/3/ - 3DCG

>>587978

I got a folder full of these but it's literally just the Polycount hard surface thread, a lot of Pinterest posts, and some 2ch thread.

I'll see if anything stands out, I can't exactly dump 552 images.

>Matching segments in Blender
Better get it right the first time x_x

>>587988
Also the Loop Tools addon in Blender will let you add some more points and then smooth it back into a circle, with or without the points being an equal distance.

>>587989
Regarding equal distance in circles: It doesn't really matter as long as they're positioned right.

As much geometry as you need, and no more.

One of these infographics was a little messy so I redid the image. The faint lines continuing from the arch will continue to the outer control loop so you can bring that in tighter if you want to remove it.

>>587979
this whole pic is irrelevant when you can just increase the "edge crease" factor.

>>587984
>>587983
i dont get it, it seems like common sense
dont people work this way ?
i stoped doing 3d since maya 2009, im having problems with so many new toys available but edgeflow and etc should be a nobrainer.

>>587999
You can't get tight and smooth beveled edges with edge crease unless you add support loops to those edges, at which point you're already doing that pic.

>>588015
It's common sense after you build up your geometric literacy, need some things for noobs too.

Quick detailing.

>>588082
lmao just use nDo2, idiot

http://www.3dartistonline.com/news/2016/10/how-do-i-combine-blender-with-houdini/

Describes his hard-surface workflow which involves getting the general shapes in 3D with sculpting first so you know how to position planes. For me I have a hard time making complex swooping forms from the get-go and need something to help guide me like curves or a sculpt.

>>588084
>using GPL software with non-GPL
just...why?

>>588083
nDo2 is for retards who can't into modelling.

>>588091
t. substance shill

>>588092
>modelling in substance
no wonder why you like nDo2, retard.

>>588093
>modelling details in 2017
>not just using images, extracting the normal and nDo2
dunning kruger effect, lads

>>588094
>lads
who are you talking to? nobody is talking to you retard
>im too retarded for modelling everyone, ill just try to hack my way using shitty images rather than doing it the right way
Maybe you're just too retarded to comprehend that you're retarded.

>>588084
>For me I have a hard time making complex swooping forms from the get-go and need something to help guide me like curves or a sculpt.

That's fine.

>>588111
>>588093
samefag

>>588119
Yeah that's what happens when two people are talking retard.

>>588094
>"extracting" normals from images for "modeling"
kek im sure even your mum think your stuff sucks

>>588162
>reading comprehension

>>587980
Put all the images in a folder, turn it into a zip file and upload it to MediaFire or Mega.

>>588064
>You can't get tight and smooth beveled edges with edge crease unless you add support loops to those edges, at which point you're already doing that pic.

Surely thou jest. Green lines are where Edge Bevel Weight is set to 1, Edge Crease is set to 1 on the outer-most edges.
Thread's cool though, I think I will go through this series of images and autistically recreate them without the extra geometry as an exercise in using Crease and Edge Bevel Weight

>>588015
>but edgeflow and etc should be a nobrainer.
The old ways, the traditions of our past, they have been forgotten.

>>588988
Forgot my pic, I am retarded

>>588988
I was thinking you were describing using edge crease on that square hole that you put bevels on to prevent it from collapsing in subdiv.

Nah, that pic's fine and I do that too.

Anyone have tips for speeding up that worked for you?

I don't need things to be perfect, I just need the general shapes with some detail and smooth beveled edges.

I think I'm slowest on making the general shapes, topology aside, and trying to get things to not visually glitch out when I put on a bevel modifier and edge bevel weights since I won't bother with solving ngons if I'm not going to subdiv or Blender figures it out on its own just fine (I check the subdiv'd edges with flat shading and wireframe). Also I'm trying to transition more toward face-weighted normals and simple bevels rather than subdiv when it's not going to be an object you can inspect up close.

Honestly it's not even going to end up in a game as a model, they're getting rendered as sprites.

>>589019
So, context is important for knowing what kind of “hacks” you can take advantage of and in which situations.
The FWN technique is primarily designed for large structures in games such as environments and also machinery or sci-fi, where you can typically get away with tiling textures for the bulk of things and add decals in-engine. Using it for prerendered graphics is a bit of a waste as the extra time you spend isn’t rewarded with better performance.
Within the context of pre-rendered graphics, you can basically throw all sense out the window and use booleans with as many edges as you need for a smooth look, as getting smooth corners is as easy as putting on an edge shader that only kicks in at render time. In this case you can be as wasteful in your design process as you want so long as the final render looks fine.

1/3

>>589915
2/3

>>589916
3/3
it pisses me off when fresh artists cant do a proper mesh for sub-d, they rely on dynameshes and booleans+quadify modifier that produces shitty geo that is almost uneditable.

>>589917
4/3
also, thank you windows 10 for rendering taskbar on the screens.

>>589917

Really nice!

Also, why do you think they would need to edit the mesh after that?, I mean after the boolean and all of that shouldn't the mesh be completely detailed already?

>>589947
with proper modelling for sub-d, its easier to have control over the edges, as well its faster to opimise meshesh if you need more/less triangles after that. You can go back with boolean and move elements around, but then you re-do the quadify/dynamesh thing - not a huge con, but it forces some backtracking. Anyway, im not saying this workflow is entirely bad, its rather tahts it produces dirty results that are less flexible. It is faster for sure tho, especially for fresh people who dont have patience for edgeloops and cut-tool.

>>588928
This please

>>590316
third this

>>588928
>>590316
>>590363
https://mega.nz/#!UwcQGRDI!WDXLTuBKxnbrge7yIxct4pz4UVkZiaQC649l83BEgoU

Here's my current topology reference archive.

>>590623
thanks for uploading
most examples are good, but some are quite wrong though.. (ngons, round openings from smoothed squares, weird use of modifiers, bad spacing creating uneven tension areas, etc.)

>>590628
Yeah, some of it's going to require your own sensibilities since some of them might have been questions about how to fix a model in the threads I scraped.

Complete idiot here, which software do these use? And which one do YOU use?

>>592147

>Which do we?

Anything, it doesn't matter, it can be anything, from blender, c4d to max/maya

>Which i use?

Maya/blender/zbrush/

>>592147
Highpoly/Hard-surface can be done with any program, it’s more of a set of guidelines and techniques, but the program specifics don’t matter as much.
Some programs are better or worse at getting a good result fast, but there isn’t anything you objectively can’t do, as the fundamentals for hard surfaces haven’t changed much in decades.
I use Max and ZBrush.

>>592169
100% true

>>592147
I personally use 3dsmax for modeling and Zbrush for sculpting organic details/variations when necessary. One of my coworkers uses Maya, it's different but it seems great as well. I'd suggest you try out a few and see what seems comfortable to you.

>>589917

How did you even manage to pass your screen through your doorframe?

>>589917
I work in fx animation and don't have any time or energy for personal projects but I've been thinking about getting back to doing something nice on the side. Modeling and texturing "cool" weapons could be one project. Your weapon is inspiring as you are aiming for maximum cleanliness etc. I always used lot of Mirai (back in the day) and then later on Wings 3d for modeling as I think it's most intuitive way to do hard surface modeling especially. With Maya I need to set up my previous workflow and I have lost all my preferences somewhere.

>>587994
>>587995
Jesus. I posted that album in November 2012. It doesn't feel nearly that long ago.
A lot of it's useful, but I could have done a lot of them so much better.

Guess this is as good a place to ask as any.

What would be the best method to round this corner? I can't just bevel it because it goes all crazy. I'm using C4D.

>>592262
Don’t use C4D, but I can imagine a lot of ways to do that. Yolo it and push the edges down manually, maybe use a spline as a visual guide. Boolean it out.
Temporarily remove all edges leaving a quad on top, chamfer the edge, and cut the loops back in.

>>589917
>>589916
>>589915


Gotta ask, why the floaters on the drum mag? The holes would have been easier to do as floaters, and wouldn't have messed with the aliasing the way a clipping floater will. (I'm assuming normal map generation, what you did makes a little bit more sense if it's just a stand alone high-poly, but even then, the 'best' option would have been to do the raised bits AND the holes as mesh)

Great model either way. just wondering if there was a reason for that or not.

>>592262
bevel stops at an edge
so your bevel will stop before it will be fully happen

>>592160
>>592169
>>592173
Alright. Thanks buddies, time to put some effort into this.

how do you plan for something like this?

>>592537
Just bridge either side.

So what's the general take on creasing. Should I be using it or should I be relying on edge loops more?

>>592637
Depends on if you want to export your work. Edge loops are actual geometry whereas the creasing is meta-data that may or may not be supported by another modeller or game engine.

>>592638
It also depends on the subject at hand since you tend to get varying results with one method vs. the other.
To me loops are just a bit more intuitive to work with, because I often preview the subdiv as I model and don’t want to mess with crease settings.

How many levels of subdivision does everyone generally go to in Blender? 2-3 or "as many as possible"? I find that some of my geometry doesn't start looking "crisp" until 5/6 while some works fine with 1/2.

>>595454
2-3 is standard
if you are working on a render, i suggest leaving it at 1-2 you can adjust the render slider to any level you want and its only gonna apply it on rendering

Theres a youtube video im looking for that I now cannot find: Basically it was a video that showed 4(?) different cubes and how smoothing groups and split UVs affected bakes and went through explaining how shading was affected and showed what to do and what not to do with UV unwrapping and baking issues.

Also if there is anything similar that would be nice. Ive always had trouble deciding where to split my UV especially around a chamfer and where to split the smoothing group.

>>592537
you should build it in when you do blockout.

>>595611
>adding chamfers to a blocking
/3/ you wild goose you did it again!

>>595621
what do you consider a blockin? 2 boxes?

>>590623
How do we achieve this so the link doesn't go away?

>>595982
The link should work as long my account isn't disabled, but you could make a Mega account and save it to your account or download it and upload it somewhere else.

>>595454
2-3. If it doesn't look "crisp" until 5-6 then you might have topology problems wherein the topology doesn't actually follow the intended shape.

>>587978
I'm a turbofaget who does most of the hardsurface stuff with booleans.

>>590623
I didn't see many examples of people going out of their way to follow the skeletomuscular contours.

>Presses smooth
>wow I'm such a good 3D artist
kill yourselves

>>597466
why does this board attract so many fucking idiots

>>597492
Part of being an artist is turning off sections of your brain and letting your imagination take over.
Some people turn off too much.

I forgot how slow the /3/ is, i though this topic is long dead.
>>589947
>>Also, why do you think they would need to edit the mesh after that?, I mean after the boolean and all of that shouldn't the mesh be completely detailed already?
Speakign from experience - there are meshes i go back several times during last 4 years fo just one client - sometimes someone want modification, rebake, certain part larger/smaller etc. Also, im old-schooled so i dont trust booleans and do most of things sub-d way, it cleans your mind naad relax.
>>592174
pic related
>>592206
thank you, im glad i've motivated anyone!
>>592375
>>Gotta ask, why the floaters on the drum mag?
Excellent question, i have no idea.

>>598525
>six displays
I aspire to have that many someday, but I'm not sure what I would do with anything past 4.

>>598548
>but I'm not sure what I would do with anything past 4.

i've been using 4screens for last 5 years or so, but i've picked some shell scripting and unix-monitoring gigs and top-left and top-right monitors usually are text terminals or debigging windows, eventually memory/cores stats when i render.

bump

any of this appllies for industrial design product modelling? or better stay wit a cad sotware

>>601678
No doubt, industrial=Just cad.

In regular hard surface modelling you just get cool shapes while in CAD You get realistic hard shapes with real measures

https://youtu.be/b4Tn9tRMeaQ

>>587982
The worst effort I've seen recently

>>592174
You know you can rotate shit around, right? Are you sure you should be doing 3D?
>>598551
3 here, I rarely find myself needing all 3 at the same time, but going back down to 2 just feels wrong somehow. The more you add the more you can't live without...

>>597492
wish i could see that post

>>602233
Obviously he just didn't set up his creases

>>602233
why is this so important again? jeezz you 3d modellers are massive nit picky cunts.

>>601850
How do you repeat that in a radius like that?

>>603353
>You 3d modellers
>/3/
I don't think you belong here.