VFX Reference Platform

Nexus VI by Fensch Toast, VFX by Romain Toumi.

When it was decided that Blender would follow the VFX Reference Platform, there were a couple of assumptions made:

  • It will help adopting Blender in a big studio environment.
  • It will facilitate the industry to contribute to Blender.

During the past two years there was little evidence of any of those outcomes. Contrary to that, for us it was perceived as a limitation with users lacking benefits such as the latest Python version in Blender.

We are aware that not following the VFX reference platform generates friction for developing add-ons and integration in studios pipelines. We’d love to hear of such situations. We’d be happy to collect useful information for studios on blender.org for it and welcome their contributions.

With that in mind, the decision was made that Blender will no longer stick to the VFX Reference Platform.

For the Blender community and eco-system as a whole it is more important to be able to use the latest Python versions and other libraries. We are still committed to preserve file format compatibility and to avoid library conflicts for integrations with binary add-ons.

This decision was made considering the arguments presented in the bf-committers list, besides a meeting with the Blender admins and Foundation chairman Ton Roosendaal. Thanks everyone for pitching in.

Dalai Felinto
Blender Development Coordinator

Amsterdam, 4th February. 2022

80 comments 22,678 views
  1. Saying that “This decision was made considering the arguments presented in the bf-committers list, besides a meeting with the Blender admins and Foundation chairman Ton Roosendaal. Thanks everyone for pitching in.” is pretty much like saying “This decision was not discussed with anyone outside of our echo chamber.”

    No one uses mailing lists these days. It’s ancient means of communication. Stop using them as channels for such important discussions. It seems you did not even actively reach out to any of the people this may affect the most. This was a huge decision, so I am left speechless by how haphazardly this decision was taken.

    All the work that’s been put into trying to make Blender mainstream VFX software is pretty much negated by this.

    It’s like if Google said “After a discussion though a few hand written letters delivered by a mail pidgeon between a few of our top executives, we decided to end Chrome Browser support for windows.”

  2. I’m a pipeline TD with 10+ years of experience in the industry, worked on animation and hybrid features.
    The last company I worked for was Animal Logic, I was there during the time when the studio transitioned from Python2.7 to Python3.7, and left before it was finished.
    I can describe roughly what it means to switch from two python versions which are not compatible for a studio:
    – Thousands and thousands of dependencies that have to be checked to make sure they have a compatible version, find an alternative if not and adapt where needed.
    – Thousands of in-house scripts that have to be modified and tested.
    – Deployment has to be done with extreme caution as you want to disturb productions as little as possible, for shows that are running full steam you just don’t transition them as it is too risky.
    To summarise, that’s a lot of organisation, work, time and money spent. Production budgets don’t account for these kind of things, it is an investment done by the studio (no producer wants to see the show’s budget go into something that won’t be seen on screen or won’t help artists).
    I’m sure there are still a lot of studios out there which haven’t started this transition and will do it when they will be cornered.
    No studio wants to go through this kind of thing every few years, as well as having different DCCs requiring different Python versions because it becomes a nightmare to manage. That’s the reason why the VFX Reference Platform is important and should be followed.
    So yes, like many people have already said, this decision is unfortunately throwing a spanner into the further adoption and deployment of Blender in studios, which I think is the opposite of what you would like to see.
    I have started to train on your software a bit before this announcement, hoping to find a position in a studio using it, but I’m not too confident now seeing that your decision impacts pipeline development, I don’t think many studios will be willing to take the risk of having to revise a working pipeline every few years, with all the financial implications it means, to be able to benefit of the latest improvements of Blender.
    I hope that this insight helps you understand the importance and raison d’être of the VFX Reference Platform as well as the consequences of your decision. Maybe that will help you reconsider.

    • No doubt this was a large change, but I don’t think comparing upgrading from Python 2.7 -> 3.7 to upgrading between minor point releases of Python.

      • It is true that this is only a part of a bigger picture of why the VFX Reference Platform exists, and I do agree that the amount of required changes is not the same between a change of major version and minor version, though you have to follow the same process of auditing your code, doing the required changes, testing and deploying. Ideally you don’t want to do that too often.
        Most studios use multiple DCCs, if these multiple DCCs are not aligned on the Python version they require and these different versions are not fully compatible between themselves, and let’s say the studio has developed python tools that are cross DCCs (with common base modules and DCC specific modules) the pandora box of managing multiple versions of a same script pointing to dependencies of different versions gets open, and that is one no one wants to open.
        In a nutshell, it may seem like a benign change, but you need to think of the cascading effect it has on a pipeline and all the factors involved (multiple DCCs used is one of them).
        That is just for Python, now let’s imagine the Linux distribution version requirement changes and that version doesn’t play well with your other DCCs.

        • I have no practical experience in a VFX studio and don’t know what I’m talking about, so all I’m saying is just theory and unqualified judgement calls.

          You are talking about a big studio environment with lots of interdependent custom scripts and software.
          The typical dependency mess that is hard, if not impossible to avoid.
          The idea that a new software, Blender in this case, might be introduced into this complicated ecosystem to me sounds relatively far fetched from the get go.
          From a studio engineer’s perspective it’s undesirable because he introduces another dependency into his existing hairball of software.
          Yes, some artists might like some of Blender’s features but if the tool needs engineering work for adoption it’s just not worth it.
          According to this blog post, there was no adoption. So why spend time and money on something that your potential users don’t care about?

          Tangent Studios adopted Blender, but they started from a blank slate. New studios and small studios imho are realistic candidates that might benefit from adopting Blender. But trying to cater to the big established players seems useless.

          I see that we might have a hen and egg problem here for big studios.
          Do you think that’s a fair assessment?

          • Thanks for providing some context to your reply, I have chosen to reply to the different aspects of it.

            – “The idea that a new software, Blender in this case, might be introduced into this complicated ecosystem to me sounds relatively far fetched from the get go.”

            When XSI was killed by Autodesk, all studios using it got forced to switch to another DCC, Animal Logic was one of them, it was not pleasant but we did it. In the last studio I worked for, the Asset departments recently switched from Maya to Houdini. It’s not far-fetched, if there is a better/cheaper/more efficient software solution, studios can and will switch part of their pipeline (it’s all about the money). If you read the comment from Thorsten below, he also explains how a software is usually introduced into a big pipeline.

            – “Yes, some artists might like some of Blender’s features but if the tool needs engineering work for adoption it’s just not worth it.”

            If it was the case no medium/big studio would ever change a software for another, that’s not the case, it is both a matter of money and staying in the competition by producing images of equivalent or better quality than other studios while maintaining the cost to the same level as much as possible.

            – “According to this blog post, there was no adoption. So why spend time and money on something that your potential users don’t care about?”

            As mentioned in other comments the development cycle in studios is around two years, which is when Blender started to follow the VFX RP, that’s also how much time a movie production usually takes. The VFX/Animation industry is a big beast, there is latency, if you look at Pixar’s USD, although many studios find it exciting, not many studios are using it (and some that do use just a few features of it) despite the fact that it has been made open-source 5 years ago and that it is coming from a major industry player, it is happening but slowly.

            – “Tangent Studios adopted Blender, but they started from a blank slate. New studios and small studios imho are realistic candidates that might benefit from adopting Blender.
            But trying to cater to the big established players seems useless.”

            It’s up to what the Blender foundation wants Blender to be or become, if they want to keep targeting hobbyists (it’s not pejorative), small studios, companies that are not in the VFX/Animation industry it’s totally fine. If they want to target the above and the whole VFX/Animation industry (medium/big studios), which is the signal they sent two years ago by following the VFX RP, they have pulled the rug when things started to take off and for what I can see, a lot of comments written by tech people working in the industry show that.
            I do understand that some Blender users want the software to be cutting-edge in its adoption of the latest version of Python or any other dependency it needs and I hear that, it is just not compatible with its usage into medium/big VFX/Animation pipelines, that’s what it boils down to. But if the Blender Foundation wants to cater for both, instead of blocking everybody to dependency versions requested by the VFX/Animation industry, there is the possibility of having two flavours of Blender to avoid the “one size fits all” (I’m not implying that it is an easy thing to do), one that follows the VFX RP platform and one that is not, so the user base is happy and so is VFX/Animation industry. I’m sure there are potentially other solutions as well.

            In my opinion it’s really about what the Blender Foundation wants the target of Blender to be, if they want it to include medium/big VFX/Animation (and I guess Video games) studios, they need to provide a solution for Blender to follow the VFX RP and stick to it, two years is not a long enough time to see the initiative bear fruits even as a try.
            If the target does not include medium/big players, it’s fine as long as it is clearly stated and communicated so there are no surprises.

        • Frederic,

          I wholeheartedly agree and thank you for sharing your experience in the VFX and Animation industry.

          As it stands, it appears to me that unfortunately the Blender Foundation has very little knowledge of how Blender users actually use the software and under what scenarios professionals in our industry work.

          Nobody just uses just one software out-of-the-box as “Blender Studio” does and even they modify it to cater to their production needs (the result is then a new Blender version for the community, which I find as a practice very shady and a one-way-road/top-down practice).

          They disregard the meticulous work even single-person professionals, small teams, let alone studios put into their workflows (!), the 10-100s of Python packages, modules and dependencies needed for the work (and not all of them will update to 3.10), so a common reference for these tools are extremely important.

          I hope that at least Blender LTS 3.3 due in May 2022 will adhere to the VFX reference platform.

  3. Hi friends I don’t work in those big companies but have used Blender for vfx for over 2years by now and been waiting for vfx update and improvement but so sad to hear this news which makes me lose hope of what I thought to use Blender for
    I always use it knowing one day we will make it to the top so God what’s happening our software please instead of improving just wants to get reed of it’s part of vfx ooh God

  4. Confused about the need for this. Python 3.9 was released just 1.5 years ago and will get security updates until October 5th, 2025.

  5. I work on games art pipelines (formerly Ubi/EA, now a startup with Sony) and we’ve been trying to target Blender as our main DCC for the last year or so (which is great!), but this move has effectively locked us out of new releases.

    New Python versions require you to recompile C/C++ modules for it, and large pipelines tend to have huge dependency lists. Python 3.10 is simply too recent, and it’s unrealistic to think that we can already expect all the libraries we depend on to have made new binaries available.

    I was testing out the Blender 3.1 RC because I was tracking down a Blender crash with file loading, but our pipeline immediately failed to load, because the Perforce API is currently unsupported on Python 3.10. This is just one of the many examples of things that will break because of this move.

    The VFX platform ensures that more reasonable time is given to all parties to make the necessary compatibility changes to their stuff, so that we can move forward together. I really don’t see anything valuable gained from a precipitated move to Python 3.10, but I can already tell that we are now unable to upgrade to Blender 3.1.

  6. This seems to me an incredibly irrational, short-sighted, premature, unnecessary, and self-defeating move.

    —m

    filmmaker and former developer with Softimage Special Projects

  7. I have two major thoughts about this change.
    First, the conversation between the devs didn’t seem to come to any conclusion about whether Blender should abandon the VFX-RP. Instead, the conversation came to a standstill (which was acknowledged by the devs in the conversation) at which point a dev literally just went “I know we’re disagreeing about whether we should just do what I want to do, but because we disagree I should just be able to do it right?”. This is not how decisions should be made, and anyone who thinks “we can’t decide between decision A or B so that means A right?” is a reasonable argument shouldn’t be part of these discussions.
    My second major thought is that people in the Blender community seem to act like there is a hard division between professional users and hobbyists and I think that is a mistake. I am a traditional artist and digital painter who recently discovered VFX work and have been primarily introducing myself to it through Blender. One of the best parts of being a Blender “hobbyist” is the knowledge that I’m using a professional quality tool, the most powerful art making tool I’ve ever used. In fact, one of the reasons I got so serious about learning Blender was that I was hearing about its adoption in professional contexts. Blender finally began to escape its reputation as “babies first 3d program”. This directly lead to myself and many others deciding to learn and use Blender. Now I am finding out that for my purposes Blender is likely to remain a software for hobbyists, and if I’m going to need to learn to use one of your competitors anyways, well you’ve just removed all of my reasons not to do that in the first place.
    Blender is truly an incredible tool and regardless of all of this, I am thankful that it introduced me to 3D and VFX. But even as someone outside of the VFX industry, this level of recklessness and open-disregard for your users makes me think my time and energy might be spent better on software that hasn’t showed a willingness to pull the rug out from under its users.

  8. Maybe we should blame the python developers they made it that way, it is not BF fault the backwards incompatibility.

  9. I am not in the VFX industry personally, but I can definitely see the advantages by keeping to industry standards for some parts. It is what is making blender’s adoption into the industry possible. And rightly so. I am saddened to hear those who work directly in the industry have made very clear and realistic paths forward that blender leadership could consider, but surprisingly blender leadership doesn’t want to listen. It does seem like a rash decision with little to no conversation with the industry directly.

    I have been a blender user almost since it became open source, and I took it up completely after Autodesk acquired then killed off XSI. I have seen the amazing shift in both blender’s development and the adoption of it into most major fields where DCCs are needed especially in the past two years. Coincidence? I think not. The industry IS responding. The Blender Foundation is blind if they cannot see this. The fact that blender has had this support both in cooperation donations, to full-on integration such as to render engine integrations through to the last installment of Evangelion which used Blender predominantly in its 3d scenes is testament to this fact.

    I can not see this decision as being a good decision for more than just The BF trying to keep up with Python. There are so many young people in animation schools who have been recommended to start taking up blender while they learn Maya (insert any other DCC software you can think of). I know because I have just taken on interns at my studio who when I suggested they learn blender stated that they have all been using it over Maya due to their teachers suggesting it will be the software to know in the future. This is from those who will be the ones to decide in the future which software should be adopted into their studios.

    This is a case of taking one step forward but then two steps back… literally. And I am probably one of your biggest fans and proponents. I have held seminars to push for Blender’s adoption into the local industry here in Japan since version 2.55. Fighting an uphill battle, just like pissing against the wind. But I now have seen progress these past few years. I am saddened by this rash decision, being made without consultation with anyone other than the main devs. I feel that this is a huge mistake that the Foundation is making purely because what appears to me, from Ton’s comment above, is a case of someone’s pride getting hurt or even worse impatience. I hope this truly is not the case.

    I am a nobody compared to the likes of the commentators here, but I am still hoping that this decision can be reconsidered or at the least Ton use his incredible leadership skills to help the VFX RP move ahead with their own restrictive hard headedness. This is sad. Very sad.

    Please give this one more serious consideration. If not for the BF’s benefit, then for that of the up and coming VFX creators and future TD’s who may just be taking up blender now because the industry standard was implemented in the first place.

  10. ok a bit late to the party….so

    I have my own thoughts about blender allways trying to be on the latest and greatest , but this is not a comment on this, but more proposal for a compromise:

    Why don’t you follow the reference platfrom just for the LTS versions ?

    Update to the new versions as you want, but that way, people can integrate it into bigger pipelines (they should use the lts anyway, shouldn’t they ? ) and give all others a way to use the latest versions ?

    • You have a really interesting perspective and I cannot agree more. That way youtubers will have a chance to get a hang of it for things like tutorials, while bigger productions can stick to the current system. If only there were a way to get a dev to see this lmao!

  11. Hi, I am pipeline TD in one of the biggest studios out there, this decision has been a mistake, I understand that you want to push Blender to use the latest and best libraries out there, but you can’t see is that Blender is getting adopted everywhere and growing, in my studio and the previous one, Blender has been requested, and our team has been working in the implementation. Now, this is gonna make Blender development stop for the time being as it is really hard to maintain our software and all the variants of the dependencies.

    I hope you change your mind. Your average user doesn’t care about this as the big studios that rely on building pipelines around Blender do, and this change is making Blender leave the industry, not because we don’t want it, but because there is already too much work and implementing Blender outside of the VFX platform is too much.

    Just as a note, it could take more than a year just to change from Python 2.7 to 3.7 for most of the big studios, change takes time, in 2 years, you didn’t actually evaluate much.

    Best,

  12. Hi all,

    I am a VFX Supervisor at what is one of the largest VFX studios in the world. Over the past 17years I have worked at MPC, Prime Focus, Glassworks, FrameStore, ILM, for Netflix, Apple, Amazon, Warner Bros to name but a few, both in commercials, episodic and feature film. I’m putting names here so the “industry” becomes a more concrete object in the discussion.

    There is a lot of discussion about industry adoption, but I don’t think the subject is well understood so I will try to represent this side as best as I can.

    For reference I gone through every page on https://lists.blender.org/pipermail/bf-committers/2022-January/051234.html on the topic to try and understand the basis for this decision. It seems that the actual decisive factors where taken in a meeting as the thread falls short of the actual decision.

    I don’t believe that the developer’s group and consequently Blender’s admin board fully grasps the reach and impact of this decision. Concern over the user base is referred multiple times in the thread but I don’t think that the impact of the decision is truly understood.

    Blender’s reach into the VFX industry has exploded over these past 2 years. Every single company I have worked for has used blender in one form or another, and so do the freelance artists we hire, project in project out. This past couple of years it has been across pretty much all departments. The VFX industry has started to adopt, use and truly love blender for what it can do.

    We all agree, that the VFX platform does need a major rethink. It’s an old dinosaur that changes according to a geological timescale. At the moment even our own development team is straying away from it on a couple of old libraries. But this is exactly what studios do. This how the VFX platform is being used, as a guide. This is what Blender was doing and very well. All other DCCs developers are doing the same, using it as a guide and straying away when it’s not sane to follow anymore.

    The thread on the topic became gradually blindsided with an all in or all out stance mainly because of python. This is completely the wrong approach. Keeping as many libraries compatible as best as you can straying away on the ones you can’t is exactly how it needs to be. Again this is what studios and DCC developers do, using the VFX platform is a guide.

    A guide which you just from one day to the next chose not to follow. This hard cut decision reads as immature and impatient. From the discussion thread this became a seemingly inconsequential decision from developers catering for developer’s needs based on assumptions and speculation about their users.

    No consultation has been made with the VFX studios. No dialog or approach has been made so we as an industry could have a voice or together with the Blender group come up with a solution. As far as I can understand not even the VFX platform group itself has been approached to understand if things are changing. No real approach has been made to the user base, if nothing else just to grasp the seriousness and the impact a decision like this could have.

    This a brash one sided decision from an amazing team that is now alienating the very industry it began to conquer. This is Blender not supporting Centos situation from 2 years ago all over again. Which thank goodness at the time you did reconsider.

    I fully understand that blender is used in much more than VFX but, once again, I truly don’t think the group understands how far reaching this is and how much blender is being used in the industry.

    This is just alienating everyone. Most likely new plugins, render engines and addon development will severely be affected. There have been hundreds of thousands of dollars poured into new development for Blender that is now most surely becoming redundant. Maintenance will become a serious cost if not impossible and will be dropped.

    I urge the Blender team to reconsider. I propose that this conversation is re-opened, that the team becomes open to the subject once again, listens to the industry consults with the studios (if nothing else at least those who are already your patreons) consults with with the VFX platform and rethinks the decision.

    Blender needs to become a reliable tool more than anything else now. A reliable tool, from a reliable group making reliable solid decisions. Decisions like this can’t be made one sided. Become open, approach the studios listen and consult with the user base you fought so hard to conquer.

    Looking forward (and hopeful) to seeing this re-considered and Blender on the right track again.

    Kind regards,

    N

    • Couldn’t have said this better myself.

    • Why should all of the other industries revolve around your update schedules?
      Why not just update your libraries if you really want to stay compatible with Blender?

  13. what a terrible decision, clearly without any regard to an industry that cares about stability beyond chasing features.

  14. As an outside observer, it sounds like there’s a major conflict in release schedules.

    Instead of playing favorites, it appears Blender Devs choose to remain dedicated to being aligned with the Python release schedule so as to not be caught playing favorites since there’s more than one industry revolving around the usage of Blender.

    Perhaps VFX-RP needs to rethink how to remain compatible with Blender since, again, it’s a multi-industry tool.

  15. Personally, I think Blender should work more with other super cool open source developers from other fields. There would certainly be completely different new and creative software developments… Far away from the “ordinary” software that is also super cool like photoshop 😉 XD.. And in my opinion also better. Good time to you all!

  16. This seems very short-sighted, and the responses from those involved seem like nothing but hubris.
    If you are making moves to exclude the VFX/animation industry and adapting to pipelines that have been years in the making to the tune of millions of dollars then you might as well abandon the Studio productions Blender Studio is making because they seem to be geared to attracting artists and studios and proving that Blender is a tool that could be adopted by the industry.
    You won’t change an industry that is predicated on having solid, stable tools and pipelines that allow for productions to be completed on deadlines.
    Let us use stable builds of Python if we want, let us use Y up for a change… if Blender is so damned adaptive then at least make that a toggle!

  17. Hey all I’m the author of GafferCycles and work in VFX, my 2c:

    To be honest, I’m hoping Blender sticks to building on Redhat’s GCC 9.3.1 + glibc 2.17 as long as Blender boots in CentOS 7 (granted it is ancient and unsupported) then that’s the most important thing for studios to use Blender. Everything else can be worked around I’d say, but that’s the main thing IMO at least for now until all the studios figure out what the next distro will be. I think VFX platform rolls at a snail’s pace and needs a bit of a kick of innovation though, perhaps the move to Python 3 will make things much better.

    If you make future builds of Blender not boot in CentOS 7 then that’s a big loss. Studios do have the resources to build their own though, but probably wouldn’t go to the effort with that inertia. Currently I see Blender making a lot of in-roads for generalist and asset/modelling departments so it’s fairly self-contained without needing tight integration of existing pipelines much like zbrush’s role, with imports/exports using OBJ/Alembic/USD or just delivering EXRs (they need to be more compatible with Nuke though, probably).

    • Yes, we’re committed to being VFX platform compatible. This means running on systems which are also supported by the VFX platforms as well as being file-format compatible.

  18. I’m a pipeline td in a midsize studio. That was the last step to end our pipeline blender tests. Multiple Python Versions are not worth the hassle

  19. I work as a server developer for a Saas platform. In my latest jobs I’ve learnt that keeping up with recent versions of our support libraries was easier than dropping way behind and having to catch up. At first it was a bit scary, but nothing awful happened really. This is more of an industry habit, I guess that it keeps us safe from breaches, works us out of bugs and stuff. We see it as being reactive. There is no way I would go back to dropping versions behind. That would be a nightmare.
    So to me it is a very understandable move for Blender. Too bad the vfx industry moves so slowly. But I’m not sure that is up to Blender to slow down?

  20. Two years it’s nothing in the VFX industry. Things in big studios move slow, this is why we are still using an old Python version. You need to be more patient. There are a lot of middle size studios, and a couple of the big ones, thinking or starting to think to adopt Blender.

    Maybe you are thinking that you can push the industry to adopt a newer Python version, but again, if the software changes his language version so quickly, the VFX pipelines departments to the things they know that will work. Just when you were starting to conquer the studios, you were going to scare them away.

  21. If the goal is to get into the big studios, you could start by allowing users to work Y up. Z up is only used in the game industry. In film, all 3D software (except for 3ds Max) are Y up, including Nuke. IT gives us headaches on a daily bases. And Skyline is the only big studio using 3DS Max. It’s not just for the geometry. The normal and position passes won’t work properly in comp. There are workarounds but it’s still a real annoyance.

    • Maybe the main goal of blender is not to being used by these “big studios”
      “Z up is only used in the game industry” that’s OK, but if you go to https://fund.blender.org/, and take a look at who is really paying Blender and his developers you’ll see, Epic, Unity, Ubisoft, Steam, Decentraland , Giants, OP Games, Activision, … It’s not a surprise that they get what they need.
      OSS is not about asking functionalities for free, it’s about collaboration, if the “big studios” want/need some functionality in Blender, they have enough resources to create a patch, fork the project or became a Sponsor.

      • I agree.
        Blender is a multi discipline tool. Let’s see if VFX studios are willing to pitch in to get what they need. Animation studios that are invested in blender have rolled up their sleeves and created addons, implemented technologies or have improved on already existing technologies to get what they need. A move that is of benefit to the Opensource community. These are not even big studios with big budgets.

        Blender is not Maya, Houdini or any of those propriety software. It’s design philosophy is different. It’s release schedule is different. It’s user base thinks differently about software development.

        I trust the direction that the development is going.

  22. What a horrible decision.
    So, you guys decided playing with the newest python version is more important than Blender playing well in an vfx studio ecosystem? This is a big middle finger to all the studios that are in the process of looking into bringing Blender into their pipeline.

    • What are you talking about? They are saying there is no evidence VFX studios are adopting Blender or making contributions to it. If there is a middle finger is the other way around.

      • So because two years in, the VFX industry hasn’t fully adopted Blender, so the BF decides to throw in the towel? Others have pointed out what an undertaking altering pipelines is. If anything, it would seem now that the BF is making it official that Blender will be excluded from VFX pipelines and the industry as a whole. Is this the writing on the wall?

  23. This is disappointing on many levels. Besides hindering adoption even more, it is also very disappointing to see you claiming the industry would not give back to Blender when Blender is using so many big OSS Projects that started in the VFX industry and are still managed and developed by the Industry. OpenEXR, Alembic, OpenSubDiv, OCIO, OIIO, OSL, OpenVDB, USD to name a few.

    This is one of some major reasons not to adopt Blender. Which is sad, because i would love to.

    As many have already pointed out 2 years is not even a full project turnaround here (Visualization and VFX). We also run big pipelines with a lot of custom code, so we won’t be able to “simply switch to Blender”. We need to be able to transition individual parts of the pipeline that work well in Blender. Which is prohibitively more effort as long as Blender does not stick to the vfx Platform.

    I get it, it’s effort that you would like to spend on other things, but this is a huge problem if you are working in a “not 100% Blender environment” as most people do.

    • – The studios can use the current and older releases for as long they wish. They use commercial apps of many years old too.
      – We maintain an LTS version that is largely compliant still.
      – It would be a relatively small investment for studios to make Blender work compiled with the libraries they wish. We’re happy to share such info.
      – If anyone from the film/vfx industry would wish contribute to Blender, we’d happily onboard that person.
      – As chairman of Blender Foundation I have not been contacted by any person in the vfx/film industry with stategical/purchasing authority to discuss using Blender, not ever. They all know my email address, and I visited SIGGRAPH yearly since 1999. I will always be open to discuss working together.

      Most importantly: I know the big studios have their own pipelines and priorities. We have priorities ourselves; to make 3D creation software accessible for everyone on this planet.

      • Hi Ton,

        – The studios can use the current and older releases for as long they wish. They use commercial apps of many years old too.

        *True*

        – We maintain an LTS version that is largely compliant still.

        *Keyword here being “still”. Which suggests moving forward we will have to drop Blender adoption as it will be no longer be sustainable to keep it compliant.*

        – It would be a relatively small investment for studios to make Blender work compiled with the libraries they wish. We’re happy to share such info.

        *Not true. There is a lot more dependent on a large scale productions than simply making specific DCCs work on compiled libraries. Even with distributed app Images through docker or kubernetes in self contained environments it is extremely challenging to cater for all DCCs.This is exactly why the VFX Platform has been created.*

        – If anyone from the film/vfx industry would wish contribute to Blender, we’d happily onboard that person.

        *Hundreds or even thousands of VFX artists, developers and post houses already do. That is what the community is. Alongside many many other people from other fields. Not quite sure what the message is here.*

        – As chairman of Blender Foundation I have not been contacted by any person in the vfx/film industry with stategical/purchasing authority to discuss using Blender, not ever. They all know my email address, and I visited SIGGRAPH yearly since 1999. I will always be open to discuss working together.

        *This is very much the wrong message. As chairman of Blender Foundation you need to encourage the Blender group to pro-actively seek and engage with the different communities and understand your user base better. Also saying you have not been contacted by anyone with strategic purchasing power form the VFX industry is disappointing to say the least. You receive dozens of thousands of dollars yearly from Activision, Ubisoft, Steam, Epic, Microsoft among many others large VFX / Game industry players. How is that not the industry reaching out and recognising and using Blender for its value? What other message would the Foundation need more than taking large sums of money from industry leaders to further Blender’s development? *

        Sitting waiting for the phone to ring to be contacted for adoption bogles one’s mind. Like many others Blender needs to lead the way. If the VFX platform is holding you back, lead the way, communicate with them and help it become a better standard like so many other vendors do. Don’t tie yourself to it but help it change, pull it forward but don’t turn your back on an entire subset of your contributors just because it’s a bit tough.

        The Foundation is doing an outstanding job, creating what is now one of the most successful OpenSource projects on the planet. But the leading the way needs to be done on all fronts. The message needs to start from the top.

        Richard E.

      • I did not get a notification for a reply hence realized there was one only just now. I agree with a lot of the answer that Richard posted so i am not going to repeat that.

        One thing i would like to double down on though: I specifically mentioned some of the VFX industry driven projects that blender happily uses and integrates in my initial post. You brushed off any contribution from the industry again. That’s just disappointing. And that also brushes away the community contributions done by people in “the industry” as if they have a blender hat and an industry hat and these would be mutually exclusive.

        And with all the “industry is bad” vibe it feels a bit ironic that the Blender homepage has a huge banner on the front page claiming it is “Part of the Industry” (whatever “the” Industry means in that context, but it lists ASWF membership prominently).

        The main point that really bothers me though is, that people don’t complain because they want to bash blender (or the FSF or anything else involved) but because they want to adopt it and there are certain road blocks that prevents that. That still leaves you and the community every right to keep things and change things as you wish. But the general tone that seems to suggest this as ill-fated bashing from a bunch of freeloaders not calling you is disappointing.

        If all you want is 3d creation software to be accessible for everyone as a hobby, that is mighty fine. But if you want to make it accessible for people that want to make this a career than it is not so simple anymore.

        I would also like to voice my support for pretty much everthing “N” posted (as of the time of writing the latest comment on this post). Especially in regards to how this decision was discussed and finalized. I would have wished for a more open discussion and reaching out to pretty much anyone outside of the Blender developer mailing list to understand the impact of this decision.

  24. My 2 cents. I think this is a step back. I handle daily 3 versions of python on 3 different OS’s (research). And it’s not funny at all. Having DCC’s align on the same python version keeps binary sanity, switching to the latest greatest just because “you miss new features”, is a bad call. Of course for most addon’s the impact is minimum, but there are many of us that need to install much more python packages that what is bundle, and build our own packages to target a specific tools we use.
    I have seen the mailing list post-mortum (and missed that on the blender chat) and if I did noticed before I would have my voice against that change. The mailing list is a blender developer centric communication channel, so even thought my company has some SWE helping Blender, not all of us keep track of that. Also having such an important discussion with few developers/pipeline TD’s over around 15 days… I find it a “little” bad.

  25. This is a bad decision, just when the adoption starts taking off, you say it’s not fast enough and cut loose? Give us C++ API and nobody will care about imperfections of the current python API.

  26. Two years is barely a product development cycle. Most productions, at least in game-dev, will not upgrade software during that time. Cutting off VFX Reference Platform alignment so quickly after starting it is short sighted. For me, personally, this move also calls into question how committed the Blender devs will be to LTS versions of the program.

  27. If you’re looking for industry adoption, I feel like there should definitely be a long term supported “Studio Build” of Blender, that sticks to the VRP, is consistently reliable, and caters more to industry needs, with again, long term support. Keep the current state the way it is for every other user.

    Kinda like how Nvidia has Studio Drivers vs. Game Ready Drivers for their GPU’s.

  28. Just by acknowledging the number of people that disagree with your decision, if you were as truly open in the development phase you claim to be, you would announce today that you step back and cancel this non-sense decision…

    Who NEEDS python’s very latest version and is in a blocking situation without it?
    Except YOU, and your endorphin level fed by having the latest version, NOBODY.

    You made an effort between 2.79 and 2.80 to embrace industry standards and if you are thankful and not blind, that’s the very main reason why Blender is way more popular today.
    Since your little success is getting to your head, you are going back to your old bad practices that made Blender a joke for many years… sad

  29. I kinda have dishearten about it but I can understand about this decision.
    A lot of people doesn’t know that Blender relies on Python as its main foundation. While most of these other software that follow VFXRP are mostly made C++ or other language so they can careless about Python being old version and sticks to that. Meanwhile Blender is made with Python and needs to evolve to the new techs, so sticking to the old Python version wont gonna give any benefit to Blender at all for the future of it and I understand it.
    I don’t know if there’s a way for devs to align with the old version while keeping Blender internal to the latest version of Python *(since im not a programmer I cannot give any useful suggestions but only to express my perspective of this situation)

  30. Good move. “Industry” won’t adopt Blender anyway, they can add “it uses newest library dependencies” to their list of excuses.

  31. This is a disappointing decision , Blender should continue to follow the VFX Reference platform or adoption by large studios will suffer. I have seen a large uptick in studio Blender users in the last two years. Maybe as a compromise LTS release could follow the VFX Reference platform.

    • Yes, but like the first lines tell. its either still on vfx reference plattform or newest python. Since Blender is not only for vfx, i understand the decision. not prejudicating other fields of blender just because of vfx

      • Another solution would be to allow hot swapping python versions at runtime or compile time. Then you could just add a path to whatever python you want. Then Blender could stick to latest and VFX studios could stick to the VFX Reference platform if they want.

  32. You are definitely not doing yourself a favor with this decision. Just yesterday I was struggling with getting precompiled binaries for pyside2 and shiboken2 on CentOS 7. With python3.7 still supported I could have taken them from one of the other industry standard programs like Nuke. Now we have to compile them ourselves which is not fun and I would rather use our development resources on productive tasks.

    How hard would it be to create LTS-like blender3 branch that sticks to the VFX reference platform?
    How hard would it be to compile blender with a custom python version?

    • Drop by on Blender Chat. I can give you the script I use to set up a CentOS 7 VM for building the pre-built libraries used all over the planet to build Blender. After all, this is about collaboration.

  33. And at least it would be less of a shock for developers if there was a dedicated discussion on https://devtalk.blender.org/ before announcing it here.

    • There have been discussions about this on the bf-committers mailing list, which is the central place for such discussions & announcements.

      • So the community is great for giving 10 bucks a month and commenting in the void on the forums.

        …But the true decisions are in the semi-hidden committer mailing lists that 99 percent of us do not know/do not understand… I see.

        • Why don’t you make a referendum system, that is as accessible to the community as becoming a donator then?

          That would be maybe way fairer for everyone, no ??

          • The Simpsons episode where Homer designs his ideal car comes to mind. Maybe some people (like myself) shouldn’t be allowed to vote on things we don’t understand XD

            That being said, as an amateur add-on developer, I find the latest python useful.

  34. Hey Dalai, 2 years is way too short. Maybe you need to find a strategy so studios adopt Blender? I already see a lot of smaller studios using it in their pipeline really hard. If it’s only Python wouldn’t it be okay if LTS follow the VFX reference platform and the others don’t?

    • I would suggest to have a call with the VFX ref guys and try to find a way

  35. Hello,

    Will this decision affect LTS version? I’d like to suggest for LTS version still follow vfx platform.

    Regards.

  36. For us, the bigger reasons against adopting Blender more widely is it’s history of breaking away from or refusing to adhere to industry standards. Have you considered that the past two years are possibly not the best ones to base this decision on? Many studios simply don’t move that quickly and a reluctant to change or adopt new software mid-project.

    I think this is premature and find it odd that you state “We’d be happy to collect useful information” given that you already made this decision in isolation from the community or industry studios.

    • Saying this decision was made in isolation from the community and industry isn’t really fair. The topic was discussed over and over again in public channels, and the critical stance of core developers was quite clear:
      https://lists.blender.org/pipermail/bf-committers/2022-January/thread.html#51234
      https://twitter.com/tonroosendaal/status/1334500207497764865
      https://twitter.com/tonroosendaal/status/1198971612101566465
      https://developer.blender.org/T84752
      https://developer.blender.org/T68774
      https://developer.blender.org/T83246

      • Twitter doesn’t looks like a good place for a discussion of this sort.
        I have read the entire thread you are linking on the bl-dev mailing list. No more than 10 different people has expressed their opinions. Obviously they all are blender developers. I don’t call it a “public discussion”.

      • This is a really bad look for the BF, and the main reason I’ve distanced from the community and its development.

        Input from a dozen or so internal developers on a mailing list followed and known about almost solely by those devs cannot and should not be called a public discussion. If my dev team made decisions like that, we’d be out of a job.

      • I dont know why you guys tweeted it instead of making a Youtube Video about it to discus, Tweeter is not the place to do those debate as most people timeline are filled with garbage post, the only people who are able to see that post are the one who are at the same time it was posted… Twitter algo arent even working properly, same as facebook…

        So I recommended to post that on Youtube where a lot of people focus so much on Blender and also maybe tag known people like Blender guru to have a bigger vocal on calling the people of the industry to get it debate, rarely people actually goes to the blender dev forum to talk about it anyway so there’s that problem too… I never even heard this until now even tho im subs to the email list and devtalk forum I never saw the thread… so you guys pretty much fail on getting people to talk (maybe you guys need to hire someone who know SEO and SNS better with those funding to get people in the bandwagon.)

  37. Very bad move for making blender closer to industry standard.

  38. I’m sad to see Blender leaving the VFX Platform without giving it adequate time for adoption. That and the Python 2 to 3 transition has been a very painful one for most studios who have a ton of tools written for Python 2, and are putting off the transition to 3. I’ve seen Blender coming into studios, and as a pipeline developer, I would love to see more of it, as this swiss army knife cuts through a lot of tasks, and with a great base of support.

  39. Very bad timing for such a decision. I work in the VFX industry and this year is the first time I have been asked to install blender by different departments. Initially as a test, but with a chance to be used in future shows. The VFX industry is slow to adapt, there are tons of workflows based on years of developments, which represent huge investments. Some times there are new technologies that push a generationsl renew of technologies and workflows. Thanks to USD, virtual production, etc. we are in a moment of transition, and probably a good moment for blender to be adopted. Exiting from the VFX platform at this moment doesn’t look wise and very short sighted.

    • No kidding. I was reached out to just yesterday to take a full-time job to stand up a Blender pipeline at Warner Bros. That’s a big fish in the animation/vfx world. I can only imagine this news will look bad to them.

  40. What are those “newest libraries” you so badly need? It looks like some rushed and strange decision . Are there clear examples of what is problem with VFX reference?

    • Here is a concrete example where sticking to an old, no-longer-developed version of Python caused problems: https://developer.blender.org/T84752

      Yes, some Blender developer(s) could have tried and back-port the fix from Python 3.8 to 3.7, but that takes considerable amounts of time, effort, and know-how, and would deduce from actually working on Blender itself. Nobody from the “VFX industry” jumped in to provide patched versions of Python 3.7 that Blender could keep using.

      The VFX Reference Platform schedule is not aligned well with the Python release schedule, which means that for a considerable amount of time, The VFXRP will be on a “security fixes only” version of Python. I don’t think that’s a healthy situation, as illustrated by the crashes it caused.

      • This is a sad and really shortsighted decision.

        Just curious, how could these addons survive so long with these crashes? They must have been completely useless for years, until the new Python version arrived …

        Crashes and bugs are part of the show. Some are simply not to fix without to break compatibility, that’s correct. This is the problem of every pipeline. But leaving the show and the pipeline to fix these unfixable bugs means to have no show and pipeline at all.

        That for the short side trip into the industry. Back to the hobbyists i guess.

  41. This will make Renderman updates to Blender even harder. I think this isn’t a good call at all in the long run.

  42. Es una pena.
    Darse un poco más de paciencia.
    Por favor!
    Gracias.

  43. I think a bit more patience would be advisable to see the big studios integrate blender in their pipeline, but I think it’s unreasonable to think a massive industry switch would happen in such a short period of time. I don’t know the benefit that this decision can bring, but i hope it’s not rushed.

  44. One of the add ons we use the most is STEPper, a file importer for industry standard STEP files. While blender stuck to the reference platform it was easy to maintain this. With the switch to python 3.10 the OCC precompiled libraries are broken again. In essence this means to stay on blender 3.1 until the add on is updated again for us. When this will be is unknown.
    Not really excited…

  45. Hi,

    Two years stretched over the painful Python 2 to 3 transition is certainly not enough to make such an assessment. More and more studios are starting to invest resources into Blender and it is a shame to see that a perhaps shortsighted decision will jeopardise that effort. Adhering to the VFX Platform gives our industry an assurance that integrating our tools will not be a major build trauma.

    To the Python point and as an example, the VFX Platform 2022 mandates Python 3.9 as the current Python version. 3.10 was released in October 2021, is there really anything you need so badly from Python 3.10 that has prevented Blender growth in the past 6 months?

    For what its worth Scientific Python recommends Python 3.9 as the minimum supported version until around mid 2023: https://scientific-python.org/specs/spec-0000/

    Cheers,

    Thomas

  46. It’s very disheartening to see Blender leave the VFX Platform compliance without giving a real push on other vendors, especially when Blender relies on so many open-source projects maintained by the VFX community.

In order to prevent spam, comments are closed 45 days after the post is published.
Feel free to resume the conversation on the forums.