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


Thread archived.
You cannot reply anymore.



File: blender.png (412 KB, 928x657)
412 KB
412 KB PNG
https://www.youtube.com/watch?v=WhWc3b3KhnY

Newest movie from the blender foundation.

Let's start the shitposting war.
>>
A thread died for this?
>>
>>675323
>GUYS
>GUYS WE CAN DO HAIR AND VOLUMETRICS LOOK
>GUYS LOOK HAIR AND VOLUMETRICS
>GUYS HAIR AND VOLUMETRICS HOLY SHIT
It was shit like all the other ones. Well, maybe except Agent 327.
>>
>>675330
It didn't feel like they improved any of those techs. This is at least the impression I got. Since Blender productions are usually geared toward demonstrating some tech improvement, what are they showcasing here exactly?
>>
>>675332
The technology to awaken giant wood spirits.
>>
>>675333
>underwear
Lads.
>>
>>675332
The technology for getting people to pay up for mediocre animated shorts.
>>
>>675333
porn when?
>>
>>675336
eww, get the FUCK out
>>
>>
That skin shader looks awful t b h
>>
It's nowhere near the Disney quality and there's hardly any improvement compared to Agent 327.
Environment and animation are what stood out the most for me.
I hope they'll put out the next "big" update sooner than in 5 years.
>>
>>675341
327 was Cycles, this is Eevee.
>>
File: 1529689652714.jpg (74 KB, 492x726)
74 KB
74 KB JPG
>>675341
>It's nowhere near the Disney quality
Just curious, did they ever advertise it as it were?
>and there's hardly any improvement compared to Agent 327.
If I'm correct this was the first movie produced with 2.80 starting the movie with the very first versions and it's still not even officially out yet so I guess they had to really fight with the software to get this movie done in the first place.

Currently the new Blender still has to catch up with 2.79 to get everything complete as that old version before taking off.
I think once we reach 2.81 or 2.82 we'll see some serious stuff, especially since they had like 3k and less per month funding and now are close to 30k and rising. Still silly to compare them to Autodesk with its millions of funding power.
Just wish they'll succeed and so far they're doing fine.
>>
>>675341
It looks professional enough for most normalfags, which is what matter.
>>
>>675346
Pretty much.
>>
>>675346
For cartoony style, yeah.
>>
File: a.png (164 KB, 409x325)
164 KB
164 KB PNG
>>675342
Is it really Eevee? It doesn't look like it. Blender Cloud doesn't mention it as well.
>>675343
>did they ever advertise it as it were?
By "Disney quality" I meant high quality and they do advertise it as a "visually stunning" short movie. Meanwhile Spring is far from it. Mediocre assets and poor animation as well as no quality control is why it's average. It's more of a fault of people working on it than the software though.
>starting with the very first versions
"The very first versions" of 2.8 were crashing every 3 minutes and half of the modifiers wasn't working. They must have started with 2.79 and jumped on the 2.8 once it was stable enough for production.
Also it's not the first time Blender Animation Studio was working on a new, unstable version of Blender, other open movies were done in similar matter.
>>675346
Sad, but true.
>>
>>675342
if it was eevee they would have explicitly mentioned it because that would have been really something
>>
File: 1466474807488.png (165 KB, 639x462)
165 KB
165 KB PNG
To be fair it's quite nice, nothing groundbreaking but it's more than passable. I don't really like the main character model, but other than that it's pretty cool. The compositing and creature design especially seem to be the high points overall.

Let's note that the artists who worked on it are quite kickass too. Julien Kaspar is a personal favourite.

>inb4 this short gets mentioned all the fucking time by blendlets now as the best thing since sliced bread..

>inb4 hurr durr RIP Pixar RIP autodesk
>>
>>675364
>nothing groundbreaking
That's the issue here, isn't it? All these short films made by Blender themselves have always been to showcase what Blender can really do with a (semi-)professional production team, in other words showing what Blender is at its best. If this is what Blender is at its best, then what is the Foundation trying to say?
>>
File: Capture4.jpg (23 KB, 596x608)
23 KB
23 KB JPG
>>675364
kek I just noticed, these eyelashes and eyebrows are hilarious

the eyes are weird also, they're quite flat both in shape (they don't even seem to be round?) and shading (quite diffuse and lifeless).. that's probably the main reason why the character doesn't sit well with me..

so yeah, the creatures and environment seem a lot more successful, quality-wise
>>
>>675365
Meh, it's good enough for plenty of things I suppose. Nobody ever mentions it in a professional context of course, but I guess it can be useful to learn the basics of 3D or something. For modeling maybe? I can't really say.. I just wouldn't expect Blender to compete with the professional tools that have been developed and pushed to their limit for the past 20 years anyway. At best a startup with a few skilled people could use it to create short TV ads or whatnot, the bar is usually lower for those.

As far as I'm concerned, since my main thing is rendering, there's no way in hell I will ever consider Blender for anything I do.
>>
>>675360
>eevee
isn't this their real-time thing? if so, that's definitely not what was used here
>>
>>675366
They had the same problem in the Laundromat short. I don't think it's because the hair system isn't up to the task but rather the artists having no idea how to make attach it to a human character without it looking like glued on pubes.
>>
>>675337
preach bruvkuh.
>>
File: 1503442890238.jpg (2 KB, 119x125)
2 KB
2 KB JPG
>>675371
>glued on pubes
cannot unsee
>>
File: victorhairsim-1200x800.png (1.1 MB, 1200x800)
1.1 MB
1.1 MB PNG
>>675373
gaze into the abyss
>>
>>675374
the main problem is that you still have to find a compromise and cannot use as many hair as there are usually on a human head. hence animators make them thicker and less of them which turns into the usual pubes look.
>>
>>675376
The main issue is that they're relying on the particle system to do the entire job of making hair look good without painting smaller hair/stubble detail where it penetrates into the skin on the albedo.
>>
>LOL IT"S NOT PIXAR
Imagine getting butthurt at a free software foundation producing an entertaining proof of concept as a way to test, debug and improve the free software they're producing
>>
>>675366
>they don't even seem to be round?
If you try to make her eyes perfect spheres, they'd clip out the side of her head because they'd have to be so huge. Poor character design when accounting for 3D honestly.

>>675371
>>675376
>>675377
Looks like Blender artists just don't know how eyebrow hairs work. You can achieve passable eyebrows by indicating some sort of flow, you don't even have to make them as dense as real eyebrows. Most games still bake eyebrows into face textures and those look fine despite being constrained to such a small space on a texture map. I wish there were closeups so I could say for sure, but it just looks like the artists here just clumped as many small hairs as they could resulting in those pube-like caterpillars.
>>
>>675342
>Eevee
No it isn't.
>>
File: 1508636961157.jpg (153 KB, 992x1080)
153 KB
153 KB JPG
>>675410
>entertaining proof of concept as a way to test, debug and improve the software
While they improved the software itself, nothing got improved in terms of the quality of short films.
And if they want to get more people to use their software, they should try to show how good it can actually be. But when they show clipping animation, lifeless skin shader and poor environments, you end up with 90% of your userbase being Blendies who will eat up anything you throw at them, since no professional is going to take that software seriously.

>>675414
>eyes would clip out the side of her head because they'd have to be so huge
>he doesn't know about the lattice modifier
Also if you wanted to make the eyes more round, you would have to make the initial sphere smaller, not larger. Poor common knowledge when accounting for 3D honestly.
>>
>>675323
I dont get it,
Does she do this every year? If so, does she not age and nether does the dog?
Why does she seem so surprised if its something that is done every year?
If others do it too why is she alone?

I mean making cool shorts is one thing, but i like my shorts to make a little bit of sense.
>>
>>675423
I guess it's some sort of ritual that's being done every year, by her people. And now it's her duty to do this.
As for why she goes there alone? idk but the answer is probably "muh tradition of doing it on your own" or something like that.
>>
>they should try to show how good it can actually be
>no professional is going to take that software seriously
it's all about established pipelines and workflows (which funnily you don't see), and not final renders (which you do see)

you can easily change the look of of the film by swapping the renderer and artists. but you will probably never change industry standards though. not much reason too, anyhow.
>>
File: 1516720025815.png (113 KB, 234x205)
113 KB
113 KB PNG
>>675426
>implying pipelines, workflows and industry standards don't change over time
shiggy diggy
>>
>>675426
>you will probably never change industry standards though
Industry standards change. Old software gets replaced by new, better software, once the benefits are clear enough and make the money talk.
>>
File: BRAVE.jpg (879 KB, 1600x1200)
879 KB
879 KB JPG
>>>675323
ECH HEM
>2012 was 7 years ago
>the absolute state of blend cultists
>>
>>675437
Are you seriously putting this at the level of Brave?
>>
>>675437
>>675438
Btw, that's the perfect excuse for a tech post.
https://www.youtube.com/watch?v=YeYW8TIWLG8
>>
File: blender game.jpg (473 KB, 1350x780)
473 KB
473 KB JPG
>>675438
>>675438
>don't compare it to old movies
>its not fair we need more time to develop it
>7 years isn't enough time
>in 70 years it will be great just wait
>when you think about it, its actually better than those old movies because its open source
>those movies cost more money than ours so it doesn't count
>we broke our everything with our update so its understandable
>yeah blender sucks therefore you should donate to the blender foundation to make it better
>>
>>675431
>implying pipelines, workflows and industry standards don't change over time
they don't really
>get some verts
>move them between frames somehow
>shade the faces
>composite the image

>>675433
>make the money talk
well the linux ecosystem is cost free, but businesses have an existing infrastructure and culture that will be difficult to switch from
>>
>>675446
>2090
>Blender 2.8 hasn't come out of beta yet
>the comments section for the release video of Maya 2091 has hundreds of Blender users shitposting about how Autodesk is going to die
>someone brings up how you still can't do precision modeling or group objects together in Blender
>"w-well there might be an addon for that..."
>>
File: 1549151115707.jpg (918 KB, 1440x1080)
918 KB
918 KB JPG
>>675449
Since it's so hard for you to understand, let me give you some rough examples.
Before 2014 texturing was done among others inside 3D-Coat and Mari. Then Substance Painter gets released and within 5 years the entire video games industry uses it to texture most of the assets, due to smart materials and automation it provides.
Marvelous Designer, first released in 2011. Before, people would mainly sculpt wrinkles from the scratch inside Zbrush. These days, the workflow is "Simulate wrinkles in MD" > "Refine in Zbrush".
>>
>>675456
Hi, those are caused by major technological advances that allow artists to work more quickly and effectively. Note that at least MD moves things in the direction of paying more money for more pro software, and anyone who moves from Photoshop to Substance is also adding additional $$$ that they didn't used to. So clearly what's happening is that artists and studios are happy to pay and integrate the best, most effective new tools into their pipelines, while cheap/free alternatives that are worse are going nowhere.
>>
>>675449
>well the linux ecosystem is cost free, but businesses have an existing infrastructure and culture that will be difficult to switch from
What point are you trying to make? First of all, Linux-based infrastructure is not cost-free: you still have to pay for all the specialists needed to operate it. And, secondly, Linux is the undisputed tech leader in many IT areas, so there's no reason for businesses to switch away from it.
>>
>>675410
fair enough
>>
File: 1518826971886.jpg (17 KB, 319x268)
17 KB
17 KB JPG
>>675414
>eyes would clip out the side of her head
dude no...
>>
File: 1417969401300.jpg (14 KB, 260x232)
14 KB
14 KB JPG
>>675446
what the fuck is "yo frankie" and who the fuck would produce such nightmare fuel
>>
>>675449
>implying pipelines, workflows and industry standards don't change over time
>they don't really
holy fuck, stop scooping in the chromosomes bucket m8
>>
>>675376
>cannot use as many hair as there are usually on a human head
Of course we can. What year do you hail from, 2009?
>>
>>675491
Yeah have fun rendering 5 hours per frame when your animation hundreds of thousands of frames
>>
File: 1486771071100.gif (436 KB, 200x150)
436 KB
436 KB GIF
>>675492
>SSS
>displacements
>volume fog
>entire clouds of particles
>complete terrain with pebbles, grass, and bushes
>OH NO, PROPER HAIR WOULD BLOW UP RENDER TIME
>mfw
>>
Anyone thought the rocks looked really bad? Like ripped out of someones Warhammer table. Looked more like painted cardboard than actual stone.
>>
>>675360
Okay, my mistake. Definitely less impressive but I still enjoyed it.
>>
>>675456
>>675483
>we use photoshop now instead of mspaint
woah calm down there with the paradigm shifts lol
>>
>assets available
oh shit, any blender loli lewds yet?
Will I have to make them myself?
>>
>>675482
It was Ton's attempt at making the Blender game engine worth something and failed spectacularly. It was a tie in with that rabbit movie and not only did it make the game engine even more of a joke, they planned for a standalone version by trying to tie Blender with the Crystal Space engine which was already on its last legs years before the deal. The CS guys took the money and ran, the project got canned, a single buggy level for the BE version got released, pushed under the carpet and no one talked about it ever again.
>>
>>675524
dear god, that's a hilarious drama
>>
>>675524
>>675531
that's actually pretty sad
>>
>>675342
>327 was Cycles, this is Eevee.
>>675355
>Is it really Eevee? It doesn't look like it. Blender Cloud doesn't mention it as well.
it was Cycles

>>675332
The big improvement was workflow. Making the movies helps them figure out what workflow changes would be an improvement, since a lot of developers will just tack on features with no concept of how they will be used.

On the cycles side not much has changed. Mostly benefits for speed, GPU rendering is a lot faster and you can now mix CPU and GPU rendering.

There's also post processing improvements to help filtering. You can also now add more samples to an image after it's already been rendered. It's console atm but it's a neat feature.
>>
>>675456
Mari still shits all over SP when you paint assets for VFX
>>
>>675546
>You can also now add more samples to an image after it's already been rendered.
Aren't samples just averaged into the values of their origin pixel? Couldn't the same effect be achieved by rendering the added samples and then blending those samples with the original image? You'd just need to bias the blend towards the higher sample image, no? (i.e. if the original had 3000 samples and the added image has 1000 samples then the original should make up 75% of the final image)

Is my understand of how samples work fundamentally flawed?
>>
>>675549
>Is my understand of how samples work fundamentally flawed?
No, you described the process they use, the bias is encoded in the meta data of the images, recording how many samples for each image.
>>
>>675549
Your understanding is correct. As long as the new render uses a different seed than the previous one, that'll work.
>>
>>675547
Yes, that's why I said "video game industry", and not "video games and VFX industry".
>>
I liked it, nothing really jumped out to me in terms of bad quality. The volumetrics, especially the clouds at the end could have benefited from more resolution but I guess they didn't have years of computing time to render it.

>>675422
>And if they want to get more people to use their software, they should try to show how good it can actually be. But when they show clipping animation, lifeless skin shader and poor environments, you end up with 90% of your userbase being Blendies who will eat up anything you throw at them, since no professional is going to take that software seriously.
I'm sure they'd be able to do much better if they had the budget to hire more artists but look at the credits, they're a tiny team. And the software is free.

>>675368
Pros love free software, look at the IT industry for example. Being able to modify existing software and getting updates from the community for free is a huge advantage for professionals. And it looks like we're about to reach a tipping point in that regard with blender. The best example would be Tangent Animation, they ditched maya and spent the money saved on licensing fees to hire blender developers to work on the features they needed to produce Next Gen, and now some of these features are merged back into blender's trunk. If I were a studio owner I'd start paying attention to these developments since the potential benefits are quite huge.
>>
>>675564
>I guess they didn't have years of computing time to render it
I would love to see them make their next short entirely in Eevee, just to force them to fix the shadows.
>>
>>675584
>just to force them to fix the shadows
There are inherent limitations to shadows using rasterizer. Or is there something particular not working?
>>
>>675564
I get what you're saying, but loving something free doesn't make it better, nor the ability to develop tools for it *if* it has issues at it's core.

To be fair, movie studios that use Maya also develop their own tools to transform it into their own custom 3D package. 3dsmax and C4D probably aren't as flexible in that regard, but Blender isn't unique, though I'll admit it is an advantage.

I still stand by my previous statement that for rendering in general (animation, archviz, etc.), Blender is one of the worst choices out there, though I'm not gonna be a total ass and claim it's not good for anything.
>>
>>675592
They work, they're just rather bad in terms of visual quality and bias issues. While they'll never be as good as a ray traced version, there are better methods of doing realtime shadows than what they're using now.
>>
>>675564
>>675596
I'm pretty sure the ability to develop plugins or scripts using the API is sufficient for small studios who need a couple of in-house tools. I'm not saying it doesn't have its uses, but "muh open source" is often a meme that gets parroted by people who have never worked in a studio or tried to develop a feature for a project with the size of Blender's codebase. Tangent for instance didn't just switch to Blender because they wanted to add something to Maya but couldn't, they did it because they wanted to use Blender to make a statement and effectively got employed as developers to fix some of the shit wrong with it that nobody bothered with before. Next Gen being made with Blender was good for BLENDER, not the other way around.

If you need full source code access with everything autistically fine tuned to the exact tech you just invented last week based on your own whitepaper, and the API/scripting system of the tool you're using is too limited for it, you might as well write your own DCC tool/physics engine/renderer at that point because the effort involved is going to be on the same order of magnitude, and you get the added benefit of having full control over the codebase instead of having to spend time maintaining it and making sure it's compatible with the changes in Blender 2.81. Or heaven forbid, release it as a standalone project like OpenVDB so it can get adopted by the industry as a whole instead of being a Blender-only feature.
>>
>>675493
kek underrated post
>>
>>675599
>they did it because they wanted to use Blender to make a statement and effectively got employed as developers to fix some of the shit wrong with it that nobody bothered with before. Next Gen being made with Blender was good for BLENDER, not the other way around.
I'm not entirely sure about that, I think that it's the right strategy to adopt for long-term gains. I'm probably a naive and overly optimistic free software enthusiast but since blender is the only viable FOSS and "complete" 3D package I do have high hopes for its future. Imagine if another "big" studio decides to adopt blender for their next project and adds or optimize some features, from the POV of Tangent and everyone else using blender it's a net advantage they'll be able to profit from and for free.
>>
>>675597
I'm not actually aware of anything significantly better than VSM. Shadow Volumes would be completely impractical. I know theres been some work on the SVO front, but is there any other more practical shadow technique?
>>
>>675603
And I'll fully admit I'm a jaded piece of shit, but I've seen too many FOSS projects over the years get praised to the high heavens because of "the community" finally taking on the big bad Adobe/Autodesk/Microsoft to then get absolutely sidelined by a proprietary competitor made by two guys from scratch over the course of a few months.

On the other hand though I won't dispute that more studios using Blender for large scale projects would be a benefit, not only for performance/workflow fixes and bringing Blender up to par with other tools, but actual innovations which could become exclusive to it and drive competition. Something like Eevee is definitely a step in the right direction. Or studios like Pixar/Sony/Dreamworks open sourcing their tools, which is starting to happen increasingly more frequently.
>>
>>675564
>>675596
> Thinking Blender is more customizable than Maya

"Muh free software" motherfuckers like you have clearly never tried to customize either of these programs before. Blender doesn't even OFFER a binary API. There's no way to create a .dll plugin for Blender. More importantly, there is no way to sell a binary plugin without being forced to release the source code. This is why render engines like Renderman for Blender will always be inferior, you will never have muscle plugins like Ziva, and you will never have crowd plugins like Golaem.
>>675599
If you want to write your own software with custom everything, you write it on top of Maya, Max or Houdini. There's no reason to reinvent the wheel with basic things like matrix transforms and parallel processing. Any custom functionality you want you can write yourself and plug into your DCC, unless you're using Blender, in which case you're limited to shitty cheap Python modules, and you can't even hook anything into the user interface since it's not based in Qt.


Blendlets continue to yammer on endlessly about "modifying the source code," meanwhile no professional CG developer in their right minds would ever want to touch that steaming pile of shit. With Maya the process of starting a new custom C++ plugin using their custom Visual Studio toolkit takes literally 2 minutes.

https://www.autodesk.com/developer-network/platform-technologies/maya
>>
>>675607
Let me summarize this. The SDKs for Maya, Max and Houdini are top notch, pleasant to use and deeply integrated in all major studio pipelines.

There is *no such thing* as a Blender SDK. It doesn't even exist.
>>
>>675607
>>675608
holy based and redpilled. blendlets BTFO
>>
File: prestomcqueen.jpg (236 KB, 1200x766)
236 KB
236 KB JPG
>>675607
>There's no reason to reinvent the wheel with basic things like matrix transforms and parallel processing
There is if you're Pixar.
>>
File: siiu.png (104 KB, 247x272)
104 KB
104 KB PNG
>>675337
do you know where we are
>>
>>675616
God tier: Write your own proprietary DCC from scratch (Pixar, Weta)
Patrician tier: Use commercial DCCs, customize heavily (SPI, Dreamworks, gamedev studios)
Pleb tier: Use commercial DCCs off-the-shelf, don't do much customization (Indie gamedev, freelancers)
Meme tier: Use meme software, fix a few of its nasty bugs on your own dime, make meme CG, say you "contributed"
>>
>>675624
Dreamworks have their own custom proprietary thing as well, although I think it's mostly for historical reasons
https://www.youtube.com/watch?v=N16Qsj6kAhw

However Disney does just use a custom build of Maya (with their own renderer)
https://www.youtube.com/watch?v=0Z-0EskUcDs
>>
File: 1545596346015.jpg (41 KB, 641x530)
41 KB
41 KB JPG
>>675625
I'd even take the default Maya + Hyperion any fucking day..
Though custom Maya + Renderman is top tier af
>>
>>675625
That doesn't look like a "custom build" of Maya, just Maya with custom additions.
>>
>>675607
>There's no way to create a .dll plugin for Blender.
That's simply misinformation.

>More importantly, there is no way to sell a binary plugin without being forced to release the source code.
This is also wrong. If you don't import GPL, you don't actually have to publish your source code. Everything in a Blender addon is GPL, but that is only the machine code part of what you release.

Why no big paid addons? Simply because GPL means people can just copy & paste without any reprecussion. It's currently all based on a honor system / Patreon model.
>>
>>675715
>Why no big paid addons?
Pretty sure Flip Fluids is distributed as a standalone binary and communicates with a blender add-on. In this case their binary doesn't need to be GPL.
>>
>>675713
The version string literally has "Disney" in it. I wouldn't be surprised if they have their own branch with Autodesk developers on site. They can afford to buy out the company outright if they wanted to.
>>
I liked the short. It looked preety good and it is amazing these movies are made with free software.
>>
>>675333
hnnnngggggh
fucking cute
fucking funny
>>
File: 1541241892372.jpg (595 KB, 2836x1395)
595 KB
595 KB JPG
>>675334
https://www.davidrevoy.com/article712/spring-open-movie-character-design-and-concept-art
>>
>>675718
that's the point though - its not amazing, its amazing for free software
>>
>>675834
>A first scene before an old Spring goes to sleep (before getting reborn).

It's fine guys, she's actually 1000 years old.
>>
>>675839
OLD ENOUGH TO BREED
>>
>>675839
No, you got it backwards. She's always a few months old at most.
>>
>>675346
I mean if Illumination can pull the stunt, why not the Blender Foundation?

Come to think of it, I'm kind of surprised Illumination isn't connected to BF.
>>
>>675376
Artists using real-time render engines in the games industry have been making good hair since the PS2 days, hair that looks better than this Blender effort.



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.