Hi
Even on YouTube, there is not a lot of quality Blender 3D animation videos. Why is it that the Blender users are not making that transition from still art onto 3D animation?
I also wanted to share this piece of art which shows the entire process of making a 3D animation video in a funny and cartoonish kind of way. Sorry if you have seen it before, I just stumbled upon it today.
It’s just that animation is way more time consuming than still image.
It’s not specific to blender…
Animation is hard, probably the hardest of the 3D disciplines. It’s much easier to make one frame look good rather than make thousands look good.
Paradoxically this is a still image explanation instead of a video tutorial…
There are people who are interested in making still images arts, just like photos or paintings. Also animation takes time to learn and a lot of practice, it’s not easy at all.
By the way, this old thread can be complementary:
Cycles is slow.
Everything is slow in animation.
From concept sketches, story boarding to final rendering - story must become alive - without considering changes, corrections, bugs…
It’s whole lot easier to shoot a movie, but even those are hard to make.
Possibly because animation as a discipline is an incredibly challenging art form. It takes years of study and practice of multiple disciplines and it never ends. Most people who have spent years practicing and sharpening their animation skills don’t have a great desire to learn a new software package. It just comes down to the pencil you are most comfortable with.
While there is some great animation done in blender, It sure feels like not a lot of animators use blender… but I feel that will change. I saw lots of talented animators pick up blender and do amazing work. The more great animation is done, the more animators will come over to try blender out.
The number of animations done is likely to go up with 2.8’s beta and release.
Toy Story
The reason why more animation will result is because various tasks will become easier and less time consuming. The rendering part alone is why some are now making their animations in game engines such as Unity and Unreal. Blender 2.8 promises to democratize animation to an even greater extent because the realtime tech. is right there in the DCC package.
So, Maya for example, has had all of this before Blender and is Maya also easier to learn than Blender?
I’d agree that this is probably just a common 3D art thing in general. Animation is a whole new set of skills on top of the existing 3D knowledge. First there’s riggin, which is a whole skill in itself. Then there’s the animation part, which is yet another complete skill in itself. So take the time it took to learn modeling, texturing and rendering and roughly apply the same time to learning rigging and animation if there’s no prior knowledge.
And that’s true for any 3D software, I think.
I wish I could do all of this 
I think not too many people animate in Blender because they are trying to eat the whole elephant in one go. With zero experience in animating most people immediately jump to trying to make the next greatest animated short. It’s very rare to see someone start from a bouncing ball, ball with a tail, e.t.c.
Blenders rigging and animation system is in some areas unnecessarily difficult to use and can feel dated.
FK/IK bone systems are over complicated in Blender while other programs can make it in one go.
so many crappy tutorials floating around which make everything even more complicated.
If you looking for high quality rigging tutorials for blender look for danpro channel on YouTube
He is the best
Can you become more specific about depsgraph?
What kind of rigs ?that it’s not possible on 2.79?
And what feature brings overrides?
For one thing, the lack of overrides in Blender has long been considered a crippling limitation.
To know just how crippling, Blender 2.79 can’t easily do the alien scene from Toy Story (which is a 20 year-old movie at this point). Nearly every movie in existence making heavy use of CG benefits from overrides, because overrides allows you to instance objects and give them unique poses, materials, and other properties when needed. In 2.79, animations have to use workarounds using Python to emulate the feature in a more clutzy manner.
I believe Maya may have even had overrides in version 1, so yes, a few of the new things in Blender 2.8 are in fact rather old features in commercial apps.
It is more or less the same with the depsgraph overhaul too, Blender’s old depsgraph was a rigid system that limited just how relations and constraints can work in a rig, a good example being how you couldn’t have IK and FK on the same bone chain. The Code Blog also had an example a couple years back of some concept known as “soft IK” which is also only possible in 2.8.
Blender can do this even. Using Blender’s own game engine. I guess you only need to know how to use it to its full potential.
If I was going to do video I would just use UE4 right now, though 2.8 and Eevee should help there. As for cycles, yeah I don’t have a render farm, or a large enough pipe from the power company for one.
With still images I don’t have to worry as much about cycles render times, and still images have the benefit of being easier to compose with geometric trickery and post processing that is just not practical between frames or with a moving point of view. The end result is what matters.
Another note on animation: I find Blender’s physics sims and tools to be slower and far more difficult to deal with than screwing around with fbx and external tools to accomplish the same thing. For example I get much better results faster if I go through the trouble to export and do a rigid body sim in a real-time game engine. Especially with compound collision primitives (any combination of: convex mesh, box, sphere, cylinder, capsule, etc.) that make up a complex shape. Heck I see decade old, more complex rigid body scenes run in realtime on Garry’s Mod. To this day I still don’t know how to do compound collision primitives in Blender. There has to be a proper way to do it. Most tutorials just use mesh triangle collision which fails in a rigid body scene at the slightest complexity beyond a simple demo. Other tutorials use constraints that tend to appear lagging behind frames of the sim, and I’ve seen mention the BGE has compound primitives but not sure how those are composed. So eh, I’ll stick with external tools for complex physics sims for now. I have been able to get really good results with Blender’s soft-bodies, but oh my… slow…
For animation cycles is very fast especially with the denoising and the CPU+gpu rendering it’s game changing
Well it’s pretty simple. Animation production is laborous. Most people f*ing hate labour.
And even if they don’t, it’s another question weather they find the time for animation production anyway.
Personally, I don’t think it’s about rendertimes at all, really.
greetings, Kologe