Why don't big film studios just use Blender?
polycounter lvl 5
Hi and Merry Christmas to all 
I want to ask this big question because I allways wondered why big studios just dont use Blender.
Is it because of the compitablilty for tools? Functionality? Because of that everyone else use Maya?
The big studios could have saved a bunch of money if they used Blender instead of Autodesk` high pricerange
tools. Have a look at the end of this page for comparing Autodesk tools VS Blender:
http://animation.softwareinsider.com/compare/1-19/Maya-vs-Blender
Its all the same.
(I do not have anything against Autodesk or intend to insult Autodesk by any means, as I use Maya myself and those guys are awezome and helpful)
I want to ask this big question because I allways wondered why big studios just dont use Blender.
Is it because of the compitablilty for tools? Functionality? Because of that everyone else use Maya?
The big studios could have saved a bunch of money if they used Blender instead of Autodesk` high pricerange
tools. Have a look at the end of this page for comparing Autodesk tools VS Blender:
http://animation.softwareinsider.com/compare/1-19/Maya-vs-Blender
Its all the same.
(I do not have anything against Autodesk or intend to insult Autodesk by any means, as I use Maya myself and those guys are awezome and helpful)
Replies
to make these studios switch and thereby abandon a good chunk of their investment, you have to offer something significantly better than what they already have. some do have inhouse software that replaces maya entirely for certain jobs btw.
That's why you tend to see adoption of open source software in the smaller to mid-size studios, because these can afford to train all of their 12 employees
You also see it being adopted more as part of a pipeline or only in a single department, and this is sadly something that Blender isn't very good at for various reasons. So usually it's something like the concept art department using blender but then the actual production work being done in maya or houdini.
The real question is why courses don't adopt FOSS, as that would be much kinder to the students' wallets in many cases, but there you get the problem that the courses wish to train you for the industry, and the industry uses certain products because everyone is trained for them. A bit of a vicious circle there.
first of, A year subscription of 3Dmax/Maya is less or about the same of an employee earns in a month. secondly why bother learning a new program when your perfectly fine with the one you got? also all the tons of resources already invested, training, decades of work habits, rigs, material libraries etc.
there is nothing wrong with either programs, so if money isnt an issue, why would you ever switch from what you have to the other?
as for the May vs Blender link, interesting stuff, but I think blenders higher score is highly influenced by its small but loyal fan base. I dont really think anyone is an autodesk "fan", im not a fan of the buss, but I still ride it to work everyday because it gets the job done =P
EDIT: One more thing, when your a big company, you are much more dependent on other people, and sending files over from different programs is not really an option, having an "industry standard" both makes, and re-enforces collaboration between team members and even outsourcing agents much easier.
If it's free/open source then the developers don't owe you anything, not even support.
But again, give it time, Blender is young and understimated.
Well, probably because it isn't
To stick with the subject, I aggre with @MmAaXx , Blender, for now, is not meant to be The Biggest Tool in a production : since it's easily accessible, I guess """Big Company""" (rather artists in it) use it time to time for specific needs.
For that kind of topic, take a look at the Blender Cloud and Blender Open Movies (Caminandes, Cosmos Laudromat, Glass Half), they do a great job to answer those questions.
Studios have accumulated years of experience and tools, and as time has gone on, the films and effects have become more complicated and pipelines have needed to adapt to that. So it's not so easy to replace and swap out for something else. It can be done, but I'd say its rare these days.
People can be retrained, and that isn't so bad, but the real cost would be replacing massive chunks of the pipeline and toolsets in order to make things work again. And this wouldn't be a short job either. That would take time and money. Sure something like Blender is free, but then there's also the cost of ongoing support and maintenance. Having to support their own 3D software on top of all the other proprietary tools they have is a big drain on resource.
Anyone (including Autodesk) offers their larger enterprise customers extra levels of support (for extra cost) which studios pay for because it provides them with extra services, like custom fixes, special builds and perhaps dedicated resource that is assigned specifically to them.
the other is just history/timing:
one of the first to really open up scripting and turning the app into a pipeline-toolset which was ultimately the killer feature if memory serves. before that generation of software, the tools were far less customizable in general.
direct successor to a product-family already established in the industry at the time - i think most of maya's general features were just gobbled up from various other alias/wavefront/tdi products and unified under one hood.
then-parent-company SGI also being the default hardware/OS vendor for pretty much all the relevant studios, cross-promoting the software and getting it out in time for some high profile productions of that day (e.g. ILM on star wars, Square on the FF movie, if memory serves).
close competitor softimage being several years late to the party (and on wrong platform for the time, probably thanks to being owned by microsoft) with a new version of their own tool also helped, i imagine.
Nuke is the same way.
Many studios used Softimage 3D, especially for animation, but Softimage couldn't get XSI out until 2000; 2 years after Maya. Studios did wait and adopt, but the momentum had already started to swing towards Maya. Same was happening in games too. At the time I was at EA, and we were switching to Maya as we were retooling for PS2/Xbox.
It's this level of openiness now that many people expect to see in any software package or technology. Features and demos can be great but a big question will be -'does it have an API and how accessible is it?' With multiple packages, applications and required integration, get this wrong and it can make or break a software technology
"What we have already works and does everything we need."
That's a HUGE hurdle to overcome. You have to offer an extremely compelling reason for a switch to even be considered.
A good example is Shake, the compositing software from Apple. The was pre-Nuke and everyone was using it, but then Apple killed it. They allowed studios to purchase the source code, which many did. The main reason in the short term was to help maintain their pipelines before considering long term options. Eventually everyone moved to Nuke. They could of tried to continue developing Shake, but they didn't. Think about why that was.
It's not that they probably couldn't, but in many ways its just not economical for their business.
You guys do know that you don't JUST have to program in C to add functionality to Blender, right? You can't make C or C++ plugins at the moment but there are loads of Python scripts out there already. And if a wanker like me can script Blender with Python with relative ease, I'm sure big studios with experienced R&D persons can do it too.
Besides, you will only benefit from source code if you have dedicated programmers. Not a luxury most small studios have. Scripting API on the other hand can get pretty deep while being easy enough for artists to pick up and make something useful out of it.
Free software usually mean less support too, If we have a problem with an autodesk software, it takes 15mins to call someone and have personalize assist from them, because we pay, because we're customer we're ENTITLE to support, someone who bought 300licences as a better service than someone who have a student license. Free product mean = no assurance, they could stop any development in a blink, they could make a broken update, they could bankrupt and suddenly ask for 3k per license, etc it's a huge risk to take for a huge studio. When you have a well oiled ferrari and the money to kept it that way, why would you suddenly change for a civic honda but still use it as a ferrari? The cost of a license is nothing to a medium-big studio. It's well invested money for them.
But regardless, this pipeline inertia shouldn't prevent any 3D artist with a focus on modeling from giving the app a try. After all, a program like Zbrush sits completely outside of game pipelines, yet is pretty much universally accepted as a powerful tool to produce high res meshes later converted to game assets. The same logic can be applied to any modeling tool, free or not.
Blender has better modeling (both polymodeling and sculpt/boolean) and UV tools IMO but that's not really enough to justify switching the whole studio over. And it's pretty straightforward to just layout UVs or just do the model in Blender and import it back to Maya. Once the threaded dependency graph becomes stable in master it might be possible to get higher performance out of Blender rigs than you might get out of Maya rigs which would be a win for animators.
Ultimately the reason a studio would want to adopt Blender is that it's open source and, therefore, any of it can be changed to fit the needs of the studio or of particular artists, including some core areas of the program that can't be changed in Maya. The big studios are gravitating towards open-source libraries because it saves money and gives higher software quality in the long run. Stuff like OpenColorIO, OpenImageIO, Alembic, OpenVDB is a huge win for standardizing the pipeline between any two programs. (Of course, all of these have Maya integration because it's what everyone is using.) At some point people are going to realize that there's a benefit to having the main program be open-source as well and not having to deal with some of the legacy cruft and baffling architectural decisions that Maya can never get rid of because plugins depend on it, all the while benefiting from fixes made by the community and by other studios, but of course it's going to take a while to effect such a change for reasons cited earlier in this thread. It's a very long-run sort of payoff and you'd have to be delusional to think that anything like Blender becoming the new industry standard would happen with any sort of speed, but Blender is getting more and more support from Pixar especially so I think it's only a matter of time before we start to see some more new, small VFX and animation studios popping up here and there that use Blender as their main package.
The savings from licensing costs are negligible compared to the better quality that happens when everyone's on an open platform. Linux is the poster child for high-quality software resulting from an open platform. It has commercial backing and source code contributions from many corporations as well as many individuals and runs most Web servers, supercomputers, and smartphones, as well as a good smattering of VFX workstations, embedded systems, and so on. It's one of the most scalable, customizable operating systems out there and it could never have been as successful as it is as a proprietary UNIX.
Odow: Not all of the bad stuff you listed is possible with a successful free and open-source software like Blender. It's possible with proprietary freeware but that's a whole different animal. Of course there isn't any official support, but the chances of development stopping entirely because Blender Foundation goes bankrupt are miniscule. Even if BF went and said, "You know that whole Blender thing? Well, we've decided that we're just going to be a super cereal VFX studio now and switch all of our artists and productions to Maya now because it supports our pipeline better and, by the way, we won't be developing Blender anymore. We're taking down our git servers, taking down blender.org, and ceasing all hosting of the Blender source code. Ta-ta, and good luck in your future endeavours." What would happen if such a random apocalyptic event were to take place is that some people who had Blender's git tree on their hard drives would re-host that source code on Github and Blender development would continue there, done by some of the people who were working on Blender before, as well as some newcomers who heard that the Blender Foundation had to shut down and who wondered what they could do to help.
The GPL doesn't directly prevent the Blender Foundation from charging $3k for Blender, but it does specify that source code must be made available upon request, so people could just as easily take the source code, build it for all the operating systems, and charge just $200 for Blender binaries. That's why Blender is never going to cost anything to download, because there would be a race to the bottom where ultimately it would be free to download and use anyway. Since Blender doesn't have a contributor license agreement, the people who write the code retain the copyright for it and license it under the GPL, so if the Blender Foundation decided all of a sudden that they wanted to change to a proprietary license or another open-source license, they would have to get written consent from everyone who ever contributed a fix to the program to do so. It would be a gargantuan task with a teeny-tiny chance of success and almost no benefit to anybody, and it would also violate the terms of the initial crowdfund way back in the early days to release Blender under the GPL. It's never going to happen.
Also, please go read up on https://en.wikipedia.org/wiki/Sunk_costs if you have the time. There would definitely be a cost to adopting Blender and making everything work, but at some point it might become lower than the cost of keeping Maya around and that's the point at which studios would want to adopt Blender. The money that was already spent on developing those Maya plugins doesn't really matter in the long run, it's comparing the cost of maintaining them, developing new ones, and having your artists continue to use Maya to remaking/porting the ones that are really needed to Blender or acquiring additional software that does what your Maya plugins did that and retraining your artists for Blender and having them use Blender that should influence the decision to migrate. Of course we aren't there yet and I'm not implying that we will be any time soon, but the gap is smaller than you might think especially once you're over the initial hump and have all of the tools you need put together. The money that was spent developing Maya plugins is technically lost either way, whether your main app can use those plugins or not. Sure, it feels bad to just throw money away and we shouldn't discount that, but let's be as rational as possible here.
There are instances here and there of free software projects selling official support like Red Hat but there would have to be wider adoption and commercial use for that to be a practical, profitable choice for the Blender Foundation to make. Krita tried to do something like that but it didn't work out.
Sorry for long post, felt like I had some things that needed to be said.
I'm a huge advocate for using Blender as a glue program, but it only seems to do the baseline very very well. Once you start using it for advanced tasks, e.g. rigging and exporting to an engine, it begins falling apart.
Open-source is not a problem, mozilla is open-source, linux is too, a lot of servers and the render farm run it, because it is better and because it is free.
I think that the AD support is not good tho, if it really is, why 3DSMax UV editor is still a piece of crap? ehehe
Actually they support you only if you have problems with installation and license, problem that you have only with licensed softwares.
If you find bugs or errors they simple tell you to make a bug report and hope that someone over there knows where to put the fingers maybe in a couple of bug fix.
that's why no single production in the world never gonna use the latest version of Maya, and sometimes they skip years before upgrade.
(oh and if you have viewport problems their first answer is to upgrade your GPU to a professional one or to a newer one...)
Like all the others softwares including Maya, I agree that Blender has some problems , as no solid UDIM support and not very good referencing system (but better than 3DSmax one imo) and no decent muscle system. yes it sucks.
Anyway while big studios are shrinking, artists sperimentig different solutions, thats why we have Blender and Modo for example, and thats why autodesk released Maya LT...
I think that for the same reason you can find MODO in a lot of studios around the world you can find Blender too, probably to became the big guy of the production (that actually means NOTHING) its gonna take a while.
Past that, everyone's already invested in a set of software, and if it works for them, fundamentally changing the stuff you're using may just be more expensive than upgrading every few years. Besides, Blender's a niche skillset. Finding good job candidates that already know it is difficult when most people were raised on Maya.
just one thing, Blender crashy? NO WAY! come on, Blender is solid rock from 2.4x series.
To address some other points...
As the thread asked about the larger VFX studios, when it comes to support these guys have a different level to those on the standard Autodesk subscription. As mentioned earlier, they can have dedicated resource and support, custom builds and fixes. Also, on top of all this, software vendors will undertake very specific work to help with productions. Many features that end up in the main products have come from bespoke work undertaken for a production. Autodesk do alot of this, but they don't really shout about it, probably due to NDA's.
Other vendors do this as well. MARI was mentioned earlier, well that was co-developed with The Foundry. That was cited together with Presto and Renderman as examples of studios developing software, which is a fair point. However you have to factor in the reasons for this in the first place. Those technologies were developed because there wasn't anything around that met the requirements at the time. Weta wrote their own crowd system (eventually became Massive), as did MPC, because there wasn't really anything commercially available.
Ideally studios don't want to do too much of this, but they have a job that needs doing, and that's where the business economics come in. Its well documented on the challenges of the VFX business, so controlling your costs is important, even crtitical. Whilst many do develop their own tools and software, pouring money into running and supporting a full 3D application when they could otherwise buy commercially isn't perhaps a smart use of funds. Sure they can save money by not paying in the first place, but that will get swallowed up very quickly on other things.
There's been some very good studios go out of business, and I'm not sure running more open source software would of saved them.
Studios are now really thinking hard about how and where they spend their money, especially when it comes to tools and technology. For years MPC have had their own coding libraries which has underpinned their own tech. They could continue this, but now they're doing more with Fabric Engine. (there's some good videos on this that are worth checking out).
When it comes to versions, many comment that no one runs the latest version and everyone's on an older version. This happens alot, but more recently, certainly in the last couple of years, studios are realising the heavy burden and cost there is to running and supporting a version that is several years out of date. After a period of time it essentially becomes no longer becomes cost-effective. So many are alot more up to date. Perhaps not to the very latest version, but probably not more than a year. Many will be on multiple productions on different schedules, so they couldn't really update all at the same time anyway. They might tag team it though in that some will finish and move up while others remain.
I'm not against Blender, I admire what they've been able to do, and how the group keep going. I think it's important they're still around and offer a choice to the user. They probably do enough to keep the main commercial vendors honest and on their toes. But I agree with the previous comment, that they're somewhat catching up and not getting ahead. And this is my main criticisim really, they're not really innovating. And in these industries and market, you need to be doing and offering something that is different to other people.
*First in the industry with LSCM unwrapping (since 2004)
*Non-overlapping UI paradigm
*Grease pencil tools that can be animated https://vimeo.com/113610809
*Sculpting with dynamic mesh topology http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.66/Dynamic_Topology_Sculpting
*OpenSubDiv https://www.youtube.com/watch?v=dzIl_S-qHIQ
And more... but I hope it gets the point across
Blender just doesnt have the portfolio of projects to prove it can be better. Autodesk development is slow though. Its an elaphant that cannot change direction quickly, which is why Blender has room to grow and develop.
https://www.youtube.com/watch?v=LLzCSzfQe8A
It's a Blender biped autorigger that was used for the Cosmos Laundromat hero rig. Everything is bones so it could probably mostly export to a game engine except for some corrective smoothing. Blender isn't quite as good at animation as Maya but I'd put it ahead of Max (although Max has auto rigs that are somewhat more straightforward to set up initially.)
You'll also always need some sort of composition tool like NUKE or aftereffects to throw all of the elements together, as well as a dedicated video editing program. Then there's also making sure you have camera tracking using some type of program or plugin to match the 3D scenes with the real-life footage and other elements like particle FX if they're rendered out as their own separate elements.
Basically even if you somehow got everyone to just use blender, you're still going to be using alot of other software for tracking, color correction, compositing, editing, etc. And that's not even getting into whether or not you're using VRAY, MentalRay, SpeedTree, FumeFX, Houdini, Renderman, etc. as well as a computer server farm unless you're doing a really simple style or don't mind waiting years for a movie to render.
One frame from the movie Avatar usually took several hours to render. http://www.geek.com/chips/the-computing-power-that-created-avatar-1031232/
So yeah, paying a monthly or yearly subscription so an employee can run software that they know and is set up in their pipeline with all of their other tools is next to nothing in operating costs as a whole.
Seeing how further it improves on every update, I believe it will be a major software and a big rival to Autodesk software in some years!
The CG industry is ever changing, the pipeline, the tools, the methodology of doing things. We're always looking for ways and shortcut to make things better, faster and of course cheaper...
I used to spit on Blender before, but then I started to use it and realized how powerful it was...it does everything you would want it to do and if not, just add a couple of addons (already included in the software) and it's as good as Maya..
Every day, new tools are popping out. Blender foundation is releasing new ground breaking demos and reels year after year...
I can't see the future but I can sense it...
Edit: I've read some really sweet stuff upcoming for 2.8..if the promises get held, it will have a nice impact out there.
One day, they'll start implementing a serious PBR viewport renderer and that'll attract a good deal of people (if PBR still stays a standard)
A point to remember here is that Blender isn't/hasn't been the only alternative. Softimage (nee XSI) was more than a credible alternative, he had history, experience and was production proven right up at the highest VFX level, in fact it still does. It has features that are still better than some aspects of Maya/Max, and also surpasses Blender and that's even before you include ICE. It was best placed to replace Maya and Max as a core 3D package, but even taking the Autodesk acquisition into account, usage didn't increase and now its been retired.
So if something like Softimage couldn't overturn the big studios, I'm not sure how i can see Blender being able to, imo.
I'm sure Blender is moving forwards with features, but don't forget so is Autodesk and it would be foolish to underestimate them. They get a hard time but there are some smart people in their product teams. They might appear slow, but some stuff does take time to get working. The big studios aren't in the dark either, they're kept up to speed on alot.
And lets not forget The Foundry and SideFX either. Houdini is a very good package and has significant advantages over anything on the market right now. Modo is a solid product, but still some way before it can match Maya in terms of a platform.
A good start for the Blender Foundation would be to redesign the interface from scratch and prioritize the relevant features, the tools are there, it only needs to make the whole software more industry and user friendly.
Anyway let's see what happen after 2.8 redesign.
For the moment I will continue to use Blender as my main tool.
ps: I'm really tired about people that know nothing about Blender but tell us what to change to make it a better.