Month: October 2021

I’m more than happy with this shot, I may even use it as the first shot on my showreel. It took me ages to build the set and I’m very proud of it. I have to say that my FMP shots have only gotten better as I’ve gone on- the cowboy shot took me four weeks and looked awful before I revised it, and this one took me about five days and I love it. I feel as if every new shot I create is my new favorite.

For the purposes of my thesis, I think it works well; the ending adds a little lighthearted factor that mirrors the ending of the cartoon version (the tiger shot). I also intentionally gave it a sort of video game style aesthetic, with Street Fighter in mind, which will allow me to further test whether adults favor a video game style in animated, realistic content, or whether it must remind them of such to not be considered juvenile.

When building the set, I also wanted to stick to that Street Fighter environment as much as possible; if it looked too much like real-life Tokyo or Hong Kong that would open the door for critique on consistency to the setting and time period. For this reason I did not use the background plate at all, but kept in the original camera track. However, I did add my own animation to the camera on top of that to fully showcase my animation. As a result, this shot incorporates my motion tracking abilities the least out of any of the shots I’ve created. For the purposes of my FMP, however, this isn’t necessary, as my focus and data collection for this project rests more on the animation styles (realistic vs cartoon) than the motion tracking.

I am now finished with the animation for my FMP. Before I decide it’s done-done, though, I want to fix the shadowmatte a little bit on the Dana shot. Perhaps I’ll revisit some of the others, too. I’ve been tweaking them as I’ve been going and they’re in good shape at least.

Unexpectedly, it took me quite a long time just to settle on the characters I wanted to use for my final shot. I knew that it had to be a fight scene, because the tiger shot, its cartoon counterpart, revolves around martial arts. I decided it should probably be a real fight this time rather than one person practicing or using a punching bag, to maintain interest and add differentiation. I chose the following reference:

Now, the only issue with this reference is that the men are clearly trying carefully not to hurt one another, and therefore are moving with no real force. This is okay, I have faith in my ability to incorporate that, and the slow movement gives me a clearer reference to study from. I decided that the fighter who flips the other one should be painted as the “protagonist”, and set to work casting characters.

Kachujin was an easy choice, she fits the theme and looks realistic while still being outfitted in sort of a fighter-esque costume. But from there, I hit a bit of a wall. Though I had four choices, I was hesitant in picking any of them, deeply considering audience bias and how certain rigs may affect my data in opinion polls.

I didn’t prefer the first samurai, as I assumed that the heavy armor may take away from the animation work itself and prevent me from being able to show off the work that went into it, and unfortunately, the rig for the fourth samurai did not work very well.

The second samurai would be perfect, except for the fact that he’s portraying a villain, and I’m hesitant to cast this character as such due to race-related biases and discomfort that the audience may feel while watching such a clip. Anything that may skew my data due to sociological factors couldn’t be used as trustworthy material.

Which brings us to the real issue I struggled over. I really preferred to cast Eva as the antagonist here, but had to accept, to my dismay, that I simply couldn’t without sacrificing the solidity of my study.

I wanted to use Eva because her costume fits the theme while staying cyberpunk-y, sort of a Street Fighter vibe, thus setting the shot apart from expectations of having to stay culturally and historically accurate. Eva as a rig would be amazing to use, with lots of controls in the hair and every accessory on her body, as opposed to the unmovable accessories and hair on both Kachujin and all the samurai. I wanted so badly to have Eva’s long braid whip around during the fight. Unfortunately, she is simply too sexualized to use for the purposes of this project. For one, I plan to collect data from a broad age range, and her outfit is simply not suitable for children. In the same vein, I fear that this character may sway audience opinion in favor of this shot for reasons other than the animation itself, especially as the fight would involve two women.

I considered for a while using her anyway and using this as an intentional factor in whether adults require animation to appeal to their sensibilities to be considered less juvenile, but it felt like the kind of thing that would require a whole separate, vast study. Perhaps related to what Mariana is studying, which I believe centers around whether audiences of animated films require their female characters to be feminine and conventionally beautiful. It’s an interesting subject worth further consideration.

That left me with Samurai Three.

I came very close to using Samurai Three, but ultimately decided that he, too, must be rejected for the same reasons as Eva: an adult audience may view a woman wrestling a shirtless man from a different lens than they would a faceless, heavily armored guard.

And so it was with much deliberation that I settled back on Samurai One. The good thing about him is that he’s got clear “bad guy” vibes, as well as being objectively “cool”, thus adding interest as well as putting Kachujin in the protagonist seat.

These rigs are both, I assume, Advanced Skeleton, which gives me a lot of flexibility in movement, but unfortunately both lack controls on their hair and accessories, as I’d mentioned earlier. Kachujin herself has no facial rig.

I could rig the accessories myself, but I am nervous about doing so, as the skeletons are already so complicated.

Conclusion: my outlook is hesitant. Also, I expect this shot to take a decent amount of time, as it involves two characters, is required to be realistic, and runs for a full ten seconds. This may be a challenge.

It took me nearly a month to finish the original cartoon cowboy shot, which was the first I’d attempted to do of my FMP, and when it was done, it looked so bad I nearly lost all motivation to continue the work. Up until this past week, during which I’ve been traveling internationally, I’ve deliberately put off revisiting the cowboy shot, as it depressed me so to see myself fail at a walk cycle. But, in between packing bags, it was the perfect project to break up the leftover time, as the pre-built set and tracked shot required no new planning, only extensive animation repair.

As I mentioned previously, I found that starting over was the most efficient option. That choice turned out to be enormously beneficial, and to my relief, I found that I could whip out a walk cycle no sweat in just a few days, as I should expect at my current skill level. I worked closely with the reference to match the timing perfectly and avoid at all costs the “drunk walk” and weird robotic horse of before.

Of course, once the reference is copied correctly, stylization is the easy part. I chose to give Sheriff Amy the bounciest possible hat, a confident march, and a goofy, stretchy bounce in her step. As with Dana- both rigs, may I mention, are by the same creator, Gabriel Salas, and are the best $15 I ever spent- my favorite bit to animate was her hair. The floaty hair gives the final touch to the weight of her walk. I did find myself wishing that her hips/waist stretched as well, to give a nice jelly-like overlapping squash along with the chest stretch.

George, the horse, was a bit more difficult. Though human walk cycles are old (floppy) hat, George is now only the second horse I’ve ever animated, and of course the first was done rather poorly. That being said, this minimal experience gave me the skill I need for smooth sailing on the horsey sea. With the reference at the ready, I focused my efforts on an up-down motion of the shoulders then haunches, in which the hips raise in each passing hoof pose, thus creating an effect of the front of the horse then rear of the horse raising one after the other.

Pleased with the effect, my main goal after discovering this strategy was to smooth out the hoof movements while keeping the weight realistic, and to exaggerate the upward motion of the hips enough to match Sheriff Amy’s bounce while preventing the horse from appearing too unruly. That took some adjustment. The last step was my favorite, of course, the hair.

Once again, the hair drives home the animation of the character itself.

This is not a showstopper piece like the tiger shot or Dana, but it’s achieved what I’d intended, a finished walk cycle for the cartoon cowboy that looks passably decent and fits consistently into the aptitude of my other FMP work while providing a stark contrast between its realistic counterpart. If I could change anything, I’d say that the horse looks worse than Amy. He still looks good, but it’s clear that I’ve only just figured out how to animate a horse, whereas with humans I’m comfortable enough to play around with stylization.

With my flight back home leaving tomorrow, I’ve had little time to work on my FMP since finishing my internship. But with packing, sharing goodbyes, and obtaining the necessary documents done, I decided to fill some extra hours by getting back into that cowboy shot that had frustrated me so much for nearly a month back in July.

I at first attempted to incrementally save vast changes to the animation, but after half a day, decided definitively that the problems lie in the very baseline timing and posing and I simply must start again, this time with a better reference, a better rig, better blocking.

Though the deadline draws ever nearer, I am confident in this decision. I do not believe I could have achieved a satisfactory result with my original characters and animation. In a span of just two hours, I’ve blocked out a new animation with a new reference, and as my tiger shot took only four days, I believe I can eat that lost time by working harder and smarter on the cowboy shot for a better result.

It’s unbelievable to me that I started and finished this in only four days and it’s hands down the best work on not only my FMP but also my showreel…. I feel giddy!!

Again, the wildest thing is how a year ago it took me four weeks to try and fail at a walk cycle, completely new to the software, fumbling around in the darkness.

I can’t come up with a critical analysis of this right now because there’s literally nothing I would change, I’m proud of all of it. I guess if I translate this to analysis, I can say again that I am most proud of my ability to demonstrate the character’s weight and force in each punch as well as deliberately timing the kicks and moment the punching bag hits him. Rigging and creating the bag itself gave me the opportunity to demonstrate weight, arcs, and timing. Between the punching bag and the tiger, there’s a satisfying juxtaposition between heavy and light weighting. It’s really all just good, good stuff. I’ve gotta celebrate this.

Next step is to create the realistic version of this- probably just a loosely related version with the samurai- and then I’ll theoretically be done with my FMP and ready to collect research data, in the final steps of my thesis…. however, what I really should do is go back and make that first cartoon cowboy shot look less abysmal. I’m dreading that because I’m so frustrated with how bad my work is there, but I know for a fact I can do well. I just did this week.

One of my coworkers mentioned to me that this was one of the “better animations he’d seen me do”, but also noted that although it was arguably good animation, it was confusing what the character was supposed to be doing. My reference was of a girl practicing martial arts alone in an empty room, and I’d hoped that his “practicing” would also read as natural. Obviously, though, it didn’t, so I decided to add in the element of a punching bag.

I downloaded the model online, then added the texture and rigged it myself. It’s comprised of four joints, and it was honestly a delight to animate. Though of course it presented the challenge of matching the swing to the punches, it gave me a chance to show off my overlap and weight skills. And speaking of weight, someone also gave me feedback that the punching bag “feels light”, which I completely see and agree with; however; I want to leave it that way because it is satisfying to watch. It can be a lightweight punching bag. I’m a bit too tired to decide what I think it’s filled with, but the best I can explain is it’s like those neon sandy pillows from the 90s. I don’t want to change it now, I’m very happy with the weight and timing of the arcs that I’ve done. I also like that he’s kind of ducking under the swing after the jumping kick, it feels more interactive.

So instead of adding in a comedic feline moment where he sees a cucumber and freaks out, I decided to make it more reliant on the punching bag; otherwise there’d just be too many elements involved. So I think for the ending I’m just gonna have it smack him in the face and he’ll fall down.

The next step for this scene is to add in the ending. As mentioned before my original concept is to give the character comedically feline behavior, perhaps by knocking over a barrel of groceries, spotting a rolling cucumber and jumping away in fright.

As of now, I’m already very happy with this work; it still needs some polishing in the graph editor but I feel that it shows good knowledge of weight and timing. This is going to be one of my strongest pieces.

I am very grateful to have the opportunity to add in ear and tail animation, as it (as well as hair) is one of my strongest suits and always adds a nice bit of interest.

This is the first 60 frames of my “tiger animation”, or the cartoonish version of the martial arts shot. I’m mostly posting this here because it’s beyond insane to me that I did this in 2 hours. A year ago, this would’ve been unimaginable even for a month’s work.

Here’s the realistic counterpart of the cowboy shot (which still needs lots of reworking)- the realistic cowboy walk. I interpreted this character as being a zombie, and feel validated in this, as when I showed it to a small child he said “I think he’s a zombie!” This shot may be my second-best work for my FMP, the first being my Dana shot. It’s miles and miles better than my cartoonish cowboy shot.

When I first began working on this shot, I struggled to find a rig that worked well for me, and was near giving up when I finally, as always, reached out to Luke for help and discovered a solution in hours. We set the rig up with advanced skeleton, and from there it was extremely easy to animate with a video reference. I got it done smoothly with no problems, and it’s one of my favorite pieces of work now, especially in comparison with some of my early walk cycles when I first started the course.

I like the idea of giving it a Red Dead/Dark Souls kind of feel, so I have asked my sound designer friend from work if he’d be willing to incorporate some footsteps, music and ambiance to the shot to really give it a video game vibe. I was thinking of even using it as the title card of my showreel, with Skyrim-esque “UI” appearing on the right with my information and a box labeled “proceed” which a cursor clicks on. I’m only indecisive because the Dana shot is so good and I know most employers make their decision within seconds.

All that’s left is the final two shots. I’ve chosen an anthropomorphic tiger for the cartoon version, who will practice some karate moves before knocking over a basket of groceries, and when he sees a cucumber roll next to him he freaks out and jumps away like a cat. It’ll be a good demonstration of body mechanics as well as being comedic if I can pull it off. I think for the realistic version I’ll be using some of the samurais we have.