Spacing analysis: regarding flattened apexes

Clip from Trois Petits Points, 2010, Gobelins.

(An insightful short regarding the post-WWI era, and war in general. To see the full movie, click here)

Another spacing analysis I thought I'd share.
It's a good example of of how animators can sometimes totally "break" physics, intentionally, and get great results. The lady in this shot is "popping" all over the place. The anticipates and start/finishes of the jumps are more-or-less as one might expect, but the apexes of the jumps are an interesting sort of hold: a hold vertically, yet horizontal motion is maintained.
Now this may come as a surprise when explicitly traced out, as done here. But we are looking at a pretty well developed method, this isn't something "out of the blue". And it doesn't "look" wrong when played out. It actually looks very right. So how can that be? Because in terms of physics, it most certainly isn't correct.
I'm no expert, I'd like to get an expert's take on this one, but here's my best effort at an explanation for now:
I think we're looking at a good expression of something perceptually harmonious. For starters, solid holds are good to have on important poses, the mind is given that little bit of extra time to "take it in" and we get that here. But secondarily, I also think we're seeing an emulation of how the eye tracks motion, an essentialization of that. It takes about 5 frames (or 1/5th of a second) for the eye / human consciousness to "become aware" of a motion and focus in on it (that is, begin to track it). These "5 frames" begin from an >anticipate< going into the motion, not (as you might expect) from the start of the motion "itself". (In fact, a good part of the consequential motion is comprehended on the basis of the anticipation alone. For example, we best understand how much force is represented by a given motion is not from the motion itself but from the anticipation that leads into the motion.)
Okay, so: in the above sequence, from the start of a jump's anticipation to the start of the "apex hold" you have roughly 6 frames in each case: 10-16 = 6f; 22-28 = 6f, 35-39 = 4f, 46-51 = 5f.
Now, once the eye HAS fixated on a moving object, it tracks the object, right? So >perceptually<, from that point, the object stops moving! Relativity swaps out: perceptually, it's the object that is now static, and the background is moving (but because it's the background rather than the focus, this is a secondary detail). And this is exactly what is being expressed with this particular method! Then, at some point, the eye presumably anticipates an upcoming contact moment and stops tracking the object (reverting over to a tracking of the ground instead, you might say); the object then impacts, and if there is another anticipate, it should be that the process continues. So what we have is an integration of anticipation-driven perception and eye tracking into the animation technique, all working toward snappy and appealing results. On one hand it isn't accurate; yet on the other hand it is >more< accurate than photographic / naturalistic / physics-correct spacing, because it takes into account the workings of our human minds and the workings of our eyes. This is akin, as always, to a similar continuum in the static arts: when comparing photorealism / photographic images to good quality abstractions / drawings: the former is "techincally speaking" more accurate, but the latter is "humanly speaking" actually the more accurate of the two: the latter having taken into account all things human, as digested by the mind and outputted onto the paper / artistic space: whether it be thought influenced by fantasy, myth, dreams; perceptual adjustments of every sort, enhancements of focus, enhancements of essential concepts, or even our "simple" human weaknesses (style is the ultimate refinement of our artistic shortcomings) etc. Drawings are to photos as cartoonish animation is to video.
A quick disclaimer as with other musings I must point out that it's "just a theory". I'm not basing this on anything especially authoritative. However, I do think this would hold up to further scrutiny. I'd be interested in learning more.
Cartoon timing naturally works very well for cartoons. But it isn't quite a cut-and-paste, what is gleaned from 2d work does not necessarily translate into 3d. Usually, the 3d side needs to be a bit more naturalistic. The level of cartoony motion that can be tolerated within the 3d realm also depends very much on the level of abstraction already being applied, visually. If the rigs in use are more realistic, the motion in use is going to have to also be more naturalistic. If the rigs are more abstracted, a greater level of abstraction in motion is possible. This particular example is highly abstract, visually; the manner of animation, accordingly, has greater freedom into extremes of abstraction.
I have already been pushing somewhat in this direction with my own work. The path of a jump often looks more pleasing if it has a longer hang time / a flatter apex, and an especially fast "in" (the motion just before contact). A parabola is the "technically speaking" correct path for a "thrown" object (or leaping character, one and the same), yet it is also, arguably, the least pleasing path to use, perceptually. Strange but true.

2011 07 20 addendum:
During Q&A at AM today, I had the opportunity to ask Pete, my current Class 3 mentor, about the technique behind this particular clip. He had some interesting things to say about the matter, along with some project-specific stylistic comparisons. I am not at liberty to go into details with regards to class discussions. A main key point he did make, though, was regarding the establishing and directing of audience focus. A cartoony style can better optimize focus. Whichever style one finds oneself working in, special care does need to be taken to ensure it is applied consistently across the scope of a project. The fictional world portrayed needs to be internally consistent, even though its laws may differ from our own.


Spacing analysis: a simple step-to-the-side

Thought I'd share a bit of spacing analysis I did today, a clip from Disney's Song Of The South, Brer Rabbit doing a simple step-to-the-side. Plotting the path of the eyes. (I devised a Photoshop action that automates compiled images from short sequences). I am seeking a more in-depth understanding of timing and spacing. This example has a fascinating abrupt change of direction, what we would refer to as a (sharply) broken tangent in 3D. It's an interesting case. The change of direction is instananeous - as far as I can see (if not, the ease-in would still only be about 1 frame). It's a good example of snappy timing. Would you have realized there was such an "abstract" path of action just from watching the clip? The result has appeal. That appeal is grounded in the good timing-spacing in particular, from among all the Principles.


Stewie Tennis Camp

I'm now into week 2 of Class 3 here at Animation Mentor, "Advanced Body Mechanics."
One option for Class 3 is Anim Jam, where we compose a set of 3 shots into some form of meaningful sequence. The emphasis must be on physicality, but a degree of storytelling and characterization can be embarked upon provided it is in appropriate measure.

I'm very happy to rise to the challenge. After much deliberation I worked out a sequence named "Stewie Tennis Camp." I thought I'd share the summary composition I did of this today. Perhaps I should have gotten this done last week, as did many of the "really organized" students, but there were a number of technical challenges I had to work through first. I did present the plan last week, but I hadn't yet gotten to "modding" the characters.

The mod process took longer than expected... as does just about everything and anything in the realms of creative effort (and life in general). I managed to break Kid's rig when setting up the constraints for the tennis racket, so I had to go back a few versions and sort that out. There were the usual frustrations with Maya (I'm "born and bred" in XSI, so far as 3D work is concerned. Maya still feels clunky, but I'm getting there).

I'm hoping to run with a background that is simply a digitally painted image I prepared last week. I can do fancy modeling texturing etc., no need to be getting tied up in that at the moment: I would like to maintain a reasonably efficient focus on animation work with minimal distractions; and this is a core learning philosophy at AM to begin with. We're here for the animation first and foremost.
I'm sure anyone who has observed my work will see, however, that I prefer a rough edge to a polished one; chaos to order; texturality and a degree of complexity. I like the look of flat artwork like this in 3D work, it's a decent style.

For the first shot, I'm working on in this sequence, Shot 2, I also ventured into the new territory of doing a pencil test. At the start of the year, I got a copy of Digicel Flipbook Lite for pencil tests ($78, and there was an AM student discount, very affordable). I hadn't really gotten around to doing pencil tests so far, however. So I did this one to help me get rolling on this shot.
One thing I'm hoping to do is to break away from naturalistic motion and posing, and to push the artistry and expression. I've always had a tendency to remain comfortably within the naturalistic realm. But hopefully I will find some stylistic distinctiveness as I go. This particular test is still naturalistic. You might think it was rotoscoped, for example, but it wasn't. A good study, but still not quite where I'd like it to be.
It's a good question: is there any point doing pencil tests like this?
It's definitely more work than simply going right at it, in blocking.
I expect it could be fruitful in finding the most essential aspects of timing and poses. It could be a good way to take a fresh look at something, free from the technicality bog that working within Maya can be. But sooner or later those poses have to be built in Maya, and there aren't really shortcuts for that in this time and place.
The week is "getting on" and I do have to block this whole Shot 2 thing out by Sunday, so I definitely am going to get on with it all directly in Maya now all the same.
I'm working on the shot in a sandwiched way: this above test involves the first bit, I have a flourish for the last bit, and then, depending on how long each of these two sets of moves takes to accomplish well, I'll "simply" fill in the inbetween bit with some fast volleying.
A note on the flourish: Pro is a girl, and I've been looking for just the right sort of "in your face" move. That has been a bit of a challenge. I spent about a day simply doing that: looking for the right move for her. Youtube is very useful. But my own acting it out was half-decent too, oddly enough. When I put ref together, I tend to have a tightly cut set of concise actions strung together, like a series of dance moves. Perhaps in time it will all simply flow straight from the imagination, because I can visualize it fairly well all the same. BUT I do make a point of looking around. There's something to be said about "authoritative motion". I learned that with my Riverdance sequence last class. Certain types of motions are highly polished when carried out by those who have the skill. In these cases, no amount of effort, in terms of generating one's own reference video, will ever match the quality of these authoritative sources. It's good to look for such sources, and to then study what comes on good authority.


Class 2 Progress Reel

I've finished my animation revisions and am finally ready to share my Class 2 Progress Reel.
Hopefully I will also manage to give it a half-decent rendering treatment as well at some not-too-future point.