Okay who wants to try making these in blender

General discussion about the development of the open source Blender

Moderators: jesterKing, stiv

Toon_Scheur
Posts: 0
Joined: Sat Nov 06, 2004 6:20 pm

Post by Toon_Scheur »

I like those discussions about the capabilities of Blender in comparison to other packages.
I think that those guys like ChojinDSL could better voice a constructive opinion about this topic because they actually used other packages.
What's really funny is that no matter what, Blender will always be bashed. Blender has now AO, it almost has HDRI, it has cool edit modes/ tools that's up to par with other packages, and new capabilities will be added constantly. So in the light of that, why do we still think/ discuss that Blender is out of the league of the big boys? Even when Blender gets all the cool stuff like SSS and so one and so one, you guys are still gonna ask the question whether Blender is capable to the things that Maya, Max, XSI can do.

My answer is: Blender is capable to do all the stuff you see in most CG movies right out of the box. If you make Hollywood work only with Blender, I'm very confident that you won't see any difference. I've seen lots of tutorials and articles about photorealistic stuff that was done in Maya, or Lightwave or Max. But as I read the articles, I see that everything they described could be done with Blender since version 2.0 . We all know what Blender lacks. And I'm sure the Blender dev team is always working on it to improve Blender. Every new version was a giant leap forward in all aspects of the design of Blender, and version 2.36 will keep us amazed of how fast things can go. Most of us can hardly keep up with the pace.
Personally, I think Blender can be improved upon in the rendering departing, as for special features, everything you see in those Siggraph paper should have been implemented yesterday! Blender really needs voxel/ grid rendering to render smoke, water and such. Clothing, hair modeling and hair animation. I am a programmer too, but I don't have so much time to work on Blender. I stopped about a year ago working on SSS, that is the method of Henrik Jensen/ Juan Buhler, but I would like to pick it up again. I would like to work on mesh optimisation, smoke, fire, water, clothing and so on. That is the great weakness of Blender. With Blender would be very hard (if not impossible ) to make animation with clothed, hairy humans/ creatures. But all things beeing equal... Maya can do, Blender can do, Max can do , Lightwave can do , ARTIST MUST DO.

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Post by Money_YaY! »

and yet no one has taken this posts challenge :? Only said that it could and explained 'why' it could but no proof yet.

paradox
Posts: 2
Joined: Fri Oct 18, 2002 7:07 am
Contact:

Post by paradox »

Perhaps they are working on their own projects, instead of trying to copy something someone did in another software version just to prove it can be done.
Paradox

Samjh
Posts: 0
Joined: Fri May 14, 2004 3:28 pm
Location: Australia

Post by Samjh »

My answer is: Blender is capable to do all the stuff you see in most CG movies right out of the box. If you make Hollywood work only with Blender, I'm very confident that you won't see any difference
Hollywood work only with Blender? Impossible at this stage. Even with high-end commercial packages, studios use multiple packages for various jobs.

Also, the crux of CG film visual effects is not the number of features, but the quality of renders for animation. Blender simply doesn't have a good enough renderer for photorealistic film work. In fact, there are few renderers that are suitable for such purpose -- PRMan, Mental Ray, Brasil, lead the way for film applications, and they are all external renderers with tight integration with their host applications (usually Maya, XSI, and Max). Lightwave is the only package with an internal renderer that has a proven history of producing acceptable film visual effects. Cinema 4D has a great internal renderer which has produced good results for television broadcasting, but not yet for films. Blender is starting to approach some of those packages in terms of general feature lists, but its output quality is far from competitive.

Blender's internal renderer has too many feature limitations, and is too slow for rendering transparent objects or highly complex scenes with realistic shadows and transparent effects. Yafray integration is simply not good enough (although it has made good inroads so far), and Yafray itself doesn't support enough of Blender's more advanced effects. In short, Blender has no chance of directly producing any feature film visual effects which is worthy of viewing by the paying audience. Free Renderman-compliant renderers (eg. Aqsis) are Blender's only chance so far, but none of them are proven solutions, and integration is still at infant stages.

The sole purpose of CG film visual effects is to produce good animations in the quickest amount of time. Blender doesn't satisfy either requirement -- quality or speed. At the moment, Blender does well enough for industrial visualisations, product packaging, print illustrations, and some limited television broadcast animations, but not good enough or fast enough for feature films.

Blender also needs to have wider and better professional support to be able to compete against the big boys. Companies like Discreet and Alias offer specialised training courses, and all kinds of support deals with their customers. Blender has nothing but a small bunch of forums and mostly out-of-date python scripts (no offense to scripters.... just a comparison with the commercial market). If Blender wants to break into the world of big-name studios, it needs to be backed up by a solid after-sales support plan.

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Post by Money_YaY! »

Simple answer, Blender needs full time paid developers inaddition to the CVS helpers. There is just no way to sway someone working for free to make what you want, they just make what they want to tinker with. If Ton will not pay someone then it will just stay as a neat little app with features that come and go.

Now if there were say two real qualified devs takeing care of all of the little drum bits every one has been begging for there would be far more free space for free CVS developers to play in as they would not be getting the constant 'make this make this' banter.

eh :?

bfvietnam
Posts: 0
Joined: Wed Apr 21, 2004 8:54 pm

Post by bfvietnam »

Samjh wrote:
My answer is: Blender is capable to do all the stuff you see in most CG movies right out of the box. If you make Hollywood work only with Blender, I'm very confident that you won't see any difference
Hollywood work only with Blender? Impossible at this stage. Even with high-end commercial packages, studios use multiple packages for various jobs.

Also, the crux of CG film visual effects is not the number of features, but the quality of renders for animation. Blender simply doesn't have a good enough renderer for photorealistic film work. In fact, there are few
Renderman doesn't either, name what defines photorealistic
in terms of Renderman, and compare that to blender.. Tell me what
in Toy Story couldn't have been done with blender.

Name in "The Incredibles" can't be done with blender.

And after you name those things, tell me what percentage of the
film used these effects?
renderers that are suitable for such purpose -- PRMan, Mental Ray, Brasil, lead the way for film applications, and they are all external renderers with tight integration with their host applications (usually Maya, XSI, and Max). Lightwave is the only package with an internal renderer that has a proven history of producing acceptable film visual effects. Cinema 4D has a great internal renderer which has produced good results for television broadcasting, but not yet for films. Blender is starting to approach some of those packages in terms of general feature lists, but its output quality is far from competitive.

Blender's internal renderer has too many feature limitations, and is too slow for rendering transparent objects or highly complex scenes with realistic shadows and transparent effects. Yafray integration is simply not good enough (although it has made good inroads so far), and Yafray itself doesn't support enough of Blender's more advanced effects. In short, Blender has no chance of directly producing any feature film visual effects which is worthy of viewing by the paying audience. Free Renderman-compliant renderers (eg. Aqsis) are Blender's only chance so far, but none of them are proven solutions, and integration is still at infant stages.

The sole purpose of CG film visual effects is to produce good animations in the quickest amount of time.
Neither is renderman, have you ever tried to render anything with
renderman in a suitable amount of time. I think how blender could benefit
is by the speed of experimentation, like if it supported a recursive rendering mode, apart form the scanline mode. Artists need quick feedback, if you need to render an animation out fast, and your animation is done, just throw more CPU's at it. But when making movies its the
speed of creation and experimentation, not the speed at which you can render out your animations on the final go.

I think given the longevity of blender, considering the development paradigm, one way or another blender will be a tool commonly used, as its the only 3D tool that you can rely on to exist regardless of the success of the industry. Its perpetual existence will drive innovation using blender a megaphone for consumer preference. That's how the 3D package developers will, and are, seeing blender's contribution. Its visible/useable research into the alternative methods of doing 3D work. And it can only become more refined, flexible and extensible. The more extensible, the more companies will latch onto it using it for the creation of in-house creation frameworks.

What software businesses strive for is the same kinds of things Microsoft strives for.. Reliance on products for infrastructure. Blender is an infrastructure that can't be owned, and it can't be leveraged.. That's the
threat of developing blender.. To some its more a threat than others.. It depends on how you see your your method of making a living. If you are to maintain and develop turnkey systems based on blender, or if you are developing plugins that work in blender, or if you are to compete with blender with your own "proprietary" infrastructure. Lightwave, 3DsMax, Maya, Softimage are just seperate infrastructures that work together because they must, not because they choose to. Their consumers demand it. If their consumers use blender, they will want to learn how to plug into blender as well..

I believe since blender is open source, it will allow for tight coupling with
commercial packages and tight customization and workflows for studios,
as well as for visualization. Its bound to happen, considering how easy it is
to download and execute..

All that stands in its way now is its ease of compiling, changing and extending. Beyond that you can see the the kind of extension that
you would find on 3DsMax, Lightwave, Maya or Softimage.

What that could do potentially, is force those vendors to try to determine
their delta of value add (what features are valueable and sellable, what features are common) to the world of 3D.. Because blender could potentially (and must become) the infrastructure for 3D development, simply because it is open sourced and free.

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Post by Money_YaY! »

Realtime GPU Blender. That would leed to super speed and test renders as far as the eye can see. IF not then blender is in the stone age still just trying to 'catch up' instead of innovate.

Toon_Scheur
Posts: 0
Joined: Sat Nov 06, 2004 6:20 pm

Post by Toon_Scheur »

I think Money_YAY is right. The challenge was toproduce the same quality head model as in the examples mentioned before. So far there has been a lot of talk. But its time to walk the walk. I will give it a go. I might take some weeks. But suppose I (or someone else) pull this of, does this means that this discussion is over? I don't think so. It only proves that Blender is a capable software. But you can't make everybody happy. Even the animators at Pixar (as I read from interviews) say that they miss some functionalities from other packages they used to use, but the current package has other things that their old package didn't have. To reply about the quality of Blender's render engine: It isn't the best render engine, but it is not bad. But if you know your stuff, like how to light up your scene correctly, you can achieve amazing result in a very short time.
Someone said something about that for a painter there is a clear difference between cheap paint and expensive paint. Well, I don't think that Da Vinci made his sketches with a $50 fountain pen with inbuilt fm-radio. I think he used just a piece of charcoal or something. The difference between cheap and expensive paint= marketing

If Blender wasn't free, say it costs $3000. And lets assume that Maya was an open source package, the guys at free-maya.org would be argueing if Maya would ever be as good as Blender.
If we look at the development story, Blender should be promoted as a development platform for all the inovators who are researching CG. If you take a look at the Siggraph paper, I get a feeling that those guys develop their code from scratch. But if they can use Blender to develop code in to test their theories, they have the added value of a shorter coding cycle and a program to test their work on a broader base.

bertram
Posts: 0
Joined: Wed Oct 16, 2002 12:03 am

Post by bertram »

Samjh wrote:Blender also needs to have wider and better professional support to be able to compete against the big boys.
[...]
If Blender wants to break into the world of big-name studios, it needs to be backed up by a solid after-sales support plan.
Agreed, but who should want blender to compete against the big players? What would be the advantage of this? There is no need for a crusade like the one of the Linux-devotees against Microsoft Windows.
Money_YaY! wrote:Now if there were say two real qualified devs takeing care of all of the little drum bits every one has been begging for there would be far more free space for free CVS developers to play in as they would not be getting the constant 'make this make this' banter.
Yes! Great idea, man! I'v also got one. Like to hear it?
How about just taking over discreet and integrating their whole range of products into blender? Clever, ain't it? :wink:
But, wait,... this would raise two problems:
1. Got enough loose money?
2. Your "make this make that banter" won't stop!
Money_YaY wrote:IF not then blender is in the stone age still just trying to 'catch up' instead of innovate.
1. Yes, this commonly is the deal with something that you get for free: It may be not the best of it's class (which is certainly not mandatory). Nevertheless it will surely offer the best cost-performance ratio (when looking to purchasing only).
2. "Catching up" and "Innovating" does not necessarily need to be mutually exclusive. The base of blender functionality is so broad that you would always find areas where blender has to catch up whereas there certainly is very much innovation in other areas.

I more and more wonder if all the posts about blender extreeeeemely needing this feature and being aaaaaaaawfully bothering weak when dealing with that task are initiated by the folks at Alias, Discreet, Avid, etc. to empoison this great blender community.

joeri
Posts: 96
Joined: Fri Jan 10, 2003 6:41 pm
Contact:

Post by joeri »

Oranges and apples.

Will blender rule in hollywood? Who cares?

What part of the incredibles could not have been done in blender?
maybe the render quality of every single frame can be made with blender.
(That's where the post started, a single image).
But that's not very interesting is it? Then we could say "can it be done in photoshop, tell me, what part could not have been done in photoshop?"
Useless question, All single images could have been done in photoshop.

But all of it, the whole production...?
The modelers, shaders, animators, lighters, the programmers working together with a director, film studio, in a package with no undo?

The hair, dynamics, softbody collision, elastigirl-IK/Fk switching, theese all need to work so animators can do their job: putting emotion in the acting, not "getting the software to do the darn thing with workaround 12 and rt: on 6"

I think Toon_Scheur is wrong:
I'm not bashing blender because it's opensource. It's just not as good as XSI / Maya. So what? Stop pretending it is, it's of no importance.
And no blender does not have "edit modes/ tools that's up to par with other packages" (Ever tried a trimmed surface?).
And for animation the whole edit stuff is unimportant anyway.

"Blender is capable to do all the stuff you see in most CG movies right out of the box. "
No it's not. Go and take a better look, or better watch a making of.
Pay interrest when you see:
- The hair on the blue monster in mosters, inc.
- The cloth in the incredebles.
If you want to animate thoose "by hand", good luck, hope you finish your movie this century.

That's not what this post was about? was it?
Why would you go for an opensourse, 6 developers, dictator ruled ( ;) ) software package when you have $480.000.000 dollar to spare?

Can you make an image of a hot chick in blender?
Ofcourse you can, put time and effort in it, and she'll be marvoulos.

I wonder how long it will take for the pile of oranges and the pile of apples to mix and mingle again.

joeri
Posts: 96
Joined: Fri Jan 10, 2003 6:41 pm
Contact:

Post by joeri »

bfvietnam wrote:
joeri wrote:
Much later the game engine (the one in 2.04) was added. That didn't have any rendering, only texture mapping. The rendering was done (realtime) on the playstation. On the sgi games where all wireframe.


Couldn't afford and Indy?

I thought Ton had a challenge?
In europe they where extra expensive, so no *sniff* indy or challenge.
That's why we loved the playstation. 500 bucks for realtime textured 3d.
Ofcourse the playstation developement kit was eating 20k out of our investment budget.

I wonder why Ton didn't make a playstation port of blender. The 1mb texture memory could have something todo with that.

alt

Post by alt »

The difference between cheap and expensive paint= marketing
Actually, difference = amount and origin of pigments (some natural hues are really expensive) and additional ingredients. These influence both the look of painting and how it stands time.

And traditional art is something where marketing does not apply much ;)

edit: even more tiresome comments removed.
Last edited by alt on Sat Dec 04, 2004 1:52 am, edited 1 time in total.

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Post by Money_YaY! »

oh blah blah blah, just use it and prove everyone wrong otherwise blah it is an app with a few developers what ever they code or make it is just code. Blender is a fine app to use, but there are to many politics to getting it to add a new whatever.


egh

bfvietnam
Posts: 0
Joined: Wed Apr 21, 2004 8:54 pm

Post by bfvietnam »

joeri wrote:Oranges and apples.

Will blender rule in hollywood? Who cares?
I don't guess it will because inhouse studios tend to focus on
keeping their system the best and most advanaced, which means they
adopt anything and everythiong that will get them ahead.. But given that
blender is already built and the source is free, its a step forward. If you look at the software Pixar used, as well as other houses, before they adopted packages like Maya, like before Toy Story was made (* hint: take a look at the Toy Story book about the production, there are photos of some of their UI, it looks like a OpenInventor hack *). Anyhow, if some studio wants to design an environment for doing animation work and they
don't want to purchase into Maya or use Softimage, they could easily get blender and roll their own, even writing or adopting plugins for blender to
integrate with their other packages. In the future may buy Maya and replace blender, but at least it would reduce the initial budget.

The advantge I think blender has over other packages is its easy to use, download, open sourced, and is file-system-less (everything goes into single files). But if you want to play the low-profile non-compete (avoid patent suits in europe) card, I could vouch for how ill-equipped and useless blender is. :)

PS- Its games night, so I can't reply to the entire post.. Gotta play BFV, yee haw!!

bfvietnam
Posts: 0
Joined: Wed Apr 21, 2004 8:54 pm

Post by bfvietnam »

joeri wrote:Oranges and apples.

What part of the incredibles could not have been done in blender?
maybe the render quality of every single frame can be made with blender.
(That's where the post started, a single image).
But that's not very interesting is it? Then we could say "can it be done in photoshop, tell me, what part could not have been done in photoshop?"
Useless question, All single images could have been done in photoshop.
You could do much of the effects with blender... I didn't see very many examples of things that couldn't be done. Most of it relies on good animators. Before SDS, it would have almost required NURBS and parametric surfaces, like fillet surfaces.. After subdivision surfaces,
it pretty much levelled the playing field.. But of course Lightwave had
it all along. "The Incredibles" wowed me, but I think good motion and
animation work is worth more than the quality of the tools, but it helps to have good tools.
But all of it, the whole production...?
The modelers, shaders, animators, lighters, the programmers working together with a director, film studio, in a package with no undo?
Wasn't that the feature that was added to 2.34.. Its not the detail oriented hiearchical object oriented undo that Maya users would expect, but its
simplistic and can be reapplied.. Furthermore, since blender's new undo feature just uses memory deltas, those deltas can be stored, making it the first 3D application with the ability to store its undo history (making it unique from all other packages).. Maya can make mel scripts, but the integration is more loosely tied than blender's because there needs to be a mel command for every action.. Whereas in blender, if anything changes, its stored and referenced.. I think that was a very clever addition.

BTW, you could take the undo history and store it in a centralized database, and have everyone shar the same project, complete with history on every object.. Maybe in the future a blender/cvs integration? Multi-user log-file style environment.
The hair, dynamics, softbody collision, elastigirl-IK/Fk switching, theese all need to work so animators can do their job: putting emotion in the acting, not "getting the software to do the darn thing with workaround 12 and rt: on 6"
Well if you consider what the animators used to do (drawing frames, doing tweens).. You could easily outsource a project to Russia where someone is being paid to do frame by frame vertex tweens.. Okay I'm kidding.. But really I've done animations using nothing but RVK.. If you can parent objects, combine/link scenes from seperate artists, layer scenes, swap out low-poly meshes for high poly meshes, which you can do in blender, thats enough to divide the work. What Pixar and EA do in house is far mroe technical (even with the artists) than is done with blender.. Its technical because environments are coded and recoded, bridged, hacked, etc..
The developers at these companies also need some level of abstraction in
their own work so they can focus on designing armatures and advanced effects rather than recoding wheels.. Blender has the potential to give them mroe access to the nuts and bolts than Maya would. But time will tell.
I think Toon_Scheur is wrong:
I'm not bashing blender because it's opensource. It's just not as good as XSI / Maya. So what? Stop pretending it is, it's of no importance.
And no blender does not have "edit modes/ tools that's up to par with other packages" (Ever tried a trimmed surface?).
And for animation the whole edit stuff is unimportant anyway.
Trimmed surfaces are of no use, they are merely curves that define where a NURBS surface is not. If you like wating around for your models to
recompute the fillets on every frame and to toy around with tangent curves and surface twisting, single-sided polys, vertex normals, etc.. Then
welcome to the wonderful world of extraneous 3D software implementations (with the advent of subdivision surfaces, beit not as precise as NURBS surfaces, you can do things more efficiently than you can do with a toolbox of Alias or Rhino, those work well for rendering still-frames, nto for animation work).. The programmers tend to go off on unecessary implementation tangents (trying to implement reality) whereas the artist would simply exploit a simple tool to do everything.. The things that concern total artists and total programmers tend to be quite different than those who can do a bit of both.

Coming from that camp, I think too mcuh work could be applied in the wrong areas, such as for implementing trimmed surfaces. I would rather just put more uniform NURBS primitives in blender than the weighted ones that produce inconsistent tesselations of quads (take the NURBS sphere).. Simply changing blender's default primitives would be signficant for doing work with blender..
"Blender is capable to do all the stuff you see in most CG movies right out of the box. "
No it's not. Go and take a better look, or better watch a making of.
Pay interrest when you see:
- The hair on the blue monster in mosters, inc.
- The cloth in the incredebles.
If you want to animate thoose "by hand", good luck, hope you finish your movie this century.
So I should care about being distracted by the hair on the monsters
and not notice the animation work? Great cloth, skin and hair will nto make an animation or a movie.. Without the animators, its just a temporary "wow, neat hair".. But you think waving the ahir around ont eh screen will keep the audience entertained? If you think so, I have some CS majors that attend SIGGRAPH every year to discuss accurate physical simulations of dogs, with artists laughing in the distance.
That's not what this post was about? was it?
Why would you go for an opensourse, 6 developers, dictator ruled ( ;) ) software package when you have $480.000.000 dollar to spare?
The question, can you do this with blender.

And furthermore, can you do this only with blender..

And eventually, could you conceive that blender would be done
to do this..
Can you make an image of a hot chick in blender?
Ofcourse you can, put time and effort in it, and she'll be marvoulos.
Well you could go borrow on from Renderosity or Turbo Squid..
The idea is, what makes blender less special vs. Maya, or product X?
I wonder how long it will take for the pile of oranges and the pile of apples to mix and mingle again.
Hey its called cross polination it happens.. However God made the apples and oranges.. Man will still have to try to figure out who made the red blood cells aerodynamically perfect.. Any good design takes a good designer.. Blender is a good design.. Don't let the big boys with their money and clever marketing belittle blender's development. It's having an effect.. Tools that are added to Maya are done so more, I believe, just to encourage people to upgrade, than to provide any real value.. Blender's updates encourage people to upgrade, only if they want to, there is not the concern of the cost, which makes the choice easier.

But what does Alias do if people don't want to upgrade? The undesireable side effects of this choice (lack of support, obsoletion) will outwieght the advantages of blender over the other packages.

But again I could play that "low-profile blender sucks" card if you like.

Post Reply