Production Report

For this project, I worked along side Jake Jones and Rozi Barnes, our final idea for the formation title sequence is very much based on the initial ideas from the original brainstorming session with other students and the pitch we gave that day. The three of us wanted to work in the cut-out style that we had pitched to Denzil, so we decided to work together as a group to continue with the idea, we emailed the student we worked with from Penwith College though he never got back to us. We took the strong visual elements and ideas we had formed in the quick pitching session, refined them and then added new ones that were more successful.
None of us had any preference of who would be director or producer, we only assigned these roles (by pulling names out of a bag), because Andy preferred us to. Despite having these roles (Rozi - director, Jake – producer), we worked as equals, one asking the other two for opinions or advice on changes and new ideas as our ideas and techniques evolved throughout production. Jake created a blog (http://flyformation.blogspot.com/) that would enable us all to upload project updates and I created a file sharing account on Box.net that enabled us to upload symbols and files we’d need that could be accessed on any computer over the internet, it also meant that we could share files that were to big to send over email.
We knew the concept of the style we all wanted to work with, but it wasn’t until Rozi, who had a really strong mental image of what the finished style should look like, had completed the storyboard that we found a style we could all work to. We chose literal cut-out shapes that had a distinct look, but were simple enough in construction that the three of us could work on them separately and still create components that worked when put together. Jake and I used the storyboard images to create the animatic and some of the cut-outs from the storyboard were used in the final animation. I created an animatic of a fly, used to test the wing and leg movements in Flash, which I ended up using as a template for the final piece.


The animatic was used as a planning tool for the actual animation, Jake and I spent a morning editing each section to the length we wanted it to be in the final version. I then used this second version of the animatic as a template to divide the animation into workable sections, then figure out the exact number of frames each should be, (doing more maths than I had done in a long time during the process). I then drew up a table with each section and the frame rates so that the group could follow it. The workload was divided between all of us, mainly by personal preference (Rozi had said from the beginning that she had a strong idea of what she wanted the lab to look like, Jake was more keen to develop his ideas for the smoke and flyman – which turned into a flywoman for the real thing – so they worked on the desired sections), but also by workload: I worked on more scenes, because they were shorter and less complicated than the scenes the others did.


In theory, creating the table meant that we could all work separately, including at home, but the animation would still fit together as we were all working from the same time scale. In practice, we differed from it slightly: the beginning section had to be longer to read the word ‘formation’ and look the correct speed; other, unsuccessful, scenes were edited or removed to compensate for this, such as the countryside scene – which was supposed to be 80 frames long – which was removed as we weren’t happy with the final result. This lead to me extending the exterior window scene which came beforehand, this meant the narrative still worked and the animation flowed better.

(BELOW is the edited exterior window sequence. To begin with the shot was static, just the window and hint of buildings, it was edited so that the camera panned slightly to the right, revealing more buildings and trees, hinting the area is of human occupation, while the flies flew left.)


Working on the scenes as separate files meant that things would be easier to go back and edit if needed, which was beneficial as the longer start meant that we needed to go back to cut unnecessary frames and shorten overly long sequences. In addition, it was – again in theory – supposed to ease the strain for the computer of working on a long animated sequence. But in reality Flash still crashed or froze frequently: I kept a tally whilst working on this project (which I kept updated on my blog), the grand total came to 26, After Effects faired no better, failing 8 times within the first four hours of piecing the final animation together. I think this is partly because of the large aspect ratio we were asked to work at (1920 x 1080), which meant that Flash couldn’t handle the large textured background we wanted in each shot and lots of symbols had to be resized in Photoshop before they could be brought into Flash without it throwing its toys out of the pram.
One of the largest problems we came across was when it came for us to export the finished files we had worked on in Flash as videos to edit together. Due to the large stage size, picture size and number of different layers for each scene (including a separate layer for every fly, which got a bit manic when there was around seven different ones, each with their own individual Motion Tween, flying around the screen at once) it became a bit of a nightmare. The exported videos became ridiculously jerky so that the action would jump around statically every few seconds rather than flow smoothly like it was supposed to. To evade this problem, we exported each scene as a sequence of .PNG files, (the choice to use these rather than another image format was due to the fact that the .PNG can export an Alpha Channel, which we needed to allow us to add the textured background Flash had hated). This then allowed us to composite all of the scenes together in After Effects, with the background and later sound. However, the large aspect ratio gave us problems in exporting the final animation from After Effects as well. Denzil had asked for the final file to be rendered as an Uncompressed Quicktime file, which came to a massive 1.8GB so the computer could not handle playing anything after the first few jerky five seconds. I then tried various different forms of render outputs, a smaller mp4 (standing at a much more respectable and reasonable 64MB) played the animation in its entirety, but at about half the speed. After sending a slightly panicked email to Denzil asking his advice and acceptance of a different size – he suggested using either an MP4 (which I had tried) or an MP2. The Mpeg2 worked and played at the correct speed, however when we added the sound clip, a piece of copyright free music that Rozi had found, it rendered the two out as separate files. We were able to resolve the problem, by exporting the finished animation and sound together, as a HD video for Windows file, which is the one I have uploaded to Moodle, my blog and to YouTube for Denzil and others to view.


Left: Original .PNG file. Right: Still from the final animation after compositing separate elements with alpha channels together on After Effects.