Anatomy of a Shot

Amidst a long-overdue vacation to the east coast, some freelance work, and an exhausting trek down to Comicon, I’ve managed to do a little personal animation. My Animation Mentor students usually ask about my workflow and methods, which I have a hard time explaining, so I thought I’d show the result, and then a progressive series of earlier versions. Here’s the preliminary version, sans sound-track tweaks, lighting, and texturing I plan to add:

[ Javascript required to view QuickTime movie, please turn it on and refresh this page ]

The rig is Boris from Jason Ryan, and is a pleasure to work with. The dialog is from a Twilight Zone episode. I’d been trying to think of some suitable dialog for Boris, and one evening I was flipping channels, heard this Twilight Zone robot talking, and thought “That’s it!” Fortunately the disc in the dvd recorder wasn’t full (I compulsively record films from Turner Classic Movies), and a millisecond after I pressed “record” this line was delivered. Serendipity!

The next step was planning the shot. I didn’t want this to look too human, so I immediately decided to not shoot video reference. There are times when I do thumbnails, and times when I jot notes. In this case I went the route of illegible scribbles, along the lines of “slumped; jerks to life in 2 stages; hand to ear, “drill” out antenna – 2 turns on “maturing”; as antenna pops up, head jerks and eyes flutter open, head jerks back. Then body jerks back, lifts foot, awkward hold, throws self forward (hit on “soon”), cont. fwd, 2nd + 3rd lumbering steps, turn to cam. on “faculties.”

So I had a fairly clear scenario in mind right away. The tricky part would be to portray the robot being awkward and glitchy initially, but not have those glitches look like animation mistakes. And I wanted the robot gaining competence as the scene progresses. The scene really isn’t long enough to really show that progression, but it was something I kept in mind as I worked. I also wanted the robot to turn itself on by using its hand like a screwdriver (the “drilling out” part above) to raise it’s antenna and really awaken itself.

As soon as I began blocking the shot, I decided to reverse the major order of events in my scribbled scenario. I thought it’d be more interesting to break the shot in two, and start with a full shot to establish the character and focus on full-body animation, then cut to a medium close-up and play with some acting. This reversal also lets the raising of the antenna become the climax of the piece. In the original conception, this interesting bit of action happens right away, and the last part of the scene is the character taking some steps.

As soon as I started playing with poses, I realized it would be fun to have the character stepping over something. It occurred to me that Boris could be stepping out of his own shipping crate. (Before this goes on my reel I’ll get the box textured so that’s immediately obvious — right now it kind of looks like a coffin!)

[ Javascript required to view QuickTime movie, please turn it on and refresh this page ]

Here’s an early playblast. I visualized the basic action, estimated the timing of the steps in my head, and blocked in the contact poses and the major breakdowns. My initial focus is on the major body parts (hips and feet especially). I work in spline mode from the start, and shift these keys and breakdowns around until the overall timing is working. The head and hands, if they’re animated at all, are “place holder” poses, though here I’ve already roughly posed the hands themselves in different poses for some contrast.

In the above playblast, I’ve done several passes on the hips, working especially on weight shifts and trying to build a framework for the next passes that will focus on the chest and on pushing the steps themselves. Getting the timing worked out on the hips is crucial before I spend a lot of time refining the leg/foot movement. Thus far I’m keeping the keys the same frames, to make timing changes easier.

[ Javascript required to view QuickTime movie, please turn it on and refresh this page ]

Sometimes I’ll ruff in the entire shot, and make successive complete passes all the way through. Because this is a fairly long shot that I know I’ll break into two pieces, I continued to focus on just the first 90-100 frames. Here I’ve added the packing crate (thanks Heather!) and done multiple passes on the hips, then the feet, then the knees, then the chest, then the head, then the arms.

I do each of those as a separate pass, though I will also tweak things from earlier passes as I see problems. At this stage I’m not focused on arcs and spacing, just on getting a good combination of movement and poses. I’m also not thinking about keys. I’ve long ago stopped worrying about keeping all the keys on the same frames, and I’m working in behavioral phrases.

[ Javascript required to view QuickTime movie, please turn it on and refresh this page ]

Now I’ve broken the shot up (actually, I just changed the camera — it’s all still one continuous shot) and roughed out the second portion. It’s fairly static — I mainly wanted to see if the timing of the ‘awakening’ and the reaction to the raising of the antenna was going to work.

I’ve also done a pass to “dirty up” the overall movement. I noticed that the soundtrack has a repeating mechanical noise about every 12 frames, and I noticed that the vocal accents happen to fall almost exactly on those same beats. So I layered in small, sharp “bumps” four frames before those beats, then “recoil bumps” two frames later in the chest. Later I’ll add subtle reactions right on the accent in the head and hands, but I’ll wait until those parts of the body are closer to final.

I’ve intentionally ignored face/lip sync/eyes and subtle head movement. I find if I focus on those areas too early, I tend to leave other parts of the body underdone. I want the scene to work with the dialog just from the body movement. Then the other stuff is icing on the cake.

[ Javascript required to view QuickTime movie, please turn it on and refresh this page ]

Now I’ve done a pass on the head and the jaw. No mouth shapes, just the jaw opening and closing. I’ve also begun developing the last 20% or so of the shot, and I’m experimenting with having “squinty eyes” for most of the shot (to contrast with the “live” eyes that will happen at the end).

[ Javascript required to view QuickTime movie, please turn it on and refresh this page ]

The above playblast is the result of several passes of pushing the body action in the last third of the scene, two passes on eye movements (I quickly decided the squinty eyes made it harder to connect with the character), a pass on the brows (which I decided to keep subtle and flat until the antenna is raised), and a first pass on mouth shapes. I also decided to change Boris’ initial pose so that he’s looking towards the camera. Initially I wanted it to feel like we’d caught Boris just rising up out of his crate, which I still think is a decent idea. But the result was that he spends most of the first 100 frames looking down.

Finally, at this point I’d decided to lengthen the scene by about 16 frames. He needs a moment to recover from his ‘awakening,’ and then I wanted a goofy half smile to register at the end. The problem was that there was more dialog at that point in the soundtrack, so I used the sound-editing freeware Audacity to clip away the unwanted dialog, and copied and pasted some ambient sound to the end. It changed the “beat” of that ambient noise slightly, but I don’t think most people will notice. (I’m still planning on using Audacity to tweak the pitch of the dialog during the awakening, on the vowel accents of “my” and the first syllable of “faculties” to better fit the antenna raising action.)

I sent the above version to an animator friend I trust (thanks Sean!), and got back the major note that the “awakening” section wasn’t reading, and because of that the screen-right hand lifting up and juddering was confusing. A couple of small lip sync issues were also noted (the “ing” of maturing, the “ly” of gradually, and the “w” of “will” all needed pumping up) and the screen left hand looked stiff at the head of the shot (which was actually intentional, but if it generates a note, it probably needs to be changed).

Addressing those notes resulted in the version at the top of this post. The big change was amping up the “awakening” section, especially in the chest and head (which is what I get for jumping into doing the facial and lipsync before I was totally happy with the body animation in this portion).  I’ll set this aside for a few days, make those tweaks to the sound, add some texturing and lighting, and take another look. I have a feeling I’ve strayed too far from the initial ‘robotic’ feel I was going for, and I’ll likely tighten things up then.

17 Responses to “Anatomy of a Shot”

  1. asterias Says:

    A great great post for newbies like me :)

  2. toyBunny Says:

    Very nice! I really like the mechanical stuttering of his left hand and the overall personality of the character. How much time would you say you’ve spent on each phase, or overall? I’m just curious. It gives me a timeframe to shoot for for my own stuff. Can’t wait to see what else you do with this. I’ll be watching.

  3. DanO Says:

    great stuff. watching it a few times and i’m almost yearning that he was a skeleton stepping out of his coffin….

  4. Paul N Says:

    Nice workflow breakdown. An important point that is only touched on briefly is that the audio used is more-or-less obscure. I’m a big fan of using audio from less recognizable sources. I feel the viewer is able to focus on the animation more when they don’t have the scene from Pirates of the Carribean, Knocked Up or whereever playing in their head at the same time. It also doesn’t invite comparison with the original.

    I’ve seen people use clips from Pirates, Casablanca, and all manner of recognizable movie audio clips, and it’s always a distraction.

  5. Dhar Says:

    Great stuff as always, Kevin. A question about anticipation at the start of the clip. I see how you moved the hips screen right first before moving to screen left to shift the weight. I wonder, do you have a rule of thumb as to how many frames such anticipations should be? Looks to me that 4-8 frames to be almost a standard. I’m still at the bouncing ball exercises in AM Class 1 and the squash anticipation before the initial jump also seems to be about that same number of frames. I usually go by “feel” but is my observation about the number of anticipation farmes correct?

  6. Kevin Says:

    Thanks for the comments! ToyBunny, I put about 30 hours total into the shot. It’s hard for me to put a time on each phase.

    Paul, that advice about using obscure audio is something I always harp on when I teach Animation Mentor classes 3 and 4 (those are the classes with dialog shots). I’ve seen some great tests with well known actors/scenes, but I think it’s an uphill battle to make those shots your own (the most popular dialog choice lately seems to be from The Office, which I think is a really tough choice because those characters are so vivid).

    Dhar, I think 4-8 frames is about right. The key is to not follow a formula, and just make sure if feels right. Also, decide how subtle it should be. In this case, I wanted the movement to be a little exaggerated, to give a feel for the robot’s awkwardness.

    Right after I posted this I decided to do a major tweak on the audio track, splicing in more time in the middle, tweaking the pitch of his voice in key places, and adding a bit more time at the end. I’m giving the whole piece a good going-over. Once I’m happy with the changes I’ll post the revised version.

  7. Nassos Y. Says:

    great post, really helpfull!

    I have some feedback on the clip
    (if you are interested).

    The eye motion when he gets out of
    the box makes him look a little sly,
    (like he is hiding something) and gives
    me a sense of fluidity of his thought process
    that is opposite to what his words are trying communicate.
    You might wanna try not animating the eyes and
    leaving them open (stiff) until the last half of the clip,
    maybe when he drills the antenna.
    And instead use the neck to make him look at different

    Also the timing of the antenna is a little too late.
    I would suggest to move it in the middle of the clip,
    (or even the end of the first half) so that you have more time
    to show the “coming to life” part of the character,
    (which is usually the more interesting)
    when he starts to realize things about his body
    and his surroundings.
    Having said that, the “hand flexing” should
    go after “antenna” (which is the pivotal point of the clip)

    Also when he tries to unjam the screwdriver ,
    you might wanna pull the hand in the opposite direction
    of the head, to make the effort “read” better.

    Well, that’s it.

  8. Kevin Says:

    Hey, Nassos, thanks for the crit. I think it’s a good note that the eye movements at the beginning might be a little fluid — as I’ve indicated, there’s a bunch of elements I plan to stiffen up a little. The order of events is going to stay the same — the idea has always been that the unscrewing of the antenna is the climax of the scene. Hopefully the next version will get across what I’m going for and will feel more ‘right’ to you.

  9. Nassos Y. Says:

    It seems that i got the wrong idea of what you were trying to get across. Sorry for the misguided tips. Hope i didn’t mess-up your creative flow or anything.
    Looking forward to your new version. :-)

  10. Kevin Says:

    No worries. I appreciate the feedback.

  11. Alej Garcia Says:

    I’m very hesitant to comment since I’m not an animator but here’s my two cents: He seems out of balance around frame 60. Now it’s true that he’s not in a static pose nevertheless on that frame almost all of his body is to the (screen)right of his foot on the ground. That pose would be in static balance only if his raised foot were massive compared to his head, torso, and hands. But the timing of the motion doesn’t convey to me that his feet are particularly heavy. Because the raised leg is in motion we actually have to consider the dynamic balance, which you can think of as a centrifugal force acting to the (screen)left due to the swinging motion of his raised foot. Again, the timing doesn’t feel to me as if this would be enough of an effect to compensate for the center of gravity appearing to be well to the right of the foot on the ground. Finally, as the raised foot starts to come down the left arm initially rises (frames 60-62) and moves further away from the center, which would further unbalance the robot. My suggestion would be to tweak the positions of the arms and hands and/or modify the timing a bit. But, as I said, I’m not an animator so take this with a grain of salt.

  12. Steve Says:

    I love this shot, especially the way he moves when he’s getting out of the crate. He looks perfectly balanced to me, like he gets just on the verge of being off balance, then pitches over the other way. I wouldn’t change a thing there. I can’t wait to see what else you do with the shot.

  13. johnchuang Says:

    Hi,Kevin! It’s a great post! Thank you so much! I’ve learn a lot from your workflow. Also, I translate this post into chinese and post it her: , I think some Chinese animator would appreciate your such great stuff. If you don’t like the reposting ,plz contact me , and I’ll remove it asap. Thank you again for your sharing and time! Have a nice day!

  14. johnchuang Says:

    Hi,Kevin, you do a great job! It’s really a great post!
    I translate it into chinese and post it here :
    ,and I think some chinese animators would appreciate your sharings. Hope you do not mind I’ve translated and repost it without your permission. If you do mind, plz contact me ,and I ‘ll remove it ASAP.
    Thank you for sharing and your time!

  15. Kevin Says:

    Hi John! I know animation in China is booming, and there’s a lot of eagerness to learn, so I support what you’re doing. Yes, you have my permission to translate the posts here into Chinese. I just ask that you continue to provide proper credit and the original links, as you’re already doing. Good luck!

  16. Travis Tohill Says:

    Hey Kevin,
    Great post here. Even at the end of AM it is really useful to see other people’s workflow ideas. I’ve actually gotten a couple great things to try from this. I totally agree about leaving hands and facial expressions till the end. I’ve fallen into that trap of spending way too many hours on that and then having bad body mechanics as a result too many times.

    I know you’ve talked to our class about working in spline from the start, which is still something I haven’t tried yet (this insane AM schedule has me too scared to experiment with different approaches right now). I tend to block in full poses, but once I start splining I focus on hips and legs before I touch anything else. Once I’m out of school and have some breathing room I’m gonna try your workflow and see how it fits me.

    Keep the great posts coming!

  17. Animation Article Database · Knowled Says:

    [...] [...]

Leave a Reply

The animation and animation-related musings of Kevin Koch