Hi, dear Blender community!
It is couple of month since me and a couple of my colleagues started to evaluate Blender. We tried to understand if it is capable to fullfill our production needs. The interesting thing is that Blender is free and it is not widely used in world wide animation production! Isn’t it strange? Why would the studios pay dozens of thousands of dollars to get Maya,XSi or Max? Moreover Blender has all features that rival industry standard Maya. Blender has some things that Maya doesn’t have. Blender is around for almost 20 years.So why do people still stick to Maya? Have you ever asked your self about it? Why don’t majority of the studios adapt free, open source, customizeable, full featured software like Blender?
So here is my thoughts and facts about Blender:
First time I opened Blender was 9 years ago. I opened it and I saw THAT interface…I’am certified Maya user, throughout my career I had a chance to work with many other 3d apps…but what I saw when I opened Blender was just insane, I just felt that I had never used any 3d app before. The interface was like the developers came from different space and time where 3d took different brunch of evolution. After some unsuccessful attempts to understand the UI logic I successfully uninstalled Blender from the system.
Second time I opened Blender was couple of month ago. I think the decision to make more conventional interface was the best decision the developers could make. It is the best, just because now Blender can bring in more people that try to migrate from other apps. I know, for people who started to learn 3d with Blender that doesn’t make any difference, but believe me if you decide to work in the animation industry, Blender’s new interface will make your transition to any other software so much easy and vice verse.
So since the moment we saw the pleasant and clear interface we started to test it. We started to dig the rigging first, our TD would apply his knowledge to build the rig. Our layout supervisor loaded our typical production scene just to understand how Blender will handle all the data. And I started to learn how the file referencing (linking),data handling work in Blender.
After one week of testing in a spare time we come up with some conclusions. Our TD was not happy with the fact that the dependency graph is hidden and there is no way to make data flow in a user desired fashion. Meanwhile it is a big limitation it is still possible to make okay rigs even with little knowledge of the software.
Our layout supervisor had difficulty to organize the objects of a big scene. I tried to use outliner and it is very limited in functionality. The strange way to select objects is so uncomfortable, there is no way to drag and drop multiple objects to create hierarchical relations. There is now way to specify start and end objects to shift select them. In addition we couldn’t manage materials all at once.
Final thing that makes Blender hard to use in production is the file linking system. It is strange and also very
limited. I’am still studying it, but I see that many things are just locked and impossible to tweak. So in order to make specific changes the artist must append the file which will destroy all the pipeline flow. Check Maya’s file referencing system…this is how it suppose to work!
Generally speaking Blender has all cool features such as python, brilliant render Cycles, hair, compositing and sculpting but it is weak in fundamental things such as good file referencing system, good scene asset management outliner and a simple global material manipulation and organization.
Shorts such as Sintel and Tears of Still prove that beautiful things can be made with Blender, but how efficient and painful was the production? Is it possible to stay in budget with all the shortcomings? Will it be easy to train other artists to get used to the limitations…The answer is not clear.
So this is it. I like Blender, I write because I care. I want this software become better and more efficient workflow wise. I want to use it on daily bases at work…but I can’t.
So if any developer is reading this post please take it for consideration.
Your issues with blender are also my issues with blender. Blender as a whole seems to make things difficult managing large amounts of data. There ARE some handy workarounds, but hacks and workarounds are not the solution. I do have hope for the future though.
I have experiences too. I working in gamedev since 2000, at many companies.
We use Blender for pc, x360, ps3, mobile games.
I made big levels in Blender too.
Blender is a very good tool with very good UI.
I agree, it’s a great tool with a great workflow. I prefer it over maya for most things. But the outliner is under-developed, referencing is solid as a rock but difficult to work with (mainly just hard to understand whats happening behind the scenes sometimes) and the dependency graph would be better for TDs if it was more exposed and editable.
But that doesn’t mean that blender’s not a good tool. I use it everyday to get work done.
just a little note; your website gives this message out for me;
Error establishing a database connection.
since I wanted to see the works, I tried archiv.org, and the demoreel is surprisingly and gladly available on its original address even with the error message.
I too hope that the issues you mentioned are getting fixed - if Gooseberry is successful and also can makes the things true it tries to optimize, Blender might turn out to be a good option for professionals.
Thank you for the kind words. You can check out our latest animation project on wwww.telluraliens.com
Regarding our website, I dont know why it is down, we should contact our provider
I’am really looking forward seen the new releases.
While i agree to the first part i heavily disagree with the second part. It’s mainly programmers art (sorry Devs), not a really well designed UI. Too much flaws, too much inconsistencies, too long mouse ways, too cluttered, too much needs for workarounds. Which results in a slow and cumbersome workflow. Watch the first two Videos from Andrew Price and you know what i mean.
Blender is a good tool with bad UI. You can work with it, no question. It is functional, and in the end you get used to everything. But it could be much better.
First of all, thanks for sharing your observations.
I have worked on Tears of Steel, and I didn’t find the file linking system to be bad at all. Maybe you have to know Blender a little bit better to understand and use it? It sounds more as if you would try to apply Maya workflows to Blender. Of course that doesn’t work.
Some things that you describe are just done differently in Blender.
But still. Of course there are shortcomings. But it’s not as bad as you make it sound, and most of the things you mentioned have not been a problem at all in Tears of Steel. On the contrary, I think the linking system is quite robust and versatile in Blender!
I do agree about the dependency graph. But that’s on the list and is one of the development targets of Project Gooseberry.
You say you want to use Blender and want it to become better. Consider supporting gooseberry.blender.org
ALL of the things you mentioned in your post are development targets of Project Gooseberry.
A robust dependency graph, an asset manager, improved linking system, and many more.
Help us to make Gooseberry happen, and all your issues will be addressed!
The thing with the BF’s open movies is, shortcomings are tackled on the spot since the blender developers work alongside the artists during production. For example, there were absolutely no tools for camera tracking and 3D matchmoving before Tears of Steel. Now Gooseberry’s goal is to improve collaboration tools needed to handle feature-length productions like the depsgraph and asset management.
Open Source projects generally take off slowly and need awhile to gain momentum. Just look at the adoption of GNU/Linux on the desktop - it’s a great OS, but many are used to working with specific programs in a Windows environment that switching over is often an unjustified hassle. Blender is currently very popular with the indy gamedev crowd since it offers modelling and rigging tools without the need to spend 3000$+ on Maya with features that won’t ever be used. However, it takes a lot more to satisfy the needs of studios working on big animated productions. Gooseberry is meant to take Blender to that level.
I don’t agree that it’s just about learning the Blender way to do things. There are some serious inherent limitations in Blender’s library linking system as it currently stands. Specifically, you can’t tweak, modify, or animate anything that’s linked. The only exception is animating armatures via the proxy system, but that’s a quirky system with plenty of issues of its own.
For what it’s worth, we suffered a lot with the library linking system on Sintel. Cloth sim on linked characters, hair sim on linked characters, wanting to do shot-specific shape key tweaks, etc. We managed to work around all of the issues, but they were quite decidedly work-arounds/hacks and not good workflow. The “Blender way” shouldn’t be work-arounds and hacks by default.
But I do agree that the core functionality of the library linking system is very good and very stable. The foundations are there, we just need to build on them further. I’m really, really hoping that this is one of the core things addressed by Gooseberry. It’s been a long time coming.
I know that the Blender GUI is perfect for some here. I just thought i add another statement here too besides the usual praises
The mouse ways are short, as long as you use keyboard shortcuts. That is what the interface has been designed for.
So you agree with me that Blender has a bad graphical UI?
How much of the 300+ hotkeys do you really know? Hotkeys are not the cure for everything. Especially with hotkeys where you break your fingers at it. The fastest workflow is the one where the artist can decide what to use. Buttons or Hotkeys. This needs a proper graphical UI too though.
And settings is also GUI.
Blender has an amazing UI. My favourite UI’s of all 3D studio software. It’s just not perfect.
There is of course never a perfect. UI is always a tradeoff. But there’s quite a lot of room in this statement betweeen amazing and not perfect. Yes, the Blender 2.5x GUI is functional, no question. And it is definitely more user friendly than the 2.49 GUI. It is even not really that bad. It has some good approaches, and in the end you get your job done. But it has simply to much flaws to call it good.
For example the whole tools and settings are too cluttered distributed across the GUI. You have too much possible locations for them. And so you dance forward and backward between left, right, up and down of the Blender surface, followed by scrolling and tabbing to adjust settings and to activate tools in menus. This is even cumbersome when you already know where the tools are. But when you start to search a tool then you are quickly lost.
What goes wrong with the GUI has beaten to death in other threads already. With a half and half result. Half of the community agreed and wants a change here and there. Half disagreed, and says everything is fine as is. So i doubt that we will get a result when whe start this discussion here again. For me personally it doesn’t really help to praise the GUI while i am much slower than in other software for similar tasks. And this is what MonteNero discovered. That’s why he started this thread.
I’m with Cessen here. I love Blender, but using the built-in asset management tools for any kind of real production is a pain. Linking (and the convoluted way it’s handled in Blender) is one of its biggest handicaps, but one that most people will never experience/complain about because they’re not using Blender as part of a production pipeline.
I hope this will not become another “Blender has an horrible GUI”,better to speak about real issues,like the asset management and the dependency graph,these are the things to improve.
Simple, if anyone tries to make this another GUI thread from here on, the posts will be deleted as off topic. This thread is about Blender’s tools as part of a real production environment. Let’s keep on topic.
I think that outliner, linking system and dependency graph would surely need some improvements and change. I am confident that project gooseberry will work on this areas as well.
