Reflective Research
I began my walk cycle by watching Sarah’s video which helped me understand the importance of timing and spacing to make a character have personality and life. I decided to learn how to make a 3D walk cycle so that I can better understand animating characters using Maya.
I closely followed along with Alec’s tutorials and blocked in the different key poses that make up a walk cycle, ensuring the first frame and the last frame were exactly the same. For the limbs, I took note of what each control’s values were in my notepad and copied them into the opposite limb’s values.
I spoke with Alec about the foot placement of my animation – the feet planted down on the ground too fast because the root control was very low. I raised it just enough and liked the confident foot placement of the character. Neel mentioned that I should make sure the head stays steady, so the character is less jittery. This made the animation flow better. If the fingers, hand and arm animation was better I think the walk would have more character – I could improve this by having more secondary animation that follows through the hands and fingers when they swing to the extremes, and I would slow the arms down. The knee pops outwards which is distracting – next time I will ensure I don’t stretch the rig too much. I’m happy that the feet don’t slide around and with the little bit of hair that moves with the head as it bobs.
I found a very useful Youtube Channel named ‘Endless Reference’. It has hundreds of videos with different body types and puts them with a grid which makes it easier to spot the minor differences in motion that make a walk cycle unique. This video made me think about how I should rotate Azri’s hips more to look feminine.
For my run cycle, the Krita exercises helped me to train my brain to be conscious of which limbs are closest or furthest to the camera, and how they operate in a run cycle.
I applied the same technique I learned from the walk cycle tutorials by creating keyframe one, copying it to the last keyframe and then making a keyframe in the middle. This method results in a much smoother animation than straight ahead methods because the small changes in each frame aren’t too extreme, although there are times when blocking can cause problems as Maya moves controls from A to B without knowing how human bodies work, resulting in hands and arms sliding and twisting which must then be fixed. I looked at different key poses we used for the ‘Jack Run’ exercise and posed Azri accordingly. I made the mistake of creating the same number of frames as the walk cycle – this resulted in a slower run and in the future I will use less frames to animate a faster motion. I think this is what caused my final run cycle to look clumpy as I tried speeding it up in Premiere Pro.
Using these key frames wouldn’t have been realistic enough and would most likely have broken the rig, however I thought it was worthwhile to take note of these 2D key poses because they helped with timing my run cycle and visualised what position I was aiming for while animating.
Before I made changes, the fingers were very straight which I liked at first because Azri is sprinting, but later I decided it would look more natural to have them bent inwards slightly. Her back-and-forth upper body movement conflicts too much with her head movement which gives a pigeon-run effect, and I think the shoulders would benefit from being rotated downwards to make them roll naturally. There are problems with her knees jittering – next time I will make sure to be more precise with my rotations so I don’t have to counter animate anything to avoid confusion. In the passing position, I could pull the feet together to make the run less robotic. Azri’s left arm is different to her right which I think makes the left drag. I can fix this by ensuring the left arm’s axis rotations are the same as the right. Her skirt looks a bit glitchy as I’ve rotated it too much on frames that are close to each other. I’m happy with her hair as it bounces when she plants her feet. I think if I had timed my approach to this assignment better, I could have made these changes and rotated the feet and hands slightly to make her run more realistic, but I am happy with the overall positioning of the rig in each keyframe.
Before I began my body mechanics animation, I wanted to ask Liam if I could use one of his parkour vids as a reference – I realised his tricks were too complex to animate, and recorded myself lifting a kettle bell at the gym instead. The video file is too large to send from my phone – it shows me lifting the kettlebell from the ground and heaving it across the floor. I thought it was a good reference to use for animating the motion of lifting and moving weight.
I paused the video at evenly spaced intervals and keyframed each pose according to what I was doing in the video at the same time interval for the frame. I tried to exaggerate the dragging motion by keeping her feet low to the ground and moving the spine in the same direction as the planted leg.
This is unfinished, but in future I will dip the shoulders and arch the spine to imply the weight is heavy. I will make sure the head is facing forward and ensure the IK’s aren’t stretched too far to avoid joints popping. While animating this I could tell that the final animation wasn’t going to turn out as good as a different motion could. I found this clip on YouTube after searching for jump references – I thought a jump cycle would look good for my showreel. I brought it into Premiere Pro and cut it into evenly timed sections and blocked in the character poses on different key frames according to each video cut to help with natural timing.
I moved on to making a jump body mechanic motion.
This video had a good range of different jumps to choose from, and I was really excited to pose the different positions of this long jump because each pose is overemphasized. In the future I will remember to think about how the final animation will look and consider if the real life motion will translate well into an animation in terms of timing. 
I found animating this jump really interesting because it’s so stretched out that in the middle, where I expected the highest height Azri reaches to be, she was at one of her lowest heights. In retrospect I wouldn’t animate a longer jump in one spot because it looks wrong. The feet slide around which I can fix by ensuring that while they are flat on the surface I don’t alter their translation. The knees were problematic and popped when I rotated the legs into the landing position which was a very small ball – I tried to reduce how noticeable the popping was by raising the root control upwards, however more work could be done to change each control that affects the knee to smooth it out further. I ensured the body didn’t jitter around by only moving the root control up and down. I liked animating the clothes on this cycle, however the hair was challenging because it’s control translation’s could be moved around which caused jitters – blocking instead of straight ahead animation would have been useful here and I think it would have made the hair less jumpy.