Fall 2011 Post #2: Child’s Play Film Full Production


With pre-production done, a clear storyboard and character designs approved by the client, and scenery research completed, we began full production on the film. Each member of the team took on at least one shot for 1-2 weeks and then rotated who worked on which shots, to keep anyone from spinning their wheels on the same problems for too long.

As a hybrid 2D/3D project, the team used a lot of different tools: Adobe Flash for 2D shots and Autodesk 3ds max for 3D shots, and Adobe Photoshop for 3D art textures, 2D backgrounds, HUD designs, and credits. We rendered everything as PNG sequences and used Adobe After Effects for putting it all together into a single film. Effects work used all of these tools, and we used dropbox.com for our production and render storage.

Daily communication became even more important once full production began. The final film needed to be collected each week on Tuesday to be ready to revise and show in class on Thursday, and this meant the majority of the work– easily 10-20 hours a week by most of the artists– had to be done Friday-Monday each week. The artists learned that they needed to get drafts up on the department forum as early as possible to get feedback and avoid problems that would only get bigger the further they revised in a wrong direction. Those who got draft up Friday and Saturday were the most successful; those who waited until Sunday often struggled with late work, occasionally holding up the entire project. As the team practiced their weekly deadlines they got better about communicating their progress and showing drafts, and the production went much smoother by the last month.

For the audio on the project we had the pleasure of working with Akash Thakkar, head of the Game Audio club at Berklee College of Music. The Production Team class has worked before with Berklee students in general and Akash in particularly and we are always thrilled with their work. This semester we are enormously happy with the results of Akash’s efforts and appreciate his bearing with us through many revisions as we struggled to communicate what we wanted from the audio for this film.

As a course dedicated to teaching good teamwork as much as the production process, we held a number of exercises to establish trust and to learn to work better with each other:

* To promote conversation on how the team could be structured and what members wanted and worried about for this class, we held discussions on past team and leadership experiences. We learned methods for facilitating meetings and rotated meeting leadership responsibilities throughout the semester.

* To discuss bad practices in meetings and how they can be overcome we roleplayed a meeting where each participant got a slip of paper with a description of a particularly bad way to treat others, and then spent a hilariously miserable 10 minutes trying to agree on a pizza order.

* To explore ways to deal with issues before they become more serious problems we learned how to speak using I-messages and roleplayed difficult co-worker discussions from imaginary bad odors to bad performance.

* To help the team establish a space safe for talking about needs and what helps each of us to be most productive, we discussed Gary Chapman’s Five Love Languages and re-interpreted the relationship concepts into ideas appropriate for the workplace.

* To examine office politics we played Witchhunt, a variation on Dimitry Davidoff’s famous Mafia game.

The only tangible way the artists get paid for this huge amount of work (we estimated 1000 hours by this team of five by the end of the semester) is in their grades. In order to incentivize both good work and good teamwork, all artists write Peer Evaluations for their teammates at midterms and finals, and a significant percentage of their grade are determined by those evaluations.


Once pre-production is done, our course demands the full animation be drafted in just three weeks, in time for midterms. This is so the film can be shown to a committee made up of the animation faculty and department chair. The committee is given no information on what the film is supposed to be about; just a link to youtube and a request for feedback. This means all major elements needed to communicate the story must be completely clear by the end of those three weeks: all revised 3D animation, 3D character and setting models, clear 2D backgrounds and at least 5 frames per second of 2D animation. Considering the exceptionally large scope of what the client asked, the team did reasonably well, with only a few big elements missing:

The film-in-progress at Midterms,
after three weeks of full production


The feedback we received from the committee helped us to focus our efforts on rendering and texturing styles that would help differentiate each sequence, and also emphasized for us the need for a game-themed Heads Up Display (HUD) that would make it clear when Emma was in a game world. After a number of attempts over many weeks we decided to use a unified HUD rather than changing it for each game, and this led to the idea of a progress bar and text that identifies the specific event in support of Child’s Play that each game is supposed to represent.

In general, we found it more challenging to reproduce the motion of 2D games than 3D games. The 3D games at early stages of animation looked mediocre but still recognizable, but when the 2D games were off they felt REALLY wrong. The Emma-Mario felt completely off when she ran with fully articulated legs, and we needed to examine the original Mario’s motion much more closely. The Emma-Pacman’s munching tended to go too fast and the pigtails we added looked at first disturbingly sperm-like. The term “Uncanny Valley” in animation refers to near-miss attempts to reproduce facial expression motion in animation, where characters in movies like Polar Express tend to feel exceptionally creepy when they look close-to-reality but fall short, and the argument goes that we as humans are programmed to know faces so well that slight discrepancies feel huge and alarming. It seemed to us, working on the Child’s Play film, that the real 2D Mario and Pacman have become similarly engrained in us, so that slight problems felt terribly wrong, and we ended up tossing lots of revisions before we felt we got those scenes right.


While the team worked all semester to animate their characters so that body motion told the story in each shot, the opening and closing sequences needed the main 3D version of our protagonist, Emma, to be able to express a lot with her face.

It is possible to set up facial animation systems on a 3D character so as to allow the animator to have unlimited options for expression. We decided not to do that. Instead, we only created the few expressions we knew Emma needed for this particular film: various degrees of sadness, pain, surprise, and joy for the first sequence and happiness for the final sequence, along with eyeblinks.

This seemed like a good plan, but in our early attempts at the first sequence–which calls for Emma to appear miserable lying alone in her dark hospital room–she did not look particularly sad. We thought we could communicate misery with just closed eyes and a down-turned mouth, but she felt only sort-of sad. Still, with so much else to complete on the film, we might have decided that was enough and never fixed these critical establishing shots.

Then we took a field trip. As a class we attended a presentation by Chris Landreth, the Oscar-winning animator behind the extraordinary face-driven films “Ryan” and “The Spine.” We spent an evening with him studying which of the fifteen major facial muscles are critical to express various degrees of human expressions. At the end of the workshop we consulted as a class and decided Emma’s glasses were the problem. They were on her face for all the opening shots, and we needed to see more of her eyelids and especially her eyebrows to make her misery clear. We decided to start the sequence with the glasses sitting on a nearby table so there would be no barrier to her expression, and have her put the glasses on as part of her surprise when the TV activated to pull her in.

Once the glasses were removed, the animator on that sequence nailed the expressions in the first few tries, and suddenly Emma’ unhappiness felt very real. Interestingly, once the face worked we had to re-think some of the body poses and motions. Without the blank face as a distraction, we noticed the body was often too-much animated. When Emma rolled over in bed, her right arm moved so much it distracted from her face. Similarly, after sitting up in surprise and putting her glasses on, she originally threw back her hands to support a backwards lean on the bed–far too much motion when we wanted the audience to focus on the growing joy in her face. The animator found the fix by letting the hands drop gently downward after putting on the glasses, a perfectly subtle gesture.


Week eleven came up fast. While the 2D sequences only needed to be at 5 frames/second for midterms (week 8), they needed to be 12-24 frames/second to show the client at week 11. Similarly, the 3D sequences needed polish and effects. The artists pulled together to get the film done before Thanksgiving and, after eleven 7-day weeks of work, enjoyed a full week off. When we came back the client had only a single timing request, that the social media bubbles in the first game needed to stay longer on screen to be readable, an issue we predicted. In addition to planning solutions for this issue, we spent week twelve on a major critique of every shot, and came up with a list of over 30 changes to the project, including adding a total of five and half seconds to the film.

Our goal was to move every shot from bring a strong student film to the level of a broadcast-worthy professional production. Most of the shots began to work around midterms, like the Intro/Outro, 2d Mario, and Pacman and so with a month to revise we were able to get them to a high shine. Others. Like Guityar Hero and Mario Kart, only began to work around week 11 of the project, and required an exceptional push by the team, but the difference in polish is still visible between the shots that mostly worked around midterms and those with which we still struggled in the last few weeks. Overall though, we are very proud of this entire film and this team of filmmakers.

Last Thursday, December 15, was the last meeting of our class. For our final Skype presentation to Child’s Play we met at the teacher’s home for dinner and games. We sent Jamie Dillan the link and studied her face as she watched the end result of three and a half months of labor. We saw her smile grow throughout the film, and at the end she gave us the best compliment we could hope for:

“This represents us at Child’s Play, and what we do.”


NEXT WEEK: The final film!


You can read more about Child’s Play on their site, http://www.childsplaycharity.org, and we love this video, hosted by their parent company Penny-Arcade:

One thought on “Fall 2011 Post #2: Child’s Play Film Full Production

  1. Ive never read anything like this just before. So nice to find somebody with some original thoughts on this subject, really thank you for beginning this up. this web-site is something that’s essential on the net, somebody having a little originality. useful job for bringing one thing new towards the web!

Leave a Reply