For the end of term brief, we were assigned to do a character walk. We could experiment with it however much we liked, although we were advised to try two different walks within the final clip - this is to maintain interest, and also just to show off talent. The other requirement was that we had to include the ball & tail rig, making it interact with our human rig, ‘Animo’.
(click images to enlarge)
I began by thinking about lots of different characterised walks, shown in part by my list in the top right-hand side of my notes, pictured above. As a class, we drew inspiration from the classic opening of ‘101 Dalmatians’ where Pongo watches a number of different women walking their dogs, demonstrating a wide breadth of walk cycles.
At this stage, I wanted to start doing my own reference walks. Taking the list I’d made of different character walks, I took my tripod and camera and filmed myself performing the walks.
As I started concentrating on my own work, I tried writing down a few rough plans for story ideas, focusing on ‘who’ - the subject of the animation, ‘what’ - the actions of the subject, ‘where’, and a few notes on ‘how’ the subject would do this walk.
After I picked which story/situation was most engaging for me (and for anyone else watching), I began working on my storyboard - which turned out to be an extremely useful guide throughout the animation.
Possibly the most useful reference, however, is pictured above. After I knew exactly what I wanted for my walk cycles, I filmed myself acting them out and took many screenshots of the clips as I watched them back later on my laptop. I then picked out the main positions of the walk and put them into Adobe Photoshop, where I annotated each still, writing what different parts of the body were doing during different stages of the walk. For this specific walk, I drew the lines of action for each position, which made animating the scene in Maya far easier as I knew exactly what to look for.
Without this guide, I’m positive that my animation would have little to no realism to it.
I did the same process for different sections of my animation - for example, the swivel transition (pictured above), and the side-step walk cycle (pictured below).
My next mission was making sure my timing (and spacing) was as realistic and as life-like as possible. I already knew that a tip-toe walk was decidedly slower than an average walk, meaning that the cycle would require a lot more frames than usual, but by putting my reference clip into an image plane in Maya I was able to assess exactly how many frames each position of the walk would use.
This turned out to be slightly harder than I imagined because in a realistic walk, the timing is never exactly the same. As a result of this, I had to scribble out lots of notes and pick an average number from three different examples for each position of both walks.
I thought that my first completed walk cycle (http://www.youtube.com/watch?v=5OQlK_jBKs0) was quite effective, however there are, of course, mistakes that I would like to clean up. For example, although it isn’t clear on a white background, both feet are sliding. I was able to sort this out later for the back foot position, by making many subtle changes in the graph editor for translating the global control. The front foot still obviously slides though, which is disappointing. Regardless, I feel that with more time and a more detailed foot model, I would have been able to make the tip-toe walk even better.
This screenshot was taken during animating my tip-toe walk cycle - the graph editor demonstrates ‘follow through’ in the torso, which alludes to realistic weight during motion.
The next step was creating the ‘where’ - in other words, the environment for my animation.
Although we were told that we’d only be marked on our walk cycles, I felt it necessary to incorporate an environment, not only for appeal, but to add a sense of realism to the walks and the situation. It is for this precise reason that I wanted to include audio in my final animation as well.
In order to create my environment, I used simple polygon planes as the floor and walls, and took an image of a carpet texture from mayang.com/textures, using it as a material for the ground plane. The assets I used were downloaded directly from turbosquid.com. Creating the finished bookcase was a feat - as I’d only downloaded seven different books, I had to copy and paste them all multiple times, arranging each one individually on each shelf of the bookcase, then shading each one differently.
At this point in the animation, I felt I was halfway there - the animation for Animo was almost complete, I just had to finish animating one of the arms that would be interacting with the Ball & Tail, amongst other tweaks and adjustments I needed to make in the graph editor.
I was then able to begin animating the Ball & Tail. As shown in the graph editor above, this involved some subtle squash & stretch at the beginning to indicate breathing - making sure that the intake of breath was faster than the exhale, which indicates the creature being asleep - and towards the end, lots of jumping up and down with added rotation.
The graph editor above illustrates something close to an ‘ease out’ in each Ball & Tail jump - as gravity increases on the ball, it travels less distance. Because I was adjusting the ‘translate’ curve, I changed the tangents to ‘linear’, as I didn’t want the ball to go back in on itself (which would have happened if I’d left a smooth curve at the end of each jump.
Above is the final animation, using just one shot so that the viewer can ascertain exactly what’s going on without focusing on specific elements of the animation. At this point, I’d say that one of main criticisms is with the lighting. I made the mistake of rendering 710 high quality frames before testing and rendering one frame in case I’d made mistakes. In hindsight, I should have adjusted the lighting and done another render, however at the time I let myself get too frustrated about it, especially considering that the render took over three hours and I was worried about the amount of time I had left to finish the project. So, unfortunately I carried on rendering different camera angles, having to put up with bad, low lighting.
What I do like about this shot however, is that you can see more detail in Animo’s
movement, and I hope, subtle changes that make Animo’s walks more life-like. For example, I had adjusted the width between Animo’s legs a number of different times during the side-step walk, because as previously mentioned, no human walk is perfectly symmetrical. It is also easy to notice that the pace of the side-step walk quickens towards the end, which I feel demonstrates realism as well. With this wide-angle shot, you can also notice the arcs I’ve included in Animo’s movement, which is, again, realistic to a human being as much of our movement is based on a circular path. This is shown most clearly in the hips and torso as Animo rises from stepping on the Ball & Tail, and again as she slides against the wall.
After lots of editing, the piece was complete! On the whole, I’d say I was pleased with the finished result, although this is slightly undermined by the criticism I have for it.
During editing I tried to stay as close to the storyboard as possible in terms of including interesting, appealing camera angles and shots. However, given that the animation is just over 30 seconds long, I didn’t want to make the same mistake I’d made with my Ball & Tail stairs animation by cutting it up too much. In using one particular shot showing the Ball’s point of view, I realised I’d made a big mistake by not including a ceiling into my scene! I only came to realise this well after I’d uploaded the animation to youtube, therefore as with my lighting error, I felt I didn’t have enough time to go back and change this.
As previously mentioned, the front feet of the tip-toe walk cycle slide forward too much, and the door is yet to be animated. Also, towards the end, I feel that animation becomes less realistic - the weight of the Ball is ever so slightly off, and the overlap in the tail is not fluid enough for my liking. In trying to demonstrate good arcs, I feel that it perhaps took away from the realistic motion of the ball, and that I maybe should have created a longer rise that travels the majority of the distance and a shorter fall that doesn’t travel very far.
I also feel that Animo is too rigid in her last walk cycle, and although she had follow through in her spine and head on impact with the wall, there is decidedly less follow through during the side steps. As well as this, I am not satisfied with the ending, as again, I was pushed for time so I had to shorten animation of the Ball & Tail.
There are editing mistakes as well, which I didn’t have time to go back and adjust. For example, there is a very slight delay in the audio of the dog yapping and squeaking, which could be subtle to other viewers, but is glaringly obvious to me! As mentioned before, audio wasn’t required for our animations whatsoever, but I personally feel that it’s an essential element that adds to the suspended disbelief of the work - it’s far easier for a viewer to ascertain what’s going on in a piece of film, animation or otherwise, with the help of sound.
I created my sound by downloading free audio samples from different online locations, cutting and editing them within Premiere Pro, and changing the dB levels for each sample. One particularly tricky sample to work with was the yapping Yorkshire Terrier - this audio clip had lots of background noise, which was obvious when watching it back as it would drop to silence between each cut. To minimise this, I added in some 'atmos'; I took a small section of the sample where the dog wasn’t making a noise, then copied and pasted it over and over to create the background noise. If you listen closely, you can hear each time the sample repeats itself - a mistake I could have avoided if I’d found a better sample or if I’d been editing in a more audio-focused piece of software, such as Adobe Audition.
Despite my slight displeasure with my final film, I am incredibly pleased with the progress I have made this term. I feel I’ve learned more in these past 10 weeks than I’ve ever done before, which, I hope, shows in my work. I’ve now gained many new skills in Autodesk Maya and I hope to continue using them and expanding them as I continue with the course. This does not mean to say, however, that this term has been without its many challenges and difficulties, but that all comes in part with the learning process; a hurdle which I’m proud to say that I’ve jumped over. Most importantly though, I feel that I have truly grasped the principles of animation, and that each one is incorporated in my final piece.
