, the horns, nails, fangs and claws are different subtools too, the body scales match the reference, ofc it could look better with different ones.
The model has 350 subtools, 250 the witch, and 15M tris and takes 2.5 minutes to load.
>>
File: 1639223782065.gif (59 KB, 397x400)
59 KB
59 KB GIF
>>897119
Wow anon, it looks very cool
>>
File: arena prototype.png (443 KB, 1581x868)
443 KB
443 KB PNG
Prototyping an arena with a bronze age ziggurat-ish touch mixed with your typical greco-roman radial configuration. Arena platform is way too far away from the crowd, 5 meters below. Gonna lower it to 3.
Also just for shits and giggles I checked how many people it can hold and the maximum amount is at 8000 more or less. The more you know!
>>
File: arena exterior.png (366 KB, 1186x658)
366 KB
366 KB PNG
>>
>>897137
raise it to 3 you mean
looks cool
>>
>>897137
also you may want to consider a steeper angle on the leadup to the seating area
right now it looks like the fighters could make their way up and escape
>>
>>897143
>looks cool
Thanks man
>>897144
Don't worry I'll put some fences leading up to the crowd, spikes, a firewall or slippery oil, I just wanna retain the low angle for that ziggurath-ish A E S T H E T I C
>>
>>897105
Actually meshroom cuz I have to work with either licensed software or free software within my own company. It's not too bad because I have a slave computer dedicated only to this kind of processing.

When you get to work with some clients you have to sign and comply with certain IP, competition and compliance standards so my office computers have to be clean for audits.

I prefer reality capture, it's the perfect balance of speed an accuracy for me but I have to play it legal inside my company. I actually do archviz 95% of the time so having another more robust program is not worth it.
As it's only for reference use I do not need perfect scans but I'm saving on one of those Einscanners so I can speed up this process.

At home I do have a pc that looks like cgpersia's server but I can't use it for professional work and I do prefer reality capture for personal work.
>>
File: rough room.png (2 MB, 1920x1080)
2 MB
2 MB PNG
>>896917
The purpose of this exercise was try to model my room as fast as possible while still keeping a decent accuracy. The intention is to take this into a 2d program (CSP) instead of sketching. I was making good progress until I got to that balcony door which completely wrecked me when I tried to box model it. Then I went back to booleaning it like the rest of the objects, but I ended up misunderstanding important parts of the structure and will have to repeat it. I spent ~3h on it today and it completely killed my momentum.

This whole experience made me realize that instead of trying to hack it in Blender, I should try Moi for my sketching needs instead, which is convenient as I wanted to use it more anyway. Hopscutter really makes things faster, but it does not give me the accuracy that Moi would. If it was vanilla Blender it would be much worse.
>>
File: 1623297638198.jpg (104 KB, 1515x1116)
104 KB
104 KB JPG
Anyone know why this is happening or an easy way to fix it? The rust part on the top is fine, but the one below looks really metallic and really orange. It's the same shade as the one at the top but isn't acting like it.
I'm not sure if I'm doing it right, I'm just using a color ramp to change the black to red because that's all I really want to do.
>>
>>897166
>that balcony door which completely wrecked me when I tried to box model it.
what part of that door did you have trouble with?
>>
>>897172
My best guess would be that is has to do with the color difference between the top and bottom of the hull. Black is [r: 0, g: 0, b: 0] or [hex: 000000], so if you add a color to it, the color will just be itself, but if you add a color to that orange-ish color, it will be different. Imagine it like regular math. If you add 37 to 0, the result will be 37, but if you ad 37 to 15, the result will be 52. You could probably fix the issue by overlaying the rust instead of adding it (or at least I'm assuming that's what you're doing).
>>
>>897176
I was thinking it might have something to do with the rust and red being so close in colour. When I reverse the orange and red on the color ramp it actually looks better, but then the rust parts from the top line up with the coloured parts on the bottom, and the same for the rust on the bottom with the colour on the top, so it looks mismatched.

>or at least I'm assuming that's what you're doing

That's what I would prefer to do but I'm too much of a brainlet. The whole thing is just one texture (obviously the bottom is a separate texture, but it's still essentially the same texture file if you know what I mean).
It's not really a big deal because it'll be in the background anyway, and I hate shitting threads up with my stupid questions, but if something like this crops up, even when it isn't relevant, I like to try and get an answer so I know what to do in the future should it be relevant.
Thanks for your help though. I'll try the overlay method which I think should work much better.
>>
Gonna try character modeling for the first time
>>
>>897188
Based skelly maker. By the way, may I suggest modeling the jaw as a separate object? Since skulls irl have it separate too, unless of course you already know that and this choice is deliberate, stylization
>>
File: basicskeleton.png (130 KB, 771x682)
130 KB
130 KB PNG
I made a super cool skeleton model! Im gonna make it look neater by aligning the bones together more
>>
File: swatsoldierDS.gif (496 KB, 256x192)
496 KB
496 KB GIF
If i try hard enough, i could use it in a videogame! That would be super cool. But right now i've been having fun making lil models as one offs. Here's a lil swat guy i made.
>>
File: hole.png (335 KB, 715x494)
335 KB
335 KB PNG
>>897188
That's nostalgic, my first non-tutorial project was a skull.
Really make the jaw separate. Also there's a hole here, it's easy to miss.
>>
File: demonhelm.png (624 KB, 1136x1239)
624 KB
624 KB PNG
>>
https://www.youtube.com/watch?v=dRlaBLJPDTg

working on comfort project to help climb out of artistic burnout. marble game
>>
File: farmer.png (806 KB, 1538x836)
806 KB
806 KB PNG
Everyone in this thread is talented.
Well, I have little to no idea what I'm doing. I watched enough tutorials to push polygons around, and add basic materials. But beyond that? Absolutely clueless. Well, I figured I'd take a shot at an "easy" model. But it's taking hours just to get this far.

I'm not trying to make "low poly" art per se. But I'm trying to be economical. Only add when it feels necessary. Toggling on smooth shading makes the model look like shit.
>>
>>897238
for you model to look better on smooth shading you either activate auto smooth or use the edge split modifier
and select the edges that should not be smoothed and mark them as sharp
>>
File: blender_jivK11X0KZ.webm (1.82 MB, 1424x796)
1.82 MB
1.82 MB WEBM
>>
>>897248
Ahh, I see. This thought has crossed my mind. But I didn't think of it too seriously. The amount of effort that requires sounds daunting. But I guess if I have to, then I have to.

Would you happen to know if this edge data can be used in a game engine? I'm not planning on making a game. But I do want to think that my model is good enough to use in a game.
>>
>>897253
BNUUY METAL
FAKK U

How in the fuck did you get the topology for the muscle definition down?
>>
>>897238
try something like this for the eyes
>>
File: Screenshot_1.png (16 KB, 387x172)
16 KB
16 KB PNG
>>897254
Just click on object, shade smooth, then click on the mesh tab and enable auto normals. Tweak the angle until it looks right.
It's that easy.
>>
>>897253
Really cool, friend
>>
File: 1621757928842.jpg (146 KB, 1920x1080)
146 KB
146 KB JPG
How's it coming bros? I know you can't really see much of anything because it's so dark, but am I on the right track?
It's supposed to be underwater. I have some minisubs made for those lights that are just floating there not attached to anything yet. It's only a test render obviously so I will tamper with the lighting and things, there's still a long way to go yet.
Also I did try it with the denoiser, but it made everything so blurry that it looked even worse so I turned it off.
No bully.
>>
>>897302
"denoiser" is a meme. It's literally just blur.
>>
>>897302
When you're rendering dark environments it's better to have more light in the scene and then composite in the darkness.
>>
>>897302
Render it with the denoiser so that we can get a comparison.
>>
>>897302

I see you want the focal point on the big entrance and you have 2 statues there. If you place the sub on the lower left corner have the light be bright enough to get the statues in view. That should somehow fix your composition problems partially at least.. Making a thumbnail sketch of what you want is TONS better than fiddling for hours to find the correct settings. It all boils down to art fundies basically.
>>
this talk about denoising makes me wonder: do blender users get annoyed by the cycle of clickbait hype generated by youtubers around new features?
with denoising in particular, it feels like a big beginner's trap. lots of splotchy images and surface details completely obliterated.
>>
File: rough room v2.png (2.06 MB, 1920x1080)
2.06 MB
2.06 MB PNG
>>897173
I started off by insetting and extruding in the wrong direction, and when I tried to fix it by moving the faces, I got face flaps as if were a tent. When I tried solving this by knifing in extra points only for the knife tool to not actually do what I want. For some reason it was not putting in the points and this started to put me on tilt. Afterwards I realized that I forgot to do the wooden enclosure around the door and gave up in disgust. Anyway, here is what I wanted to do with that door. I was in the zone the whole day today, but I really wasn't fast enough as it took me 5h to model it. At this distance you can't really see the curvature of the grills properly so that was a wasted effort. But apart from that I do not think I did much that isn't visible. I'll get better at this. I've studied a lot over the past 7 months and now is the time to refine that knowledge into actual skill through practice.

I am going to bring this into CSP and try painting over it and see where that gets me. I want to develop a hybrid art style for illustrations. Maybe it will turn out that I need to do more (or less) on the 3d side. But modeling each prop to perfection is not the way to go for this.
>>
>>897329
may i ask if you've spent much time doing no-frills poly/sub-d modelling before jumping into moi or hopscutter?
>>
>>897331
I don't think I've done much modeling, but at this point it does not feel like I am a beginner anymore since I have a lot of knowledge and some experience. I can't be an expert due to lack of actual practice though.

You'd really need to be an expert at either style to understand their pros and cons though. Nobody will claim that C is easier to program in that Python.
>>
>>897001
How would you model that wood tape?
>>
>>897344
It is easy to draw a curve along the surface in Zbrush, in Moi I could also place the curve control points on the surface and it will align the normals properly followed by a sweep. In Blender I could use the face snapping functionality with alignment turned on to get the same thing. Once you have a curve you can do a sweep in Blender as well. I am not familiar with it enough to say for sure, but you could probably draw a curve using the grease pencil as well.
>>
>>897343
i think you'd benefit from a few exercises limiting yourself to traditional box/sub-d/poly modelling and very strict (overbearing) constraints like no booleans at any stage, no non-quads ever, no creasing etc.

it will be very frustrating, but at the end of it, many modelling tasks will become trivial for you, and you'll avoid the downsides (clea up) that come with the other workflows.
>>
>>897326
denoiser is really only useful in like archviz interior scenes, which are already prone to be quite noisy due to the nature of the lighting itself and there usually isn't anything with fine surface detail that would suffer from a particularly aggressive denoise alg

but in 2022 for archviz I'd consider using UE5 rather than an offline renderer, so meh I guess.
forget about denoising character renders with skin pores or such, unless your character is literally a plastic doll, the denoiser will turn it into one.

but of course, LOOK AT LE DENOISER THINGY IN TEH NEW BLENDER VERSION OMG GUYS SO KEWL SO HECKIN FAST AND VALID LIKE FAVORITE SUBSCRIBE GIVE ME FREE MONEY XDXDXDXDXDXD
I hate youtubers so much.
>>
>>897305
That's what I was thinking. I did it ages ago and it was bit like when you use clone in photoshop and if you don't do it right it looks sort of patchy.

>>897317
Ah I didn't think of that, thanks a lot I'll do that.
It's hard enough as it is because I have to model things in Eevee and then render in Cycles. My computer doesn't have the ability to model in Cycles. Just for that I posted, I'm not kidding, it would take about 2 minutes to render in the viewport.

>>897321
I can do it again, but it won't be the same because I don't have the lights or nothing in anymore. I just added them in for the timebeing to see how it was looking with them.

>>897324
The one on the lower left was supposed to be lighting my meme sub but it just kinda looks like a rock and you can't tell what it is. I could easily add a few more subs. I think the other anon is right in that I should darken it in the compositor. It'll be easier that way.

>Making a thumbnail sketch of what you want is TONS better than fiddling for hours to find the correct settings

Believe it or not it's the first time I did do that. I only really did a basic gist of it though on a scrap piece of paper that's like 2"x2".
Like everything I just try to find a starting point and then try to wing it from there.

Thank you for your replies friends.
>>
>>897279
That's neat and all. But I want the eyes to look 3D. I'm not going to go for a spherical style eyes, because that would be too realistic. But something that does convey depth. I'm not sure how I'm going to do that. I'll likely hollow out the eye sockets, and paint the inside white. If it works how I think it might, then animating the pupils inside of the blank eye cavity will create the illusion of whole eyes.

And besides. I haven't learned how to texture yet.

>>897288
Thanks, I'll see how it works out.
>>
File: blender_2ZONnf4Abl.webm (2.27 MB, 768x790)
2.27 MB
2.27 MB WEBM
>>897255
sheer.
>>897294
thx

Tried to 'brighten' up the face.
>>
File: wip4.png (104 KB, 270x456)
104 KB
104 KB PNG
>>
File: farmer 02.png (1.01 MB, 1736x842)
1.01 MB
1.01 MB PNG
>>897362
Alright. Focused on the face today. I tried to make smart loops like how I saw in a tutorial. Loop around the eyes and mouth and all that jazz.
Set auto smoothing to 70%. That seems to get the edges I want. But I also looked up a toon shading tutorial. I don't think the tutorial I found will get the exact results I want. But it will suffice for now, Until I find something better.
The eye shape is practically what I want. But I don't know how I'm going to trick the viewer into thinking they're solid rather than hollowed. I guess I'll have to look up a tutorial for anime eyes too. So be it.
>>
I dont recommend posting your work to a known white supremecist board
>>
>>897447
But what if I'm a white supremacist?
>>
>>897447
Did you hear the news about the white power shooter, come to browse because you were curious about the big evil Four Chans, stumble into /3/ and decide you had to warn us about the site's reputation?
We just post cg stuff here. Calm down.
>>
File: Why.png (499 KB, 1248x1500)
499 KB
499 KB PNG
Why, when I set the beveled edges of my little cube mesh to Shade Smooth, do I get this ugly shine accross the front?
I tried marking the rest of the edges as sharp, but it doesn't seem to make a difference.
>>
>>897453
One or more of the side polys share the same smoothing group with the front face.
idk how it works in blender but this should be a one button fix.
>>
>>897453
You need to turn on autosmooth for the sharp edges to work. Object Data Properties -> Normals -> Autosmooth. Alternatively, one of the faces might be non-planar. In Edit mode, select all and go Mesh -> Clean Up -> Make Planar Faces.
>>
>>897456
Autosmooth fixed it instantly. My thanks.
>>897455
It truly was a one-button solution lol
>>
>>897447
Fuck off nigger fbi chud
>>
File: Shenwrongrender.png (2.26 MB, 2010x953)
2.26 MB
2.26 MB PNG
Shenron Blendlet again.
The issue I was having with clouds is that I was trying to go for a more interesting background but it didn't look right with realistic clouds. I feel like it looks okay now but I'm taking a break to play Path of Exile.
>>896917
Incredible to be put into the collage, first time I've been included in anything. Means a lot.
>>
>>897220
>>897219
Absurdly good.
>>
>>897431
This has a slight cavewoman vibe. Pretty cool I think.
>>
File: intestines.jpg (612 KB, 1920x1080)
612 KB
612 KB JPG
>>
File: 2B.jpg (1.77 MB, 3840x2160)
1.77 MB
1.77 MB JPG
>>
>>897220
you can probably already use that in a video game
>>
File: 1644229678066.png (316 KB, 496x438)
316 KB
316 KB PNG
>>
>>897329
I tried using this as ref and panting it in 2d, but trying to draw the bed creases killed me. My 2d skills are just too weak. I am going to have to look into toon shading and making use of freestyle edges in Blender. I've decided that I want to lean more on 3d and need to find ways of making use of that modeling info. What are some good resources for going from 3d to 2d anime style art?

The plan is to do some research and do further modeling to complete the room further and make it more accurate. I'll redo the beds in Moi, bring them into Blender (or maybe Zbrush) and then come in with a clothes brush to do the creases.
>>
File: Scn 16.jpg (536 KB, 1574x841)
536 KB
536 KB JPG
>>897344
Checked. Chair anon here. I used a spline with a special compound object in max called loft. There you can control the profile and bank of the fiber. I do it with a smoothing modifier stacked so it looks good when it's at final res.

Al so made a small script called select by ring so i can hide sections of the fiber and solve knots around themselves.

Knots have to knotted just like the real thing because some images are meant to be factory references so yeah.
>>
File: farmer 03.png (1.19 MB, 1817x845)
1.19 MB
1.19 MB PNG
>>897438
-Worked on the face some more.
-Modeled ears(that took some time)
-Watched a few more toon shading tutorials. Found a different way to handle coloring. The previous method involved selecting specific colors. However, this other method can mix with textures. I haven't made any textures though, so it's just mixing a base color for now.
-For the outline, all the tutorials suggest the same method. I'm not really liking this method, because the lines look disconnected and all warbly. I'll have to hunt for a better method. But for now, this one is fun to play with. I was able to fashion some eyebrows by extruding out from the brow area, and the lines automatically draw around the extrusion. Pretty neat.
-There's the start of a hat up above. But I got distracted with more face and head stuff.
>>
>>897664
Damn, you done channeled that Toriyama. Nice.
>>
>>897664
Very neat Anon, you're using the inverted hull method for the outline? A popular alternate approach is to do the outline entirely by compositing it on top as a post process effect, search for UE4/Unity shader tutorials with keywords like "post process toon shader", but I'll quickly summarize it here:

1) Choose a render pass to use as the base, color/final image usually has too much noisy info, world normal or depth is better.
2) Run and edge detection kernel over the pass, laplacian and sobel are popular choices.
3) Run a dilation kernel for X passes until satisfied with line thickness.
4) Multiply on top of original image.
>>
>>897672
Thanks. I'm going for old school Harvest Moon. I guess the style is similar to Toriyama's.

>>897676
Full disclosure: I've never rendered anything on purpose. I've accidentally rendered stuff by playing with the options. But I really don't know what I'm doing there at all. I'll need to look up how to render before any of that makes sense. A "dilation kernel"? lol

But I'm really hoping for a method that can be done in real time. Because the goal is to make a model that can realistically be used in a video game. Not that I have any real aspirations to develop a game. But the *idea* that a game could use the model is fun for me. But I guess it can't hurt to learn what the post process toon shader can do.
>>
File: Spooky.jpg (1013 KB, 1200x900)
1013 KB
1013 KB JPG
Are you afraid?
>>
>>897664
Love this lil guy. Inspirational.
>>
>>897795
Yes I am afraid of turds in my pool.
>>
>>897795
>>897823

Actually it looks like diarrhea foam.
>>
>>897824
Look at this anon you could get some inspiration.

https://www.youtube.com/watch?v=LekZbZ0ksyA&t=92s
>>
File: 1646743766001.png (818 KB, 585x724)
818 KB
818 KB PNG
>>897599
>>
File: 1613790518704.jpg (34 KB, 300x300)
34 KB
34 KB JPG
>>897664
He's so fucking cute and I rarely say this kind of shit, props to you anon.
>>
>>897823
>>897824
yes it was supposed to be a turd.
>>
File: Tupandactylus.jpg (32 KB, 938x900)
32 KB
32 KB JPG
>>897834
Cool, a flying drone thing or a helmet?

Trying to put fur on this bitch but it's melting my pc. Want to upgrade from GTX 1070 to a RTX 3070 TI but is it worth it right now?
>>
File: canyon.jpg (792 KB, 1920x720)
792 KB
792 KB JPG
Trying my hand at geonodes for terrain creation. hashing out some bullshit sci-fi concept art.

Blender is kinda fun
>>
>>897210
>>897232
>jaw separate
Fuck I should have done that, wish I looked back in here earlier. I could try deleting the jaw bits and make the jaw separate
>>
File: 1631497206438.png (949 KB, 737x839)
949 KB
949 KB PNG
>>897855
Yeah it's a helmet
>>
>>897907
Looks cool but at a design level, I'd remove these long screws or change them, they make the thing look fragile/cheap
>>
File: farmer04.png (929 KB, 1525x836)
929 KB
929 KB PNG
>>897814
>>897835
'preciate it.

>>897664
Bah, I lost my original post. Don't feel like retyping it all. Here's the QRD
-Crude hand. Bad topology. Needs work.
-Thinking perhaps working with odd faces is better. Dividing faces from 1 to 2 to 4 is impossible without creating triangles. But 1 to 3 to 5 can ensure quads.
-Crude hat. Don't how to fit it on his head snugly. They're separate objects.
-Got lazy with the hair line and created a bunch of triangles. I should probably handle his hairline with textures instead. But I don't feel like learning textures yet.
-Played with lighting more. Added some diffuse lighting. Ultimately want to go with a watercolor look, but don't know how to pull it off. Need better tutorial.
-Don't know how to get the eyebrows to show over top of the hat.
Don't know how to make the lines look connected. Hate how they "hang" off the surface like that. The gaps are ugly.
-I need to learn more, because I'm uncertain how to do a lot of things.
>>
File: Day 1 Render.png (1.45 MB, 525x691)
1.45 MB
1.45 MB PNG
Been tooling around with blender for a few weeks. Decided to go for a cliche floating island as my first project I actually intended to finish, so here's my first pass. Thankfully I realized I hadn't activated CUDA *before* I tried rendering.
>>
>>897946
>Don't know how to make the lines look connected. Hate how they "hang" off the surface like that. The gaps are ugly.

A couple of ideas:
- Stack multiple outline shells instead of one, that should fill in problem areas like the eyes.
- Another angle of attack: instead of filling in the lines just "cut away" the tiny outline slivers. The outline shader could have logic to drop the opacity of the outline to 0 if the angle between a polygon and camera is close to 90 degrees.
>>
>>897447
You'd accomplish a lot in your life if you dropped your pessimistic outlook on life. The only person that fails will only be you
>>
>>897883
On another note, can anyone answer me some questions regarding the ocean modifier:
Is it always the best choice when trying to create water surfaces, even when it's something as simple as this?
Or does it make more sense to stack some noises using the displacer and get cracking with a decent shader instead of fucking around with the foam settings?
Which one is better in regards to render times, especially when including shader fuckery for both to look equal?

I'd try that shit myself but given the shitty specs of my system I can't exactly run tests all day, as I need that shit to work, so I'm hoping some of you fags have some experience in that area.
>>
File: farmer05.png (397 KB, 573x786)
397 KB
397 KB PNG
>>897948
Stacking sounds like a bad idea. Because the solidify modifier already creates a double of the mesh. Here, I stacked 3 more, for a total of 4. And so now it's essentially as if 5 meshes are overlapping each other. That's an expensive solution. And you still see gaps. It's just slightly less noticeable now. There has to be something that can calculate the distance between faces and fill the space with a solid color. There just HAS to be.

I don't want to cut away the slivers. But what is this logic you speak of? How do I do such a thing? Altering opacity based on angle of poly and camera sounds VERY useful.
Imagine if I could create color that's only visible at a specific angle. Please tell me more.
>>
>>897908
I initially tried to have a single big chunky latch but it looked a bit too ridiculous so I settled for something smaller. To be honest I don't really have a direction for this project, I just want to familiarize myself with different hard surface techniques and hard surface topology.
>>
>>897603
Yeah, the bed creases are really killing me. I have no idea what I am doing when it comes to 2d and am just searching for a mental triggers that never come. If it is just laying down flats and lines I could manage, but if it is just flats and lines I could do that in 3d. I'll have to pick up drawing as a hobby, and forget about painting for the time being unless I want to just straight up trace over the reference. I am really not into that, if I had take such an approach I'd be better off just getting a programming job.

Found some interesting things while looking for resources.

https://www.youtube.com/watch?v=lx6B7XRu3nk

This guy makes decent looking illustrations and I'll get his course in a few days. It seems Zbrush has some capability for comic style rendering, I hadn't known about this before today. Blender can give you freestyle edges, and separate out the passes for later compositing in Photoshop or CSP, so it would be a waste of time to take a screenshot and trace over it. By now I am familiar with texturing, but not so much with non-physical based rendering (NPR) so I'll have to look into that. I've made up my mind, so rather than use the sketch I posted, I am going to go all out in modeling the room in 3d.

https://blendermarket.com/products/mangaka

Instead of using PS filters which just work on single images, it might be worth looking for specialized programs that can take advantage of 3d information to create drawings such as this 4$ plugin. This is the first such program I've seen, and I bet if I looked beyond Blender I'd find more. Data such as depth can easily be exported from 3d programs, and there are bound to be ways of taking advantage of it beyond Blender compositor nodes.

(1/2)
>>
>>898101
https://youtu.be/fNQVdfRfMWg

I found it interesting to learn how pro mangaka do their backgrounds, towards the end he applies to process to a Blender model. The process is taking the image into PS, grayscaling and leveling out the image and then touching it up by hand. I am enlightened by this, I did find it strange how some mangaka who can barely get the proportions right on their characters could do such good backgrounds. Once the midtones are leveled out of the picture, that makes it easy to integrate manga style characters into the scene.

https://youtu.be/MtznoRM5Z0I

To cap the post, here is the documentary on Inio Asano demonstrating his workflow with commentary by Naoki Urasawa. His style is into realist backgrounds and cartoon characters. On the 3d side he is into Unreal Engine. It is too bad the video does not go into his 3d workflow, instead he is shown taking photographs.
>>
File: 18-05-2022.png (194 KB, 758x845)
194 KB
194 KB PNG
I don't think I'm competent enough to finish this project
>>
>>898111
Is that supposed to be Cell? What is the problem?
>>
File: StillWorkingOnHips.png (1.66 MB, 1590x878)
1.66 MB
1.66 MB PNG
I'm glad this thread is still around.
How else would i never get anything done and just talk about my projects.
But doing 3D modeling for fun is hard when your day job consist of 60% CAD, 20% telling the design guy to stop requesting impossible things, 10% getting yelled at by management for over-engineering things with way to expensive parts. The other 10% are spend on the toilet rethinking my life choices.

But i ma still chipping away at this space suite thing. With some sporadic sprints where major progress is made and other days where i just shift around some vertices.
At this pace i will get done when the chip crisis ends. So most likely in 5 years or so.

Also some thoughts on this cluttered mess, that is supposed to look like an exoskeleton assisting in high gravity environments, would be appreciated.
>>
>>898115
Also here an shitty turntable.
Love it when converting to webm tipples the file size.

And it most likely still won't loop, but 30fps animation and 30fps screen recording should line up.

anyway, i feel like i have way more detail in the back than the front so far. Let me know what you guys think about that.
Also i am trying best to keep everything human shaped, so nothing sticks out, everything is just adding thiccness.
>>
File: noeds.jpg (179 KB, 1931x859)
179 KB
179 KB JPG
>>898043
I don't use Blender so it might not work as is, but I did make a mock up of the logic in the shader editor. You said you've followed some toon shading tutorial so I assume you're already familiar (used at least once) with it.

Briefly: the angle between the camera view vector and surface normal can sort-of be found by taking their dot product, since both vectors are normalized (have unit length) the value returned is the cosine of their angle.
The cosine ofc can be transformed into the true angle by taking it's inverse cosine, but that's not necessary here.

If the vectors are pointing in the same general direction the cosine will be positive, but if they point in opposite directions it's negative, so to avoid blasting away the entire outline mesh we need to make sure the sign is never negative, so we take the absolute value of the cosine.

Then just remap the value as desired, I did a stupid and made almost everything black, really you'd want mostly white with a tiny crunched black region to the left.
>>
>>898115
>>898118
It's improving, but it's lacking that SEX factor, you know what I mean?
>>
File: fan_wip.png (1.9 MB, 1600x1200)
1.9 MB
1.9 MB PNG
Working on this fan after the lamp and monitor now. Thinking about trying out hard surface modeling or modeling even more complex props after this. Making environments was cool but I don't think that they are good for starting out as I should stick to some fundamentals first. Thanks for reading my blogpost, cheers.
>>
>>898132
Just wait until i attempt to make a face for this. It will kill more boners than than the Chinese people in reeducation camps.

Yeah i get it that its not that sexy, but thinking about it. I have many components, that are inspired by them skimpy armor sets found in many games. Like the Orange chest piece. The Straps running everywhere.
Also my blockout for the leg mechanism tries to exaggerated the overall curvyness of the design.
I Think a big problem is the stuff stacked over the buttocks. Completely obscuring the rear end. And my plans to add more hoses and such to the suite to further break up the lines won't help.
At least the helmet lends itself to shove in a pretty face to compensate for a lot, only downside i suck at that. And i don't feel like putting some face form the internet into that, making the model less than 100% my Work. I know nobody cares about licenses if you obscure stuff good enough but nah. Better keep it clean.
But if anybody has some ideas to make it more coomer oriented let me know. A bit more sex appeal cant hurt.
>>
>>898132
From the thumbnail I thought she had extra sets of nipples like a dog.
Not gonna lie, I'm a bit disappointed. She is kind of a bitch after all.
>>
File: noedz.jpg (167 KB, 1139x1231)
167 KB
167 KB JPG
>>898043
>>898120
I was curious and had to double check if it works, turns out the camera vector and normal vector are in different spaces by default (camera space vs world space) so a transformation is needed before the math can be done.
>>
>>898136
well, you've got a fan now
>>
File: farmer06.png (535 KB, 1351x892)
535 KB
535 KB PNG
>>898120
>>898162
I really appreciate the help. You're pretty clever.
Your post introduced a lot of words I'm unfamiliar with like vector, cosine, and dot product. Along with many nodes I hadn't used before. So it took me over an hour just to get a vague idea of what you meant. There was a lot to unpack, as they say.(I never get to use that line. Seems appropriate now) I didn't get to cosines in school. My math knowledge extends as far as solving for X in basic algebra equations. I use to know how to solve for the area of circles, but I forgot that.

Still, the information did start making sense to me in a vague way. I believe I get the broad scope of what's happening. However, it's a good thing you followed up with part 2. Because I was stuck, and there was no way in hell I was going to figure out the vector transform thing on my own. I recreated your set up, and I got it working. But now I don't know what to do with it lol. What was the point in all of this again? The idea slipped my mind.
Perhaps it was to somehow extend this effect out from the mesh.
>>
File: Farm-Line-Animation.gif (94 KB, 548x710)
94 KB
94 KB GIF
>>898162
>>898168
Alright, check out this gif. This is my most desired goal right now. I would love to learn how this is possible. There HAS to be a way. It makes too much sense. If only I knew how to find the solution. It's tough digging through tutorials, when most of them explain the same process.
>>
File: Tunnel Door.png (205 KB, 694x672)
205 KB
205 KB PNG
I've been working on a Metro 2033 fan game that I started in for my capstone project. I wanted to do prerendered environments so I'm forcing myself to learn blender, however my knowledge of modeling comes from solidworks so I'm really struggling making the jump. Here's my very baby tier attempt at one of the hermetic doors. Door and surrounding tunnel are separate objects.
Here's the reference: https://i.imgur.com/2bBRZZF.jpeg
If someone would humour some questions: at a higher level what's the best way to organize this scene? There are a lot of elements in the scene and I'm not sure which objects and their details should be separate beyond the different elements that need to be animated. IE: door and door hinges, screw assembly. locking bolts, wires, ect.
How do you model things with dimensional accuracy? In SW everything is relational and drawings can be bound to an axis or other reference geometry. In blender it seems a lot more like: "looks good enough".
Any other helpful tips are welcome or direction to where to post if this is the wrong place to ask.
>>
>>898179
In blender each one of those squares you see on the grid represents one meter irl.
You can group objects that belong together in the organizer (top right). Of course you could also just combine all the objects that make up a door into a single object. You can still edit them individually in edit mode but when you select the door in object mode in 3D view all components are grouped together. Combining and splitting meshes out of an object is easy to do.
>>
File: noydz.jpg (428 KB, 2907x1288)
428 KB
428 KB JPG
>>898168
>>898176
You grasp things remarkably quickly for someone who's yet to be exposed to trigonometry and vector math, kudos.

The original idea was to cull polygons from the outline that are at grazing angles relative to the camera. I got the culling to work in pic (just the reverse of what was originally proposed), had to google how to make Eevee render transparent objects kek. DESU I'm not sure if it's such a great approach as it will cull polygons from all over the outline, not just the problem areas, might look fugly in the end.

I'm not sure if this idea would work, but maybe scale the displacement of the outline by the dot product of the camera vector and the normal? See my poor drawing. That way you can get a narrow band along the "equator" relative to the camera that should face it from any angle and thus be more visible. IDK if the shader graph can change vertex positions like this, most likely you'll have to deep dive into geometry nodes to get it to work.
>>
File: Door Breakdown.png (1.01 MB, 800x600)
1.01 MB
1.01 MB PNG
>>898181
Thanks, so would this be a good way of breaking up the door? Where each of the outlined parts is a separate object, grouped into the door collection? Also that covers rough unit size but what about angular accuracy? In SW it's easy to grab the dimension tool and set the reference angle between two lines to whatever you need.
>>
>>898182
>sees the little "displacement" socket in the material output node
Ok forget geometry nodes, it should be doable with just a standard material
>>
File: Door Indentations.png (631 KB, 500x671)
631 KB
631 KB PNG
>>898183
Another question: should these indentations in the door be modeled as part of the door, be a texture + bump map, or some other strategy?
>>
>>898207
If it's supposed to be for something real time like a game then it should not be modeled. Unless that model is baked down to a lowpoly after modeling.
>>
>>898211
It's for a prerendered asset in a 2D game.
>>
>>898183
I would model each part individually and then parent them to the door, move into the same collection (if I'm keeping my project in Blender).
I'm exporting to a game engine, CTRL+J to join them, uv unwrap so it uses the same texture/material.

>>898207
I'm a big fan of adding details on a high poly version using a multires modifier and then baking it down. I would use a displacement modifier with a displacement map made externally (photoshop) to get those straight lines for the grooves.
Then use a baking addon (because native blender baking is AIDS) bake PBR textures (diffuse, normal, etc) and use the low poly version.
>>
File: BKn_S017.png (952 KB, 1963x1655)
952 KB
952 KB PNG
wip character sculpt I´m working on still very basic but im kinda struggling taking it further, if i should add details or retopo and add details later. Before that i still need do define the base shapes of the figure more
>>
>>898252
>I would use a displacement modifier with a displacement map made externally (photoshop) to get those straight lines for the grooves.
Lol, why? Literally just model them in. Should even take a minute.
>>
File: RAMEN SHOP WIP.png (993 KB, 1211x746)
993 KB
993 KB PNG
saw a anime wallpaper the other day and figured i should invest some time into practicing and really getting a realistic look out of my renders. started on this last night.
>>
>>898254
I want to fuck your sculpt, anon... in other words, great job so far :)
>>
File: file.png (72 KB, 465x441)
72 KB
72 KB PNG
Im afraid of fucking this up, specially in the legs
>>
>>898258
Because I've given up on baking from high to low in blender. If I don't use a multires modifier then I end up spending more time tweaking the extrude and ray distance than actually making the thing.
It's that sad.
>>
File: ramen render small.png (1.24 MB, 1210x749)
1.24 MB
1.24 MB PNG
>>898297
got some textures started.
>>
>>898345
>>898297
Looks good, I don't like the wood texture and the ... papers? need more work
>>
File: balcony door v2.png (797 KB, 1920x1080)
797 KB
797 KB PNG
>>897329
Did some refinement of the balcony door model today. Outwardly it is not a lot different, but I worked a lot on the smaller details. I also redid the door (not shown in the image) as the last time splitting it into two did not work correctly. I needed some advice and the Moi author was very prompt at answering my question. Thanks to that and the practice I did today, it feels like my skills have gone up.
>>
File: balcony door v2 clipper.png (452 KB, 1920x1080)
452 KB
452 KB PNG
>>898354
Here are some smaller model pieces.
>>
File: balcony door v2 hinge.png (1.05 MB, 1920x1080)
1.05 MB
1.05 MB PNG
>>898355
Here is the last one for today. Initially I made a mistake drawing curves in the air which is tough instead of creating a box and then projection cutting them like in Zbrush. In Moi it is possible to set a hotkey to set the construction plane to the current view with makes that possible from any perspective.

When I am done modeling the room in 3d, I am definitely going to take some time to practice drawing on these models. I've gone over my failures and concluded that getting good at 2d is simplifying things to their fundamentals. I might not be able to paint, but I can get good at drawing primitive shapes directly, as well as the curved surfaces like the ones here. The matter of leveling up my 2d skills will simply be a matter of composing those. I can't really think of a way to get better at painting in a similar manner, so I can only conclude that I am in over my head with that.
>>
>>898297
I did a ramen shop a few years back and posted wips here.
It looked pretty similar in terms of layout and dish placement, I had to do a double take.
creppy
I never did actually finish that project, I really should though.
>>
>>898254
Ah, I remember the granny thighs well on my own first sculpt. The trick to fixing them is to realize that in pretty women they should be taut towards the sides instead of falling evenly down the middle.
>>
>>898343
height to normal. there is an addon for that too
>>
>>898254
what >>898334 said
>>
File: Test 3.jpg (2.76 MB, 8640x2430)
2.76 MB
2.76 MB JPG
>>896917
I fucking love chairs here's 120.
>>
File: articulatedspine.jpg (1.11 MB, 2448x3264)
1.11 MB
1.11 MB JPG
>>898115
As someone who uses CAD and knows 3D modeling I have a question for you. I have been trying to articulate a sculpted body to make a figure, but I have a hard time trying to mesh a workflow between solidworks, blender, and zbrush. At what point should I be using solidworks or blender for trying to make an articulate figure for 3D printing?.
>>
File: zbrush 001.jpg (369 KB, 1118x839)
369 KB
369 KB JPG
Started this in 3d coat. Abandoned it out of boredom.
Got my first 3d printer (a used photon zero) and I want this to be my first print (after the recommended test prints).
>>
File: PizzaTime.png (283 KB, 1148x534)
283 KB
283 KB PNG
Pizza Time
>>
File: farmer08.png (1.03 MB, 1440x836)
1.03 MB
1.03 MB PNG
>>898182
>>898184
ok. So, I couldn't make sense of what you were saying. And basically I didn't even register what a "geometry node" was. My brain categorized it as something that was too in depth, and I forgot about it.

So I went on a search to try and fill in that eye gap. And I thought that perhaps if I can extrude all the edges of my mesh and then make only the faces facing my camera visible, then it would look good. So I checked all of the modifiers.(Fucking skin modifier caused blender to freeze) I figured wireframe was the closest to what I needed. So I played with materials to try and get the wireframe modifier to only be visible when facing the camera. And I think I got it working? But that just created a bunch of ugly lines throughout the mesh that faded in an out of existence. It wasn't working

So then I searched harder to find a way to extrude edges properly. I was trying to look for more modifiers. Seeing if there were any add-ons for modifiers. Eventually I put the correct sequence of words into google to come up with "geometry nodes." That's when I finally took notice.

I did the blender guru's geometry node tutorial with the sugar coated gumdrop. But then I noticed there was no extrude node. Turns out, I needed to upgrade Blender to get that node. So I finally upgraded to 3.1. And applied the extrude edge node. It was going terribly. I couldn't get it to do what I imagined. I eventually realized I was reproducing the reverse hull technique with geometry nodes, which got me thinking about how the normals are reversed in the extruding geometry.

And then it struck me... I was being stupid! All I needed to do to fix the eyes, was bend the mesh in a way that the reverse is visible. All it took was one edge loop, and a push, and the reverse normals filled the gap! It's not perfect. But tolerable. With only small gaps at the ends. I have to think about the inverse.

Part of me still wants to dive into geometry nodes to find a better solution.
>>
>>898437
Good job m8, looks slick. Nice fit-for-purpose solution too.
I was going to implement my idea and post a follow up, but it seems Eevee can't do vertex displacement (wtf is this 2005?) so I dropped it.

Geometry nodes look pretty cool and if you just move vertices around, not create or delete them, then it will be 100% replicable in any modern game engine vertex shader, so the effort in learning it is definitely not wasted because the fundamental ideas transfer over.
>>
File: farmer09.png (547 KB, 1461x824)
547 KB
547 KB PNG
>>898437
Also, here's the inverted hull method using geometry nodes. It works using the same concept. But seems to get slightly different results. I'm not sure why. But in any case, maybe there's something here that can be worked. With geometry nodes, perhaps one can build a better way to draw lines.

Also, also, I kind of feel bad for posting so much in the thread. It probably seems spammy to post for every small update. But typing out my thoughts is helping me stay focused. And I'm secretly hoping to get advice.

>>898444
>then it will be 100% replicable in any modern game engine vertex shader
That's what I'm hoping for. When the model is done, I want it to look like something a game engine can actually handle. So people are like "oh shit! Someone is making a new HM game?" And then they get disappointed when it's only a mock up.
>>
File: Unity_7OqJVg6Xn3.webm (1.43 MB, 1056x818)
1.43 MB
1.43 MB WEBM
Feast your face.
>>
File: file.png (87 KB, 206x156)
87 KB
87 KB PNG
>>898550
>>
File: blender_mBKgJKqubk.webm (1.72 MB, 916x794)
1.72 MB
1.72 MB WEBM
>>
>>897000
Absolutely incredible. Great job
>>
>>898445
I know "UE5 remake of an n64 classic" is a meme but I'd legitimately be so pumped for a HM64 that kept the isometric and just made it beautiful.
Your updates have been a delightful insight into how much work goes into the little things for someone like me that's super new to modeling. I say keep posting.
>>
File: ramen shop final small.png (1.73 MB, 1281x792)
1.73 MB
1.73 MB PNG
This is the maximum my pc can handle, I almost wasn't able to render this because my PC kept crashing :/ I also couldn't add depth of field cause the program crashed every time I tried.
>>
>>898581
doesn't look complex enough to cause those issues on a modern computer.
maybe you need to work on your optimization
>>
>>898581
Looks damn good anon. Is your pc bad?
>>
>>898582
I have a 1050 lol
on a 6 year old PC
>>
>>898585
old but shouldn't be too old to render that scene
got a wireframe?
>>
>>898587
I might throw one up later, but im afraid to open the file ': o
>>
File: WrinkeldTubes.png (740 KB, 1181x640)
740 KB
740 KB PNG
>>898402
My workflow goes like this:
>Take / Make the free form 3D model to meet the visual design. (Blender)
>Prototype all critical Dimensions for the Manufacturing process. (CAD)
Your pic looks like resin printing. So make sure your printed joints work and such.
>Fit Visual Model to mechanical model. By exporting to CAD.
Depending on the model, like a figure you can get away with having the solid parts as Meshes and modeling the joints Parametric for accuracy in between.
Depending on CAD package you you need to join your freeform model and your precision parts into printable meshes.
>Optimize for manufacturing. Like reducing Detail for a clean print. DFM Design for Manufacturing.

Tl;dr Use what ever tool is the fastest / most convenient. And you only need precision modeling for some critical dimensions.

Nobody cares if your figures legs are .0 something mm off. Your snap fit joints need to work. And i would prototype them separately in CAD.


Also for my model i'm currently experimenting with some wrinkly cable sheathing to make it look less modern. Opinions on that?
>>
>>898390
I have reverse autism and those chairs are too perfectly aligned.
>>
>>898607
not that guy, but i like your work anon. nice n clean. your cloth's folds are a bit too evenly distributed tho imo.
>>
>>898613
also, their directionality is all one-way. some break up would be nice.
>>
File: HOse.png (710 KB, 1923x634)
710 KB
710 KB PNG
>>898614
currently its just a shitty corrugated hose modifier set with some displacement noise thrown at it. To see if it will fit.
Its a bit very bulky but at this point, its going to be a very thicc suite anyway so why not double down.
>>
>>898550
'ate furfags but your character is pretty cool, nice job
>>
>>898581
You either have to much polys or too large textures. Are you using 4k textures for everything here? To get the poly count turn on statistics in the overlay.
>>
File: 1635350574428.png (221 KB, 995x903)
221 KB
221 KB PNG
Making some totally tubular props
>>
File: Screenshot_1.png (678 KB, 1005x753)
678 KB
678 KB PNG
I'm making a muppet-like character for someone.
It looks clean because I'm 100% following reference for the topology, not trying to innovate or try new things since it's for someone else (except the 4-finger hand).
I feel like a fraud.
>>
>>898614
If you look at spacesuits and their umbilicals the one way directionality is actually accurate.
>>
File: F_2(2).png (2.13 MB, 803x694)
2.13 MB
2.13 MB PNG
Im making a likeness of a certain celebritiy, I started with a low poly basemesh only and sculpted by hand from there.
Can you guess who it is? Is it close enough?
>>
>>898712
angelina jolie?
>>
>>898712
I can't figure it out. But tell me, what do you use to sculpt skin texture? Im guessing texture brushes right? What are a good source for those?
>>
>>898771
I only used default zbrush alphas, it was the first time I've done it all by hand and honestly I was super scared it would look horrible. But because I was trying to make a likeness I nedded a lot of control. I used the dam standard to define the big wrinkles at the mouth, the lips, and the corner of the eyes, for the small ones I used mainly the standard brush using the spray color mode and an alpha that looks like wrinkles, is it alpha 6? alpha 60? something like that, you can't miss it. If you drag it in one direction it gives you a directional look but if you drag it in circles it creates a leathery, more natural look. For most of the body I just used generic surface noise with low scale. next time I do want to get a tilleable pore texture to use as a general base for the whole thing.
>>
>>898750
nah Angelina is far more beautiful. The lips and cheekbones do look similar. This one's an adult actress, I should have specified
>>
>>898787
>This one's an adult actress
peak coomerism
>>
>>898787
Is it that punching bag that War Machine was married to?
>>
File: bed.png (755 KB, 1920x1080)
755 KB
755 KB PNG
>>898358
One more day of this and I'll have enough stuff in the scene to move on to the next step which is doing some sculpting on these Moi models to make them more lively. The bad looks a bit drab here even though the bevels are nice.

(1/4)
>>
File: shelves.png (543 KB, 1920x1080)
543 KB
543 KB PNG
>>898804
I am trying to figure out what this furniture piece is called, in the interim I'll be calling it shelves.
>>
File: fax.png (512 KB, 1920x1080)
512 KB
512 KB PNG
>>898806
I get the sense there is something wrong with the normals for this fax model. Here it is from a good angle.
>>
File: hdd docker.png (407 KB, 1920x1080)
407 KB
407 KB PNG
>>898807
This one is taken out from my PC case and has been standing there for years. Moi is a good choice for modeling these objects as they've been created in CAD software in the first place.
>>
>>898807
Nwm, I thought headlight would be the default like in Blender, but it is not. It is possible to change it in the lighting options to headlight.
>>
>>898712
Lisa Ann.
>>
File: BKn_S018.png (1.19 MB, 2289x1615)
1.19 MB
1.19 MB PNG
>>898334
>>898379
Thank you, hope I dont disappoint
>>898362
Thank you for the feedback, I tried to follow you advice and have the thights move a bit more outwards
>>
File: mov.webm (1.15 MB, 1120x864)
1.15 MB
1.15 MB WEBM
i dont know why i made this
>>
>>898712
Looks like Michelle Obama, 2bh
>>
>>898823
because it's neat.
good job anon
>>
File: 8117811574.jpg (91 KB, 1246x1093)
91 KB
91 KB JPG
I DID IT BROS
>>
>>898821
Definitely better than last time, but I am not sure if that is what I meant. Let me give you a reference. I can't be sure, but your problem might simply be that your reference is not good looking enough.

https://youtu.be/VzMAh66ofq0

I think the hips need more work, also the back of the thighs and the ass. Think about how as people get older the muscles and the skin start to sag, and look at the back of the things and see that happening in your sculpt.

The things also give off a masculine vibe with how they are constructed, if I didn't see the back I imagine you going for a masculine construction on purpose.
>>
>>898821
You look like you could use some references. So I'm sharing a few of my personal favorites.

I like this one because the shadow hits the contours of her middle section in a nice way. You can see the structure of her hips and stomach without it looking like an anatomical sketch. Also, her tits are large, yet perky. A good example for drawing big well structured tits without them appearing fake.
https://files.catbox.moe/5xg056.jpg

This one is just a nice back and butt. You can ignore the weird line pattern. I don't know what that's about.
https://files.catbox.moe/5ie73p.jpg

Back and shoulder rotation. Allows you to see the muscles working.
https://files.catbox.moe/86sd83.jpg

Basic skinny girl with a nice bone structure. You can of course deviate from the reference to give her fuller hips. But skinny people are good to reference because you can see the structure that is underneath the curves. You want a good underlying structure.
https://files.catbox.moe/gbjjn1.jpg

Classic huge tit model busty merilyn(she has various names) Photographed standing at various angles straight on. So you know what huge tits looks like without deceptive angles.
https://files.catbox.moe/ycoeoi.jpg

Another classic busty model, Katie Owen. Just spectacular tits on a small frame. You should look up more of her for a great tit shape in relation to her torso. I think her figure is the one you're going to benefit the most from studying.
https://files.catbox.moe/mvs36q.gif

Also, look up Lucy Wild. Because she's the full package. Amazing tits, amazing figure with wide hips and a full butt. A few things to note about giant tits, is that in order to make them appear natural, you have to sacrifice the visibility of the rib cage. I'm sorry, but it will just look uncanny to have HEYUGE tits, sitting atop the ribs. Naturally, their size would overlaps the upper rib and ab area. Otherwise, they just look like bimbo plastic tits.
>>
>>898821
>>898855
You should make her navel smaller.

https://files.catbox.moe/l6wezw.png
Also, I don't know how to explain this with words well, so I apologize for drawing on top of your image. But there is this zone around the pelvis. I don't know the name for it, but I call it the panty line. Because it has the same shape as panties. And quite literally, a line forms there. It's a subtle line. There are a couple of lines that form on the stomach. And when you're skinny, you can see two lines indicating a separation between the groin area, and the navel area. I guess these lines are from the fatty tissue that cover the muscle. Because you don't see them in muscle anatomy charts. But look up any photo of a skinny person, and you'll see these contours.

I think you should build that panty line area to make the stomach and hips look more attractive. As it is now, her stomach is looking rather droopy. But if you taper it a little, build the sides up more so the curve of the stomach is more subtle, and create the panty line contour zone, then it should be good.... I think.
>>
File: F_5.(1)png.png (1.5 MB, 1080x1080)
1.5 MB
1.5 MB PNG
>>898819
That's it! Thank you for noticing, I hope that slapping some hair on her makes her more recognizeable. If anyone has any constructive feedback feel free to give it.

>>898788
Not really it's more of an artistic study of a non traditional face, adult actresses have a ton of reference one can use, and from more extreme and diverse angles than normal celebrities. I mean I still coom but theres easier ways to do it.
>>
>>898855
The disadvantage of using real girls as reference is that they aren't peak attractiveness.
>>
>>898900
Fake girls are just reinterpretations of real girls. It all filters back to reality. Some day, a Japanese dude back in the 80s was taking inspiration from fashion magazines. He used that as his basis for a beautiful girl. And then a generation or two later, new artists take inspiration from him. And then a generation or two later, new artists take inspiration from that guy. And that's how you end up with uncanny uguu alien faces. Because a style is built off a style that is built off a style.

Be the guy who takes inspiration from reality, in order to avoid your style becoming incestuous and deformed.
>>
>>898870
good likeness the skin texture is a bit too extreme though
>>
>>898901
Most girls are fake and just failed reinterpretations of some rare "real" girls.
>>
>>898901
Beauty is objective even though there are some variations between people in their appreciation of it, ugly people do get born in the world far more often than the sense of aesthetics varies. It is all about finding the right mental triggers built by millions of years of evolution and activating them. So for that reason there is no need to worry about staying true to reality. It is better to follow your sense of what is aesthetically pleasing and only use reality as a guideline instead of something to adhere to.
>>
>>898909
Beauty is not objective. It just has a lot of overlap that creates the illusion of objectivity. But I do agree with you in a sense, because that overlap exists.

Now...No one is saying to *adhere* to reality. I can't believe I have to state this. But of course I do, because everyone always assumes the worse, and takes everything I say to the retarded extremes, and then I have to come in and explain the nuances. The thing is, biology has already figured out most mental triggers for aesthetic sensibilities. Through thousands of years of natural selection. So you can learn a lot from reality, even if you deviate from it in some ways.

If you re-read my post carefully, then you'll noticed I never say that the references are the correct way to do things. I say that they're good examples for specific things. I even outright state "You can of course deviate from the reference".
>>
>>898911
>Beauty is not objective

DAS RITE! PEEPUL ONLY BE SAYING DAT BLAQ KWEENZ IS UGLI CUZ DEM SOSHUL CONSTRUKS!
>>
>>898948
>conflating beauty with sexual attraction
meme
>>
File: PSX_20220522_015254.jpg (2.23 MB, 1127x2289)
2.23 MB
2.23 MB JPG
I'm trying out sculpting on nomad and I feel like this is not looking the way I want it to.
>>
>>898870
doesn't look like lisa ann, has orc proportions.

eyes too wide set and large
lower third too long. cut some of the philtrum, she
>>
File: router.png (463 KB, 1920x1080)
463 KB
463 KB PNG
>>898806
I made various errors while modeling the router which caused it to take much longer than I thought it would. The problem was that the sides had a slight angle to it, so I could not construct the router using straightforward techniques like I'd anticipated. I should have just separated out the surfaces and scaled them a bit before lofting them back, instead I did all sorts of crazy things to achieve something that should have been quite simple. It took me the entire day to do these two, but they might not even be visible in the scene. Instead of being faithful to my own room layout maybe I should put them on the top shelf where they could be seen.

I am happy with how the model came out, but I really am going to have to learn how to draw as 3d is too slow if you are doing everything from scratch. Maybe I should be kitbashing, but when I was doing a different scene a few months ago I checked out online asset sites and they did not have that much of what I would've wanted.
>>
>>898901
>>898909
>>898911
>>898948
>They're arguing about theory
>>
File: usb.png (186 KB, 1920x1080)
186 KB
186 KB PNG
>>898961
Did a bit extra today. Working in 3d space is pretty easy, if time consuming. Modeling these objects is a good way of internalizing their form. There is no reason why I should not be able to draw an object as simple as this. Let me do the top shelf and then I'll be able to start that kind of practice.
>>
>>898132
I guess the sexy factor is getting mangled. Making the shoulders wrinkly did not help.
But here is my problem. If i search for image reference on "sexy space suites" and similar its mostly just latex fetish gear and nothing else. Like there are only so many possible permutation of skin tight suite with some umbilical connectors tacked on.

Currently trying my best to keep the silhouette curvy as best as i can.

But i need some help deciding if i should go more modern or more retro in the design?
And if i should keep the mannequin or attempt to make a face?

Also some overall critique would be helpful.
>>
>>898901
>Be the guy who takes inspiration from reality, in order to avoid your style becoming incestuous and deformed.
Perfect explanation of what anime became
>>
File: LA_BeforeAndAfter.jpg (246 KB, 2170x1080)
246 KB
246 KB JPG
>>898870
Update

>>898906
Thanks, I prefer to crank the detail and then dial it down, also once you add subsurface scattering it smooths the skin a lot.

>>898959
Thanks for the feedback, I tweaked the eyes, the shape of the head, and the position of the lips. They were very subtle changes but added together I feel like they made a huge improvement. Maybe it's not 100% there yet but I'm very happy of how it's turning out.
>>
File: Tunnel Door Hinge.png (76 KB, 644x442)
76 KB
76 KB PNG
How bad are Ngons for hard surfaces? For this hinge should I slice the body to create quads or does it matter? I am a literal potato and don't know best practice.
>>
File: kotsculptfinal.jpg (126 KB, 1482x2160)
126 KB
126 KB JPG
I hate myself.
>>
>>899000
If it doesn't affect the shading or your topology flow or your modifier stack then it doesn't matter. ngons or triangles on flat surfaces are usually fine.
>>
>>899000
Just add a triangulate modifier on top if you want to send it to a game engine or zbrush and you're good. As long as it looks good.
>>
File: Door.png (205 KB, 898x657)
205 KB
205 KB PNG
>>899004
>>899008
Thanks, I appreciate the advice.
>>
File: Skeleton Model WIP1.png (387 KB, 939x680)
387 KB
387 KB PNG
I like how the skull is shaping is so far and I had planned to try to make this into a full character model. But I'm not sure if I should just work on the skull by itself or try to make it a full body character cause I've never made a full body yet and I'm not sure if I should work on something like that on a whole different project or not.
>>
making daxter from jak and daxter
>>
File: daxter6.jpg (33 KB, 925x816)
33 KB
33 KB JPG
>>899018
>>
File: daxter6.jpg (25 KB, 816x811)
25 KB
25 KB JPG
>>899019
still very early
>>
>>899003
Why? Are you making furry porn?
>>
>>899063
That anon is right!
2D furry porn >3D furry porn
>>
>>898984
In order to be sexy, the suit should be form fitting (ex. Evangelion plug suits). The more stuff there is on it, the harder it will be to make it sexy. Somehow the suit gives an impression of being hard to move around in, possibly because all the exo-skeletal aids.

Might you, it is not that the model itself is bad by any means, it is a pretty good space suit.

It is just for example - take a look at the thighs, and imagine this was a woman, it would be a woman over 30. Not sexy.
>>
>>899076
To add to this, if you can't make the thighs sexy, then it would be better to obscure it completely with a skirt. The problem with the kind of spacesuit you are going for is the same as if you took a slim model and extruded it along the normals, it will end up looking fat regardless of how good looking it was before.
>>
File: smug jug.png (829 KB, 720x766)
829 KB
829 KB PNG
>>896917
made smug jug, did what I could in substance to get a glazed pottery look, probably finished with this now since I really just wanted to sculpt that goofy ass face
>>
>>899088
I really like this image.
>>
File: smug jug.jpg (89 KB, 500x575)
89 KB
89 KB JPG
>>899088
Now I'm looking at the reference a few hours later though I can tell how off it looks, I think this is why I don't usually finish something in one sitting desu

>>899089
and I like you anon
>>
>>898345
This is super adorable/calming.
>>
>>898407
>>898414
I love this place because we have varying skills like this posting in here and I love you both. I hope I can get good enough to make 3D renders of my partners drawings and then print them out so I can have them hold the engagement ring box.
>>
>>898581
Use a render farm maybe?
>>
I hope anyone reading this thread know that I look at each and every WIP and I'm happy that you're all posting. Keep up the good work. Meme or not you guys keep me workin.
>>
>>899090
I activate my magic card, Jug of Smug.
>>
File: Damnpokeys.png (290 KB, 484x962)
290 KB
290 KB PNG
How would I go about getting a better :D kind of face for this Pokey here? I know sculpting is an answer but is there one without sculpting? Same with the eyes, I did what I could with scaling in certain ways but it messes with the mesh a bunch.
>>
File: BKn_S019.png (1.08 MB, 2123x1607)
1.08 MB
1.08 MB PNG
>>898861
thank you for the extansive advice, the draw over is definitely helpfull, I made a qick first pass to get the panty area in.
>>
File: top shelf.png (612 KB, 1920x1080)
612 KB
612 KB PNG
>>898982
I think this is good enough for now. Modeling wise, the room is 70% complete. I'd want to do more detailing on the objects in this image, and after that go into sculpting mode for stuff like blankets, curtains and bed creases as well as various clothing articles, but I'll leave it like this for now. What I will do is scatter a bunch of cylinders on the bed on the Blender side, and then start drawing practice. I think at the path I am on currently, there is nothing stopping me from becoming a good 3d artist, but I know enough that 3d on its own will never give me the speed that I want. I want good results as fast as possible and that makes developing 2d skills unavoidable. If I can't develop them I'll abandon this path and get a programming job. I need to tackle my fear from my school days and prove that my lack of art talent is just a lack of interest and poor instruction from the teacher.
>>
>>899101
This looks a lot nicer now. Overall, the model is a bit boney, so it might use some smoothing out.
>>
>>899088
I think that looks really cool. If the handle was smoother I would have been tricked into believing it was real.
>>
File: room 70%.png (2.33 MB, 1920x1080)
2.33 MB
2.33 MB PNG
>>899130
Geo nodes make it easy to scatter objects in Blender. Right now the room has enough complexity for me to stop modeling it for a while and start exercising other aspects of art.
>>
>>899088
Genuinely the best thing I've ever seen in one of these threads

>>899100
Boolean modeling
>>
>>899076
Some more detailed thigh critique pls.
I'm not sure which way to go. More thicc or less.
Or just give up and put some baggy pockets over them.

Without the gear, and the silhouette bulking it up i'm not that far off compared to a plug suite...
>>
File: Thighs.png (1.51 MB, 1584x869)
1.51 MB
1.51 MB PNG
>>899166
forgot pic
>>
File: try again.png (1.52 MB, 1445x1041)
1.52 MB
1.52 MB PNG
Going to redo an old base mesh I made, anyone want to tear into it a bit or offer some advice? I've already taken my own notes but other perspectives are always helpful.
>>
>>898993
to my eyes the updated one is a much closer resemblance.
>>
>>899088
nigger that unironically looks like a real (smug) pot
>>
File: ite chill nigga.jpg (41 KB, 500x483)
41 KB
41 KB JPG
>>899088
dude make this
>>
File: 0001-0099.webm (1006 KB, 960x1080)
1006 KB
1006 KB WEBM
I've "finished" the dragon, i sculpted the cloth and I added these small details to the wings, it is a model to print so I'll join remesh and split it to fit the printer and that will be it, I'm scanning a bunch of poxwalkers now, the first ones looks like shit so instead of cleaning them i will use them as reference to sculpt.
>>
File: Poxwalkers.png (2.15 MB, 1920x1080)
2.15 MB
2.15 MB PNG
>>899182
And these are the scans
>>
File: Mk3.png (355 KB, 835x739)
355 KB
355 KB PNG
Making a new model to stream with whilst I'm banned from tw*tch
>>
>>898826
What did you do?
>>
File: room in eevee.png (2.29 MB, 1920x1080)
2.29 MB
2.29 MB PNG
Ugh, I didn't have any experience with indoors scenes so I wasn't prepared for the rendering nightmare those grills would give me. I guess there is a reason every indoor scene that I've seen rendered in 3d had big windows. What is happening is that the noise is huge in the rendered image and the room is much darker than it should be to boot. Literally, the only way to render it in a sane amount of time on my rig will be to open up those grills wide because right now they are trapping all the light rays inside the room.

I wish I could post a screenshot, but it seems the Cycles ones are over 4mb, probably because of all the noise. Here is one in Eevee. Ambient occlusion makes some areas brighter so it does not get much points for realism. I am going to have to think what I want to do here. Since right now I just want to get some drawing practice in, maybe I'll just get hide the balcony doors and let the light stream in. After that I'll put some real effort into getting this to look good using non-physical based rendering.
>>
File: spacesuit thighs.png (116 KB, 413x334)
116 KB
116 KB PNG
>>899167
I've marked the inner thighs with red curves. Note how they are almost flat as they travel down to the knees.
>>
File: asuka thighs.jpg (54 KB, 532x810)
54 KB
54 KB JPG
>>899199
Now note how the thighs curve for Asuka here. I won't highlight it, but you should get what I am getting at.

>>898827
Compare your own model's curvature with the ones by Flycat, and you'll see how stuffy your thighs appear.

You have an A pose and he is doing a T pose, but your model had closed legs, it wouldn't have this willowy look.
>>
File: BKn_S020.png (1.34 MB, 1905x1609)
1.34 MB
1.34 MB PNG
>>899131
I added some more meat. I tried to give her some skullcrushing thights like Chun Li. But I guess next I should moveon and retopo to make her usable.
>>
File: ThighsCurve.png (519 KB, 1000x462)
519 KB
519 KB PNG
>>899199
>>899201
i don't fully get it. But yet another excuse to look at porn for "reference"...

I found some examples, but they seem a bit overly exaggerated. But i think some middle ground between will help.
I guess this is one of the problems that arise form using a rather low poly base mesh and subdividing / smoothing all the form away. Too smooth.
But ill see if i find some more references.
>>
>>899217
your model is like the extreme end on thigh side. Guess i have some shaping to do...
>>899218
>>
>>899217
cottage cheese body
>>
>>899217
The heel bone is still a bit too prominent. Also, the ass is dropping inward, and it seems to be asymmetrical. Am I just imagining the last one? Also, if you look through the thigh area from the back, the crotch should be slightly visible.
>>
>>899218
I'd recommend getting a copy of Anatomy for Sculptors, it has a lot of neat diagrams and cross sections to get a better idea of how limbs look like from different angles. Also you should watch how and where sculptors add and remove volume from different areas, zbrush's official youtube channel has a bunch of videos by Sakaki Kaoru which I found really helpful even if you're not sculpting. Although you should learn sculpting, it's just a better tool for anatomy. In your case I think you're only missing a couple of clay strip brush strokes and a smoothing pass over the major muscles.

Good luck, you're GMI.
>>
>>899217
I know that this mainly a styalistic thing. But it really bothers me when people are depicted with forearms that are larger around than their upper arm. That almost never happens. It's a popular trait for cartoony figures. But I think it looked kind of goofy. If I were you, I'd Bulk up her triceps area more. That would give her arms better muscle balance.
>>
>>899250
subdivision
also this is not the dumb questions thread
>>
>>899252
Good point. Thanks.
>>
>>899088
how much do i have to pay you for that smug jug model
>>
>>899171
jesus christ man start with a normal model before you let your coombrain take over
>>
File: wheel topo.jpg (298 KB, 1168x1131)
298 KB
298 KB JPG
>>896917
Rate/roast my topo. It's a wheel

What's wrong with it? what could be done better?
>>
>>899019
>>899020
love it anon, very cute.
Used to love playing Daxter as a young 'un.
>>
>>899226
Some dubious links later, and i now actually own some "professional" reference pdf.
(1) The original. I now see the issue, if i hide all the crap build around, you can put a ruler next to that.
(2) My half assed attempt to replicate the Real anatomical reference given with the book (somewhat high body fat, why no female body builders but only males to show off muscle groups ??), looks a bit lumpy.
(3) Some more smoothing on 2 and less ugly topology flow.
(4) Hentai reference, maybe a bit too unrealistic.

Need some opinions on what to refine.
On the topic of sculpting. I know the tools, use my phone with weylus as a drawing tablet, it works but is janky. I'm way faster / less performance heavy using proportional editing on regular geometry.
Everything on this model rather lowers sub-D workflow and i rather keep it uniform. I know multi res is a thing but i'm to dumb to make that work in a way i'm happy with.
>>
File: 1653392606166.png (499 KB, 1056x594)
499 KB
499 KB PNG
test render on maya
i think that neat
>>
>>899302
thanks bro (:
>>
File: robo_v1.png (1.43 MB, 1920x1080)
1.43 MB
1.43 MB PNG
Trying to make a robot, i'm struggling with hands... and also legs. NGMI feelsbadman.png
>>
>>899198
It really very hard to get the lighting the way I want it to. Reality really beat Cycles as a renderer. I only got it to work in a reasonable amount of time with Cycles by turning on the Fast GI Approximation. But is not even close to being good. I've put in a lot of effort into this, but now that the day is over I realize that my original plan of simply putting in a HDRI and having it work was the most sensible one. It has the background and the sun, but it failed and now I find myself messing with all sorts of things to get it to work. Those attempts are all hacks.

I got some good replies to a thread I posted on on Blender Artists. What I will try doing next is LuxCore. It's an open source renderer that is supposedly decent at complex scenes like my own. If that fails I'll just get rid of the balcony doors and let the light stream in.

Some light setting tips I've discovered today:

* https://blender.stackexchange.com/questions/118568/is-there-a-way-to-drive-lamps-similar-to-how-you-can-drive-and-control-a-ca
* https://engelik.gumroad.com/l/DQNYI

Also, the following vid is a got overview of various NPR techniques.

https://www.youtube.com/watch?v=X8YkWdhty7I

I'll want to try out Malt before I wrap up this adventure. I want to start the drawing practice already. I do not at all enjoy messing with lighting.
>>
>>899312
> if i hide all the crap build around, you can put a ruler next to that.
Yeah, I meant to suggest that you should hide the props on the model, but never got around to it.

I am not sure to suggest what to refine specifically on your model since I am not sure what a spacesuit should look like. But sculpting is a matter of perseverance. The only real sculpt I did was copy Flycat's base mesh. It took me almost a month while it took him a day, but in the end I got it to look good as he did. The beauty of 3d is that you can constantly refine your existing mesh.

I myself am reflecting why I failed to draw those bed creases in 2d, and the most likely reason is because I lost sense of the refining workflow and essentially just started scribling lines. If I was having trouble sculpting the bed creases in 3d, I'd have zoomed in and added extra geometry before getting it just right. But in 2d, if I zoom in I start seeing pixels and stop thinking. I should have taken a step further and opened a separate canvas and painted that tiny bed crease in full resolution instead of thinking something as simple as that would be beneath me. After that it would have been easy to downscale it and transplant it to a 50x50 grid on the actual painting.

The refining workflow is the secret to making good art. I have it in programming, I have it in 3d, but I need to grasp it in 2d. Once you have it, you can steadily get better through practice. Digital art is not about dexterity, if you lack accuracy there are all sorts of tools that can give you that. It is really about planning, much like programming.
>>
File: StylesStayTheSame.png (1.27 MB, 2088x915)
1.27 MB
1.27 MB PNG
>>899364
Went with a blend between 3 and 4, and a bit toned down curve. Somewhat aiming for an not too much above average body.
> I am not sure what a spacesuit should look like
Looking at some Latex Fetish gear, it smooths out the more fatty parts of the body.
Realistic Spacesuits are so baggy body shape vanishes completely.
Looking at the Martian Movie, i guess they had some good ones with good mix between spandex and bulky equipment.


But i dislike sleek and modern ascetic of the sci fi stuff around. I want to make some cassette futurism type stuff initially but somewhat missed the mark. Now its just a project ballooning in scope.
Initially started out as another of these "this time i'm going to finish something" Projects, recycling assets form abandoned projects. That i ended up remaking anyway to fit the techniques learned along the way.

Somewhere someone said about moves that an director always makes the same movie, telling similar stories in different settings.

At only thing that keeps me going on this is the quick prototype that serves as a rough guide on what needs more modeling. At least i see some improvement since about 2 years back when i started to get serious about 3d.
Had to find this old model in my badly organized data hording mess. But i had to simile to see a few bits made it through several projects and stuck around on the new model form the old one.
Best improvement is the detail and speed. Back then not knowing the tools it took ages to make anything. Now i just blitz through most technical stuff.
Made a bunch of stuff in between these 2, but the space suite had been bugging me for years now.

Also to anyone else struggling. The best way to improve is trying to make what ever comes to mind and recycle the bits and techniques that worked for the next project.
>>
File: mokrender3.jpg (465 KB, 1920x1080)
465 KB
465 KB JPG
>>899373
Also excuse the shitty weigh painting.
Was "good enough" for the final render.
>>
>>898993
she simply doesn't look like Lisa.
>>
>>899088
love it anon
you should add on a shapekey of it grinning and render out a lil animation it'd be kino
>>
File: sycra_anatomy.jpg (278 KB, 1024x1325)
278 KB
278 KB JPG
>>899312
I don't know if starting to add anatomical detail to your model at this stage is a good idea. You're busy modeling a space suit for the character. I think it's best just to stick with your original model and finish the character. She has a simplified anatomy but that's okay.
Like that other anon mentioned, 3D involves a lot of planning.
If you want to create a sexy looking character with better anatomy it's best to start of getting your base mesh right from the start. See pic related. Many people sculpt a character by creating all these muscle groups and then merging them together to create a more realistic character design. Of course it's important to keep in mind the differences in proportions between the male and female bodies as well.
>>
File: hhead 001.jpg (91 KB, 616x750)
91 KB
91 KB JPG
>>899094
Good luck anon.

Here's an update. Re-did the top armor thing completely nd I'm feeling this one more.

By tomorrow I'll probably want to change it again.
>>
>>899416
I know the getting done part is important, but this model is supposed to push visually quality to the best if my ability. Sunken cost fallacy, sometimes you have to throw out work no matter how long it took to make in order to improve. I'm still progressively refining any part if I feel like it.
If I have a busy day I don't always start on a new bit and just tweak an old one.
Minor corrections like this are still doable with next to no effort.
For commercial projects its wasted time, that is true. But for hobby projects, in order to grow as an artist it's imporatnt to be able to change things when ever needed.
>>
File: damnyouroxy.jpg (210 KB, 1625x864)
210 KB
210 KB JPG
I have no real clue how to make that first bone piece for her staff. I tried blocking it out but I feel like I'm just not seeing the obvious answer. How would you guys go about making such a weird yet subtle shape.
>>
>>899434
Block it out and use the scrape/flatten brushes in sculpt mode.
>>
>>899437
I meant to ask if sculpting is my best answer or if there's another way to achieve these kind of shapes.
>>
>>899438
You could do it with subd modelling but achieving a good organic look would be a pain.
>>
>>898827
>Following this guide.
>Doing pretty ok.
>Takes a few hours just to cover a few minutes of the video, but this is ok.
>Think to self "I can probably improve on this area here, if I just move a few polygons.
>End up mired in a topological knot that I can't fix.
>Spending hours trying to reform proper loops.
>>
>>899440
Struggling means you're learning a new skill you don't already have or you're improving in a (hopefully) better way. Keep at it bro.
>>
File: smugspin-1.gif (3 MB, 512x512)
3 MB
3 MB GIF
>>899384
You mindreading fucker, I was just making this as I checked the thread again and saw your post
>>
File: smugspin.gif (2.66 MB, 395x401)
2.66 MB
2.66 MB GIF
>>899445
And a lil turnaround too

>>899263
Not sure anon, haven't sold my shit before, does 20 quid sound reasonable? Might put it on turbosquid or something
>>
>>899446
i'd buy it for 15 USD
>>
>>899440
What I did my model, I did not do the kind of box modeling Flycat is doing, but instead sculpting with dyntopo. That way is easier.
>>
File: Goodenoughfornow.jpg (250 KB, 1580x1195)
250 KB
250 KB JPG
>>899434
>>899437
Couldn't exactly get scrape and flatten to do what I think you were trying to suggest so I had more luck with just cylinders and blocking out more correctly. Had issues in the end but was still experience with extruding and scaling and beveling.
>>
File: BKn_S021.png (1.44 MB, 2097x1687)
1.44 MB
1.44 MB PNG
did some more reshaping
>>899446
that looks pretty good also it reminds me of the pot of greed
>>
>>899446
>>899451
I'd pay so I could 3D print it and leave it on my desk at all times
>>
>>899316
Hey, I know that bridge…
>>
>>899445
Based. I'm saving that for smug reaction purposes.
>>
File: room in luxcore.png (2.95 MB, 1920x1080)
2.95 MB
2.95 MB PNG
>>899363
Luxcore is slower than other renderers, but it can handle scenes the others can't once the caches are on. It handles those grills like a pro. Here is my reference, the only light source is from the HDRI. With this I should be able to start drawing practice. I regret not beveling the edges while I was modeling them. Neither Eevee nor Luxcore have a bevel (or round edges) shader, and Cycles which has it cannot handle this scene.

I tried out Malt as well, but right now it is too raw. It hit 1.0 so it should start stabilizing, but right now there are no learning materials for it. I'll wait a couple of months before checking it again. Right now I have no incentive to dive into it.

Phew, I am glad this is over. Not having beveled edges on a model is bad from a 3d modeling perspective, but it won't matter for a drawing reference. I'll try to tackle drawing by immitating a 3d program. Start with the wireframe, then hatch the face normals in the UV direction, after that the shading. I'll try avoiding tackling deformed surfaces like sheets until I am ready for them. I'll also try to avoid tracing the shapes outright as it would defeat the purpose of the training.
>>
>>899445
it came out great. i love it

>>899472
your doing a good job but they have an unusual fat distribution. if that's you're intent or the look your going for all good, but if not it's odd seeing well defined pecs but the back of the calves being smooth.
>>
>>899090
This should be a real product.
>>
>>899496
are those polygons on the bed?
oh no i'm going innnssssaaaaaanne
>>
File: slutt.png (627 KB, 1299x790)
627 KB
627 KB PNG
Here's some whore I sketched out today. It ain't much but it was some good derusting, I haven't touched Blender in too long
>>
File: sculpt.jpg (115 KB, 1291x893)
115 KB
115 KB JPG
took a break, back to grinding
>>
>>899567
Grind? You've already achieved perfection.
>>
>>899063
nah I just hate myself in general
and because I suck
>>
>>899496
Based and luxpilled, but what were your issues with Cycles? Did you try using the full GI preset?
Also if you want to lose a bunch of time you could bake Cycles' bevel shader.
>>
File: robo2.png (1.53 MB, 1920x1080)
1.53 MB
1.53 MB PNG
-_-
>>
File: smug1jar.gif (2.38 MB, 256x256)
2.38 MB
2.38 MB GIF
>>899088
>>899090
>>899445
>>899446
you killed it man love the jar here's my loop mod
>>
File: actors.png (2.45 MB, 1190x1806)
2.45 MB
2.45 MB PNG
>>
>>899591
Premium content
>>
>>899558
Randomly rotated and scaled cylinders. I put them in place of creases, later I'll just go in with a clothes brush.

>>899587
Cycles isn't smart enough to sample the rays in the right directions. This makes the noise high, the room dark, and the grills pure white in intensity. Using Fast GI approximation does make things better, but the shadows have poor quality. Installing Luxcore and turning on the caches as well as cache friendly noise patterns is both easy and gives the best results visually.
>>
>>899591
Oh shit it's old school animation anon, great work son
>>
>>899472
make that hips EXTRA THICCCCCCC please
>>
File: reduced_size.jpg (173 KB, 2688x1512)
173 KB
173 KB JPG
First attempt at sculpting a self-portrait. Very satisfied with how the eyebrows and hair in general came out. Skin didn't come out too bad but I have to work on giving it less of a "wet" look. Might remesh if laptop doesn't melt kek
>>
>>899607
Skin needs some detail normals for skin pores as well as painting an appropriate roughness map.
Painting a subsurface map would help a lot as well, just don't apply it evenly to the entire head like some amateur.
>>
File: Poxwalkers.png (1.39 MB, 1920x1080)
1.39 MB
1.39 MB PNG
>>899184
Today i've worked a bit on the first model, I'll have to retopo the coat, maybe to split it in multiple pieces could work too.
>>
File: DaxterFront.jpg (221 KB, 1500x1125)
221 KB
221 KB JPG
Daxter!
>>
File: DaxterSide.jpg (192 KB, 1500x1125)
192 KB
192 KB JPG
>>899619
>>
File: DaxterBack.jpg (197 KB, 1500x1125)
197 KB
197 KB JPG
>>899620
>>
>>899619
>>899620
>>899621
That's fucking awesome anon, great work
>>
>>899622
Danke anon means alot
>>
>>899591
is that you, world4jack?
>>
File: 7hh1.gif (432 KB, 480x270)
432 KB
432 KB GIF
>>899227
>it really bothers me when people are depicted with forearms that are larger around than their upper arm. That almost never happens.

Fat American detected.

Go and look at people doing dumbell curls from infront. The forearm will almost always be larger than the bicep from this angle. Sometimes not by much, but sometimes by quite a lot, especially in women.
>>
>>899628
In your own cherry picked reference, his forearms are in balance with his triceps. The forearms are a little bit bigger due to the angle. Yes, angle matters, but not that much.
You'll notice that when palms are turned forward, that's when the forearm looks the broadest. And this is due to the bone untwisting and making a convex curve. In the guy's model, her palms are facing down, so the convex bone shape shouldn't be very pronounced.
The area where the bicep goes into the forearm muscle creates a noticeable contrast in size. As you trace your eyes down the arm, it sort of tapers then flares then tapers again. But if you trace your eye up the arm, then you notice that the triceps flare out too. To about the thickness of the forearm.
So what I'm getting at, is that the flaring out of the forearm makes a kind of sense. But it's only relative to the tapering of the bicep. And it shouldn't be so pronounced that they look like Popeye the sailor's forearms.

In women the forearm is especially *thinner* not thicker. Women tend to taper more in the limbs. But it really varies a lot in body shape. For example, in especially skinny women, then forearm flare is going to be especially noticeable since they're all bones and sinew.
but I dare you to find an athletic woman who has forearms larger than her upper arms.
>>
>>899607
No way! We have black people on this board? That is so cool i didnot know that.
>>
>>899656
There are black people on every board, anon.
>>
File: BOOBRENDERIN.jpg (137 KB, 1480x842)
137 KB
137 KB JPG
BOOB RENDERIN'
DEFAULT CUBE DELETIN'
NO SKILL HAVIN'
>>
>>899496
Change of plans, I am going to try 'drawing' through sculpting. I'll never get into proper 2d, I feel completely lost with just a piece of digital paper in front of me. I've been thinking about it, the the way Zbrush for Illustrators guy does is how I should be doing it. I've started watching his course, and the way he made a crappy sculpt of Darth Vader work as a drawing is inspiring. I've decided that I should pick a toon shading style and make my models so they fit it. So far I've been learned how to do it in photorealistic fashion, but that way is too labor intensive. The models have to be good, and I have to do texturing and time consuming rendering on top of that. I could achieve a lot more if I took a step back from that. The benefit of the style will be that it both looks good and can be done quickly.

Even if it means pestering its author for how to use it, I am going to give shading and rendering the room using Malt a serious try. After that I will do some anime style chara modeling. Hopefully that kind of practice should get me to the pro level. Though I am disappointed at my drawing ability, since the year is 2022 I do not really need it, and I am not worried about my modeling and sculpting abilities since they are easier to develop. As much as it stings my pride, I'd hate it a lot more to become some kind of tracer even if believe that would be a viable path towards becoming a 2d artist. Since I am going for illustrations I have lot more flexibility than an animator would have, so I should be able to make my work look just the way I want it to.
>>
File: Slim B.png (896 KB, 1090x822)
896 KB
896 KB PNG
>>899669
That's cool.
Speaking of boob rendering: I made a pretty good boob shape yesterday. That was fun. Been spending all day today trying to copy a hand from reference. Wanting to get the topology right, but it's complex and confusing. Feels like solving a rubik's cube. At the rate I'm going, it will take a whole month just to get the mesh up to where I want it to be. The wait is killing me, because after the mesh is done, I want to learn how to make the body jiggle.

Toggled subsurface on to smooth it out.
>>
>>899674
>subsurface
oops, meant subdivide there.
>>
File: 1649539463879.png (257 KB, 787x841)
257 KB
257 KB PNG
>>899674
show wires
>>
File: Slim B 02.png (805 KB, 798x849)
805 KB
805 KB PNG
>>899687
Not sure if you're joking. Posting anyway.
I don't want to show more than this, because a lot of it is still unfinished. Legs are all wonky. Formless feet. Arm and hands misshapen. But it's coming together nicely enough. I'm building the topology by stealing from various references, and a little bit of my own "ingenuity" here and there. Mostly pushing points around manually. Occasionally going into sculpt mode to smooth things out and push around large portions.
>>
>>899588
You did that with a displacement shader or with geometry nodes?
>>
File: New Canvas.png (98 KB, 775x399)
98 KB
98 KB PNG
>>899619
he's got kind of a "gumby" pelvis. He could stand to have more division between his thighs and crotch.
>>
>>899674
just make big boobed lobster women and you won't need to worry about hand topology at all
>>
>>899688
i dunno if doing a quad patch like that early will work if you want nipples
>>
File: 191829.jpg (1.11 MB, 4256x2832)
1.11 MB
1.11 MB JPG
>>899632
>I dare you to find an athletic woman who has forearms larger than her upper arms
>>
>>899710
he said woman
>>
>>899701
Looptools -> circle -> problem solved
>>
File: untitled.png (1.49 MB, 1920x1080)
1.49 MB
1.49 MB PNG
>>899690
Hard surface modeling.
>>
File: CarRender.png (2.19 MB, 1853x907)
2.19 MB
2.19 MB PNG
>>
>>899723
youd still have poles
>>
File: Slim B 03.png (25 KB, 184x209)
25 KB
25 KB PNG
>>899701
>>899748
Guy who made the boobs here: I did this. >>899723 They look ok. I didn't know about poles until you said something. I just assumed they were ok. A bunch of small polygons converging seemed like the recipe to make a cone. So I didn't see a problem. It's good to know to avoid them.

Still, I didn't actually create a pole for the nipple. I created another patch to round off the end. I don't remember how or why past me thought to do that. Maybe it was in the tutorial? I don't remember it being in the tutorial.
>>
>>899747
Good job, but those normals are killin me.
>>
File: blender_WOBIkk2Vhr.png (780 KB, 1144x752)
780 KB
780 KB PNG
>>899763
try something like this, you generally dont want poles for cylindrical stuff
>>
File: Slim B 04.png (1005 KB, 1479x999)
1005 KB
1005 KB PNG
>>899774
Eh? Am I misunderstanding what a pole is? Your example is what this page told me to avoid. https://resources.turbosquid.com/training/modeling/poles/
I'm told that star pattern at the end of the nipple is bad.
I guess we're allowed to post nipples on /3/? I assume because it's academic or whatever, it gets a special exception? Well, let me just show you the whole boob, so you know what I'm working with. The front patch seems to work out well enough. That's how the tutorial shows to do it. Which I guess is to reduce the number of polygons that go on to make the nipple. It's probably more for triangle efficiency than anything. Just guessing.

The only thing I don't like, is in the base mesh, you can see the edges of the polygons that transition into the nipple. However, when subdivide is turned on, that edge pattern is tiny and nearly imperceptible. So I'm not too worried about it.
>>
>>899774
Pointy tits are hot and rare in the wild.
>>
>>899776
hit image limit, so ill use catbox
https://files.catbox.moe/eom2k5.png
picrel is what im talking about
>>
>>899778
Oh I see. Yeah, I kind of get it now. Those are the cause of the wonky shadows around the nipples on my model. I will consider changing that. But for now, I think I'll keep it the way it is. I like the idea of reducing the polygon count. It pleases that OCD side of me that seeks efficiency. It will probably take me some time to warm up to the idea.
>>
>>899781
is there any reason for reducing poly count other than autism, e.g. is it being used in a real time application? even then dont you think it'd be worth it to spend poly budget on something that the viewer will focus on
also shading trumps polycount
>>
>>899784
Its easier to control the workflow than having trillions of lines and points.
>>
>>899785
anon the alternate approach wouldnt be any more denser in polygons than what you have right now
>>
>>899788
I'm not that anon... he will answer u i guess.
>>
>>899784
>>899788
Well part of me wants to make a highly refined realistic model for posing and animation. And the other part of me wants to make an efficient but still realistic model that could work for games. I'm teetering between the two. I might just spend the extra polygons on the nipples. Because like you say, it's a feature people focus on.

>the alternate approach wouldnt be any more denser in polygons than what you have right now
The cylinder has 16 edges. The patch reduces the count 8 edges just before the nipple. If I remove the patch, then it will remain 16 edges all the way through the nipple. So the nipple would have double the amount of polygons. Or twice a dense.

>shading trumps polycount
True. But like I said before, when the patch is subdivided, the bad shading is reduced so much, as to be nearly imperceptible. Perhaps when I get around to texturing this thing, and giving it a realistic shader, then the patch will be completely obscured by the detail.
>>
>>899778
You could probably subdivide those specific verts and be good.
>>
>>899589
fukken saved

glad you like it :^)

>>899488
>>899451
brehs I had a look at shit like turbosquid and sketchfab and they take such a big cut so I'm out of ideas.
>>
>>899672
Done for today. It is really tedious to just watch tutorials, but tomorrow I should get to the NPR and Photoshoping parts. Today, I've been boring myself watching him model and the only thought on my mind was that could do all of this just as well as he could. Right now my problem is that I am split on whether I should dive deeper into shading arcana with Malt, or whether I should just extract the freestyle edges and the rendered scene and play around with gradient maps in CSP. I'll know soon enough where I stand. Malt strikes me as particularly powerful, but I know that messing with shaders is not the most direct way to go about this. I know that if I really dove into it, it could eat up a month or two of my time. If I was doing animation it would be worth it, but right now I have a lot leeway, too much in fact. I am really opening up to how much 2d stylization makes things easier - models can have bad topology, I do not need to bevel, UV unwrap or texture. Even the geometry does not have to be good, as long as it looks good past the NPR filters, everything is fine. If I want to make it, I need to focus on taking advantage of all of this. The question is merely, what is the most efficient way of doing that?

When it comes to rendering, my hunch is that just letting Cycles or Luxcore will be hard to beat, as they don't need fiddling around. I won't have to let them run for hours in order to clean up an image like I have in the past. Since I will be crunching the midtones, the benefit of that should be minimal. A few minutes followed by denoising should be fine to get me the correct lighting.

I am sure that the closer I can get to a workflow that focuses purely on modeling, the closer I will be to winning.
>>
>>899778
New thread when?
>>
OP here, working on the next thread image now.
Didn't think we'd hit file limit so fast (or at all), usually never happens. Guess it's a good sign that people are working hard.
>>
>>899921
Plot twist, someone beat me to the punch.
My bad guys!

New thread
>>899902
>>899902
>>899902
>>899902
>>899902
>>
>>899608
Thank you kindly, will work on that
>>899656
Kek anon. Well there are quite a bit but the topic doesn't really come up on certain Lithuanian saltmining radiostations





Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.