I’m new to animation, but I’d like to know in advance how far can a small team go with Blender in regards to a full featured animation movie. I know of Sintel and Cosmos Laundromat, but these are shorts.
First, my thinking for the workflow. I individually create and rig the characters and props, in separate files. I then import them into a scene with a few cameras, creating the world, possibly with a terrain of about 2-3 kms. I start animating cameras, characters, props, and particle effects. I render, and continue with a video editor if necessary (I know Blender has its own module for that).
My questions. On a decent PC (4GHz quadcore i7, 16GB RAM, GTX 980), can Blender handle a detailed world this big at once? I mean, can I have the whole world loaded, and move the cameras freely? Should I divide the movie in scenes for better performance and, if so, is there a scene manager of some sorts for that? Any recent tutorials you’d recommend on the subject?
Home movies are shot ‘seat of the pants’. Everyone knows how horrid those are. You don’t proceed with a ‘full featured animation’ by building a world and stocking it with props and actors. You proceed by writing a story, developing a storyboard, making a shot list, and then modeling what you will see in the camera in each shot. If the clerk behind the counter is only going to be seen waist up, why model his legs and feet? Wasted time. And animation is one place where you cannot afford wasted time.
Most of the techniques of animation, in fact, are developed around avoiding wasted time, since animators time is expensive, and planning is cheap. Now, you may not be paying anyone to do your animation, but you are spending your time, and you don’t want to waste that, either.
Animating with Blender: Creating Short Animations from Start to Finish is a pretty good guide to making an animation (the only real difference between a short and a feature length animation is the length, not the methods). There are some tutorials out there as well, on YouTube and elsewhere, but I don’t vouch for their quality.
I agree with you. There are exceptions as always, but generally that’s true.
I think that’s obvious, I was asking about the technical aspects. Good to see we’re on the same level, though.
I wasn’t clear enough. The camera will eventually scan the whole terrain (more or less). So why would I want to model the scenes piece-by-piece instead of building the world all at once? How possible is Blender having performance and stability issues?
Again, I don’t think is wasted time if he’s going to reveal his legs in a future scene.
Blender as the software is not the limiting factor in achieving what you want.
Time - super computer = fast render, shit computer = more time needed
Attitude - maybe you do not have the level of commitment required to make such a project and you give up after a couple of months because you get bored
Skills - maybe you and whatever team you have do not have the necessary skills to achieve your vision
Story - Take at the very minimum the first 12 months getting this right.
Planning - Take at the very minimum the next 12 months getting this right
Then you could consider starting to model something
Again, I don’t think is wasted time if he’s going to reveal his legs in a future scene.
Why have the same model for both scenes. This comes into the planning phase. Maybe the desk scene is in closeup and the walking scene is in mid/longshot. Level of detail required in each is different so use a high detail model (without legs) and a lower detail full body model for walking where a high detail model would waste render time. Maybe make the planning 18 - 24 months
That’s what I wanted to know, if I can rely on Blender when processing large amounts of data. Regardless of the approach, and assuming the PC can handle it.
Naturally, these are the basis and foundation, although times may vary considerably according to the level of charisma and expertise on a particular field.
I see what you mean, but don’t you think making and using variations of a model will add up time as well? Why not try something in the middle, model everything in the absolute minimum quality desired, then refine any characters and props according to their closest frame in the storyboard. No low resolution duplicates, yet no fully detailed models on everything, either. It may cost some rendering time in the end, but I think it’s more convenient when, for example, something comes up and you have to change the location or distance of the camera. The more prepared you are from the early stages of preproduction, the less frequent this will occur, but these things happen. This of course means that, if the guy behind the desk stays put the whole time, he doesn’t need legs or anything.
It’s always nice to see how other people think and work their way out.
