I know you’re very into modeling right now, and that’s good for now, but one thing that I’d suggest you do fairly soon is to determine the proportions of the room (vs. that of the character), and then, within that room, put together an “actual running-time animatic” using preview renders and, as necessary, simple geometric shapes that are of the right size.
Also, if it’s going to be to-music, put some kind of music in there, especially with regards to the rhythm and tempo of it. Even if it’s a recording of you saying “boop, boop, boop,” or a click-track.
By “actual running-time animatic,” I mean that you actually build a set, to scale, and you put actual objects into it (including your perhaps-incomplete and/or un-rigged character), and you have the characters making those moves. (The picture of the girl, for instance, might be nothing more than a box … but, a box of the correct size. Likewise, the bed is just a series of boxes, but the mattress is the right size, the right distance off the floor, and the bounding-box of the headboard is correct. You know exactly how many feet/meters corresponds to one Blender Unit.
You put cameras in, and, with OpenGL Preview rendering, “shoot some film.” Shoot lots of it. Preview renders can be completed very fast, such that you can easily afford to say, “well, what if we do it this way?” You can easily afford to shoot more coverage of a particular move, before taking the strip into the editor, and decide from there how you want to “cut” it all together. Just like is done with real film.
Then, go into VSE or something else and actually edit it down to what could actually be the “final” cut. (I happen to use Final Cut Pro, but you can also use iMovie or something else. Or, of course, VSE.)
Then, you start doing take-offs from that, working to replace each clip with the real thing, using the camera placements and so-on, linked directly from the rough shots.
I do this, basically, because “I can’t draw worth a damn.” :mad: But, when you do it this way, it’s a real eye-opener as to the timing of the scene, the camera-angles you might actually use or not-use, what’s going to be important (because you can see it) versus what isn’t (because you won’t see it), and of course, exactly what strips you need to render “for real.” Also, it makes the whole thing “feel alive, real,” much more quickly, which is a boatload of good encouragement. Your focus is put onto the show, very early on, and so it feels like you’re watching the show, even though most of the pieces are stand-ins. If the pace and timing’s off, as it usually is, you can tighten things up here and relax them there.
Use object-linking, too. Start with that, from the start. One by one, those boxes and cones become real assets of the same outward dimensions.
Rig up the set, put cameras in it, name those cameras, and include file-name, camera-name, and timestamp in each shot. (If you shoot with one camera and decide to move it, duplicate the camera, instead. So what if you have a bunch of “extra” cameras lying about.) Later on, you can link-to those exact cameras, reference exact object dimensions and coordinates and so on, so that the completed footage exactly matches the animatic … which is, of course, much more than an animatic now. The coordinates, angles, camera f-stops and so on … are real.
The frustrations of trying to sandwich “completed” shots into a coherent form late in the game, when you have so much time invested, are reduced by moving them early in the game, when you have options and real creative choice. You find that there’s a lot of “slop” and that, by cutting a second here or two seconds there, suddenly it becomes a lot more “tight.” And, you’re doing it when you can easily do it. The preview renders, lower-res though they might be, will exactly match the high-res replacements. (OpenGL’s capabilities these days have become quite stunning.)
This workflow has been described: “Edit, then shoot.”