AM Class 5, Shot 2, latest iteration

Most of the changes this week were made to the second shot in the sequence.
Assuming I don't make any major changes, I've now resolved the most difficult issues, and the rest of  the polish process is a more relaxed matter of working out irregularities and adding details.
I would like Nono to interact a bit more meaingfully with Lillian, though, in terms of her own reactions and focus.
Just starting speech animation here, jaw rotations only so far.

AM Class 5 Shot 2, coming along

This last week's submission. The first and last shots are now becoming well developed and body mechanics has been generally resolved; Charles, my mentor, has indicated that I must now focus on performance. I've already re-filmed video reference for various details.
We've only got two weeks left for this assignment! Time flies. I must also get the facial animation rolling this week. 
I'm really happy with how Nono's run-away bit has turned out. That quality level needs to be brought across the board now.
A challenge with this shot is that the motion is naturally somewhat soft to begin with. I need to tighten it up, especially in the middle. A "soft section" is a section that's wide open for subtler performance and polish work, so this region is one big blank canvas with lots of room for exploration.
Hopefully I'll even get a bit of that "jiggle settle" and "residual energy" cited by Andrew Gordan in a recent lecture on polishing our shots. These descriptions get at the more complex nature of a good settle, something more sophisticated than simply overshooting the action.
And of course, there's still a lot of the emotive and expressive content to be put in.

Working in XSI

Just thought I'd put up some screengrabs of my Softimage | XSI workspace as I animate.
I've got years' worth of customizations in here (mostly written in JScript), a very familiar working environment. One thing I struggle with in Maya is dealing with what seems to be a more limited degree of freedom in terms of workspace customization. It's alright in general, but I struggle in particular with the tiny space allocated for custom commands. As you can see, I occupy an entire computer monitor with custom toolbars within XSI (these are collapsible as well, so I can use the monitor for reference etc. as needed). I also do very much enjoy working in 3D, something I've done for a long time via a custom 3D camera setup that I usually use with all of my cameras. This setup uses the Interest null/locator of the camera to define the focal point for the 3D, and maintains a constant (adjustable) angular displacement based on a custom trigonometric expression. The 3D is simply accomplished via crossing my eyes and focusing. Low tech, and one of the the oldest methods of achieving the effect; but of course it being a volumic environment it is effectively as "true 3D" as any hologram, perceptually. No expensive goggles or screens needed; with just a bit of tinkering anyone can do it with their existing setup.

Two run cycles

Two run cycles. Studies toward my current shot's ending. Which cycle is better? 03 or 04?
It's an interesting comparison.  The cyclic motion of the first is given a node on the hips; the cyclic motion of the second is given a node on the head. It feels as if the first is more typical of a starting acceleration, and the second is more typical of a high-speed run. Perhaps because of this, the second "feels" faster, even though both are traveling at the same speed.

I was happy to find that Nono could take this kind of motion without looking too odd in her outfit. Rigging that dress was a study in itself.

AM Class 5 Shot 2, early rough stages

So here's my Class 5 Shot 2 in its early rough stages.
I'm really eager to move past this point. However, I do want to ensure I cover enough ground at planning stages to enable decent results. This clip shows video reference with drawover planning for keyframes and breakdowns. The drawover work permits me to think it out a bit more carefully before I actually launch into the posing process; aiming to put thought before actions.
After the basic poses and blocking ideas are in place, segments can also be given additional layers of detail sourced from further processes of re-filmed ref or analysis of other sources. And then, around that point, it also takes on a life of its own and most refinement continues via reflecting directly on the scene, and putting in changes based on feedback.
The blocking here is still very rough, lots left to be done before blocking is complete. I'm so glad we have a few more weeks for working on this assignment!

Connect Four Breakdown Analysis, a set of solutions

So… here you have it: my main set of answers to this “connect four” challenge:

 

Here are six solutions to this challenge:

Circle evenly timed, Circle favoring alternate points,

Square evenly timed, Square favoring alternate points,

Star bounce, Star bounce cartoon edit

Many other solutions are also possible, though there probably aren’t very many more that are appealing. (e.g. In terms of making a complete study of transitions, anticipate and overshoot would need to be represented. These would be “somewhat floral”, with a small loop at each knot) For now (shortly), I’ll be focusing on these six timings at greater depth.

Solution: Stepped

 

 

Before getting into the above breakdown situations, let’s take a look at the blocking for this cycle (in stepped keys).

The keys are distributed evenly in time.

So for 24fps, this puts key poses on 1, 7, 13, 19, 25

(or 0, 6, 12, 18, 24, it makes no difference, but I went with odds here)

First and last frames are identical, playblast less one frame for a loop.

 

A note to clarify what I’m showing: the upper left panel shows a drawing of the path of action, the upper right panel shows the result, with ghosting, in 3D, and the lower panel shows the animation curves that produce these results in the graph editor / animation editor.

As we have just seen, in blocking the four key poses are always the same, and are also always spaced and timed the same.

So this challenge has only one blocking solution, and this blocking can lead into ANY of the above timings.

(Lesson: blocking in stepped keys can be interpreted into many final timing solutions. Blocking is a simple representation of many timing possibilities.)

But… “Animation: it’s all in the timing and the spacing” (said Grim Natwick, according to Richard Williams)

To a beginner, the above blocking might simply be converted directly into splines and that would be that. This is exactly what leads to that typical “beginner’s look” with animation, also known as drifty, floaty, or underwater feel.

Next CFBA posting: getting into the splines.

The Connect Four Challenge, an Introduction

Here is a trail of animation analysis I’ve been exploring, that I’d like to share.

I intend to post it as a series of thoughts, which lead into some potentially useful insights into the basics of timing and spacing, and patterns of keys in the graph editor, accordingly.This posting is the first in the series.
 

The "connect four" challenge

The basic challenge is as follows:

Take a set of four dots in the same pattern as seen on a dice.

Consider each dot to be a key pose in a second-long animated loop.

Connect the dots with keys that are evenly distributed both in timing and spacing.

Simple, right? After all, both the timing and the spacing are clearly defined. So if you key those poses at the right times and places, you get your answer and you’re done, right?

Well, maybe. For one solution. But there are many solutions.

So the challenge is this:

Consider a variety of solutions and implementations, making an effort to resolve each solution with just those four keys in the graph editor.

(To limit obvious alternates, all solutions assume a same start/finish point, a clockwise path direction, and the use of the orthographic XY plane)

It’s an intentionally simple scenario, yet oddly enough there are still many answers to the question. Initially, I meant only to examine the answers in 2d form. But when I applied the answers to 3D effort, I realized that there was more to the matter yet again, because for each answer there were also many ways of implementing that answer in the context of rigging and animation curves (Lesson: something which seems simple will typically become complicated in ways that were not anticipated.)

When I ran the challenge through some of the most obvious answers and implementations, I was able to confirm a variety of animation rules of thumb. It was also a generally decent technical study good for solidifying some basic concepts in 3D animation. 

... more to come later! Feel free to comment if you'd like to have a go at the challenge for yourself (before I give away my own set of answers!)