Anatomy of an Animation Project: K is for Kids
We just finished up two spots for Netflix—the intro and outro sequences for a content advisory film—and wanted to share as much of the behind the scenes process as possible. Rather than focusing on both spots, I’ve selected mostly stuff from the intro spot to spotlight. There’s mentions of the second spot, but I don’t get into the process of it very much. The first half of this article explores how we approached the project internally and the second half focuses on our communication with the client. Also, at the bottom you’ll find a lot of links to things like storyboards, drafts, frame-by-frame animations, and more. Here’s the film and you can learn all about how it was made below!
Let’s talk a bit about how we organized our team on the project to maximize productivity in the short amount of time we had (learn more about the timeline in the second section).
We don’t subscribe to the idea of assigning specific shots to specific animators. As a studio, every one of our team members brings something unique to every project and [throughout this process] we wanted to utilize those strengths throughout all the shots. This started with the idea—we chose to combine a lot of different styles for two reasons. First, it fit the brief from the client. Second, it would allow every member of our team to work in their strongest medium, thereby maximizing our efficiency throughout the project.
Maxwell started modeling and animating the 3D assets immediately, as they would be the base on which everything else was built. Here’s a few examples of those early models.
Max did work on the project alone, roughing out all the animation for about 4–5 days. We’d kick out super low-quality renders to get the timings right and then do a full Octane render each night to see how the updated textures and materials were looking. Here’s a little information about his process in his own words.
“In wanting to keep the models on the more playful side, I decided modeling the 3D elements with big soft edges would be key. This led to predominantly using box modeling with a subdivision surface to strike that look. Working with lower poly cages allowed for a quick iterative process, which was super helpful.
While not always forgiving, Octane is a pleasure to work with. You can get beautiful results quickly, but there is always some challenge that presents itself (this time: z-pass). This project was also the first time I got to work with Octane Scatter, and it was incredibly helpful in filling in the landscape while keeping render times manageable.”
Soon there was enough of a base for everyone to get involved. Taylor had previously done a lot of keying for Sesame Street, so she took over keying the puppets. In addition to the keying she made some of the frame-by-frame blobs that populate the backgrounds of the whole video, modeling some of the trees for the earth sections, and built the 3D space suit. She was constantly working with both Max and Zac to pick the right sections of characters to key and for adding her 3D elements to scenes. When asked about her favorite part of the project, here’s what Taylor had to say.
“I loved that everyone was pumped about this project from day one. I was so excited to come to work everyday because it felt like everyone was utilizing their best talents in such a weird and fun way together. Each new render we saw of everyone’s combined efforts looked better and better day by day, and this made me want to keep pushing my work towards perfection. I got to do frame by frame animation, 3D modeling, and after effects animations for this project and being in all those arenas had me feeling the perfect balance of challenged and motivated.”
Nayt was consumed by another project for the beginning, but jumped in make the frame-by-frame versions of the characters being fried by the rocket booster and shot by the laser gun near the end of the project. He used rough animator on an iPad Pro to create the animations, which he exported with alpha for compositing. Also in rough animator, we created and animated most of the little creatures that you see scattered throughout the whole video. Our direction was basically, ‘make weird little creatures that aren’t like real animals’ and he knocked it out of the park. Nayt had a few words to share about the project and working in Rough Animator.
“First of all, K really is for Kids. For this project in particular, I started with two pages of ideas, and then took the characters that Sam liked best (for kids), and just went crazy in a blend of pose to pose simplicity. As an animator, we all want to have freedom to make something that makes us feel something, and honestly, it’s amazing how much tools can get in the way of that quickly! So as we began thinking about how we could make as many krazy little guys as possible (for kids), I had to suggest RoughAnimator. One of the HUGE advantages to RoughAnimator is that you get a really streamlined process, on a really streamlined device allowing you the freedom to work like you would on a computer, but feel like you’re a little bit closer to a page (for kids). So if you like animation and fancy yourself an animator, save up 4.99 because that’s a premium that’s tough to beat! for Kids!”
Michael was working on making the multiple explosions and the rocket booster during this time. Zac pointed him in the direction of video game explosions and drew a lot of inspiration from this spectacular video. Around the middle of the process, he painted a few UVs for the 3D objects, which were ultimately scrapped (find out why below). Near the end, we made a decision about changing the background from images to painted (also below) and had him work to bring those together.
“I was always told clouds and smoke were some of the hardest things to animate. The rumors were true! Getting the explosions and smoke to move and act in certain movements while having it dissipate was a fun and challenging learning process.”
Finally, Zac brought everything together in compositing and added a bunch of his own flare to the spot. Rather than writing up exactly what he did, I just asked him to write a paragraph on it.
In our short history at IV we’ve never had the opportunity to combine so many different styles and mediums together into one piece. While this was enjoyable, it also presented a challenge in the compositing stage. We had elements from every single member of our team (even Austin our Relationship Director!) and they all had to come together into one cohesive piece. Stop motion, live action, frame-by-frame, puppets, 3D, 2D… pretty much everything. The core of project was built in Cinema4D and rendered in Octane. After bringing the renders and camera information into After Effects I began layering in the puppets and frame-by-frame elements given to me by my team-mates. We rendered a z-depth pass from Cinema which helped us get some nice depth of field blur using Frischluft’s excellent Lenscare plugin. The z-depth pass also allowed us to composite our 2d elements throughout the 3d scene without having to render out specific object mattes using After Effects built in 3D Channel Extract. Late in production we decided to add a few extra 3D elements (ie… 3D clouds and a glowing rubiks cube), but instead of re-rendering another pass from Octane we opted to quickly add them with VideoCoPilot’s Element3D. After many hours of color correcting each element to fit in the scene we added a bit of Reel Smart Motion Blur from RE:Vision on top.
As I read through what I just wrote, it feels a little silo’d, which is strange, because it really wasn’t. It’s a little hard to describe the amount of collaboration that happened between everybody without sharing our entire slack communication logs from the whole project (trust me, you don’t want that). As the wise old man from my hometown always said, ‘ah well.’
Last Minute Changes
As the project progressed, there were a few things we had originally planned that weren’t coming together like we hoped. Rather than trying to shoehorn everything into our original ideas, we decided to call a few audibles that seemed to be where the project was naturally progressing. We were lucky enough to have a client who was totally fine with last-minute changes, so most of these were implemented on the day before final delivery.
We had painted a bunch of UV’s for the 3D objects in an effort to add texture to the otherwise very plain 3D models. We stuck with this idea very deep into the project. You can still see the slight textures on the car, the road, and the ground in the reference image here. This screen grab is from the last draft we sent before delivery, so we definitely waited until the last minute to make the decision.
For anyone who has ever created anything in color, you know how frustrating it can to find the perfect combination. As we neared the end of the project, we were pretty happy with our space colors, but the ground and the studio colors weren’t quite clicking for us. We didn’t hate them, but they didn’t seem as adventurous as the content of the video. We were doing lots of weird stuff with the puppets and the scenes, but the colors felt very conservative.
As with most creative problems, we found ourselves searching Dribbble for some inspiration. After browsing countless pages, we found a few pieces that we felt captured the idea of what were were going for with the clear, bold, and somewhat crazy vibe. Colors are hard.
We took this inspiration and made some tweaks to better work with our lighting and environments in the 3D objects. Did we make the right choice? Were the original colors better? Those are probably questions we’ll be asking ourselves for the rest of our lives about every project ever. What we do know is that we think the new colors better captured the feeling of the overall project.
One of the core ideas was mixing different mediums to create a very eclectic world. From the very beginning our ideas for the background was always to use real-world images. This wasn’t something that we ever really came to dislike, but at the very end it started to hold us back on other changes that we wanted to make, like colors and integration between the earth and space. Because of this, we decided to have Michael paint our backgrounds instead of using images. We think there’s virtues to both looks, but ultimately the painted background worked better for how we wanted to project to flow and feel.
This was not a last-minute change, but as you can read in the boords, part of our idea was having the car and the initial spaceship be practical papercraft assets. We did build and even shot some tests with these assets, but ultimately decided that we weren’t going to be able to achieve the level of quality that we wanted with the time and resources we had.
When it comes to things like this, we take the Pixar approach to failure. That is, failure is to be celebrated, because it teaches you what works and what doesn’t.
Piggybacking on the demise of the papercraft, the opening scene would need to change. We still liked the idea of keeping it in the real world, but also explored the option of making it entirely 3D. For the 3D approach, we stuck with the same idea that we had storyboarded—we’d just model a dashboard and make a 3D hand instead of a real one. We really liked where this was heading, but always planned on using the real-world approach if it worked out, which it did.
As soon as we decided to scrap the cardboard dashboard, we started thinking about ways to make it even weirder. Nayt had brought up the idea of throwing the tape in a microwave and have it catch on fire. That particular idea didn’t make the cut, but it sent us on a non-sensical journey into weirdness.
The journey culminated in a visit to a local antique store, where we found all the items you see in the intro. We just wanted an eclectic mix of old technology that in no way works together. We just wanted it to be weird.
Every project comes with it’s own set of limitations, which we try to view as a positive thing—creating within limitations often produces better work than a directive with no bounds. On this project, we had two major limitations in the form of a short timeline and pre-shot footage.
One of the things we were most excited about from the very beginning was the idea up placing puppets in this animated world. However, it came with the caveat that we wouldn’t be able to shoot any puppet stuff to fit our ideas, we had to work with footage that had already been shot. There’s a few instances where we could tweak the puppets a bit given the chance, but we’re overall really happy with how we shaped the idea around them.
As you’ll find out in the next section, we didn’t have a lot of time to create this project. From beginning to end, we barely scrapped by with three weeks. We had to really hone our workflows and make sure we had actionable plans, because there was no time for much experimentation. As a producer, I’m never quite sure whether I love a quick timeline or a relaxed timeline. There’s a certain determined creativity that emerges when everyone on the team knows there’s no room for big mistakes. This can quickly be lost when you feel like there’s no rush.
There’s not as many visual aides, but this is where you’ll find how we communicated with Netflix from the beginning to end of the project. I would classify our relationship with Netflix as extremely casual, meaning this isn’t very representative of how we normally communicate with clients. I’m a big believer in matching the tone of our communication to that set by the client. If they are casual, I will be casual. If they are to-the-point, I’ll be to-the-point. It’s all about finding the right balance with every client.
July 12th, 2017
It started with a phone call from Andrew House, in which he explained that Netflix was making a content advisory film for Netflix Kids producers. The kicker — it was a puppet film and it needed a title sequence. The two words given as creative direction were ‘psychedelic’ and ‘absurd.’ Also, it didn’t have a budget that reflected the deliverables and did have a very quick timeline. As most calls like this end, I said we might be interested, but the budget and timeline meant I needed to talk to the team first. A little later that day, I scheduled some time to talk Zac through a couple project inquires that we’d received. I explained the Netflix project and I believe his exact words were, “puppets and psychedelic? I mean, we definitely have to do that, right?”
And we were off to the races. Sort of. It took a few days as you can see below.
Thursday July 13th, 2017
On Thu, Jul 13, 2017 at 12:40 PM Samuel Cowden wrote:
Hey Andrew! I think we’re in! Can we get $xxk?
Could you email me your deadlines?
Sunday July 16th, 2017
On Sun, Jul 16, 2017 at 1:45 AM, Andrew House wrote:
Awesome. $xxk it is.
Let’s talk and nail down scope and deadline Monday. It’s got a lot of moving parts and I want to have everything lined up for you.
Monday July 17th, 2017
On Mon, Jul 17, 2017 at 8:02 AM, Samuel Cowden wrote:
Hey Andrew! Sounds great! Today is a little crazy for me. Do you have time tomorrow? I should be able to make anything work.
Tuesday July 18th, 2017
Andrew called me and we talked through some details. He said he’d send me proxies of all the puppet footage they had captured on green screen that we’d be able to use for the title sequence.
And then follows up with an email to summarize the whole project.
On Tue, Jul 18, 2017 at 4:50 PM, Andrew House wrote:
I am going to assemble a lot of things into this email and we can use the thread to communicate on the project going forward!
Need: The best opening title sequence for a puppet show of all time. Something that will grab the attention of the audience and harken back to a simpler time where puppets taught us all how to live life. The sequence needs to be visually jaw dropping. It should feature our puppet cast in some way. Blue sky past that. Our references have been Avenue Q, Sesame Street, Yo Gabba Gabba, Sid and Marty Croft shit, and Showbiz Pizza.
Deliverables: (1) One opening title sequence and (:45 or so) (1) Finale closing sequence (:30 or so)
Timeline: Need all finished and dusted by 4 August. Can guarantee quick turn arounds on feedback.
The rough of the song is attached — the new version comes over tomorrow and I will share that as well when I get it — the composer would love to work with your team as well to flesh any pieces or flourishes out with you that you need for the composition.
I know this is an aggressive schedule, but we have no stakeholders outside of the team, and are committed to ensuring you have what you need to pull this off.
(I can’t actually include the rough of the song here, but it sounded mostly like the final version, just rougher)
July 19–20, 2017
There were a few email exchanges on these days that contained a bunch of clarifying questions and answers based on the previous email. I’ll summarize them here, because the actual emails are booooring.
The actual music tracks were a little longer than the deliverables, so we wanted to know which portions we were actually animating underneath.
Andrew sent us two new rough versions of the songs.
We asked if it should be irreverent or somewhat tame. He explained the the actual film was irreverent, but “the opener should just be big and showy to get everyone interested.”
Thursday July 20th, 2017 (later in the day)
On Thu, Jul 20, 2017 at 3:12 PM, Samuel Cowden wrote:
Andrew! I have a storyboard for you. Read through my descriptions below and then check it out.
Our primary goal as we approached the intro was to create a narrative structure that we could then add crazy weirdness onto to create something awesome. What you see in the storyboards is that narrative structure. We wanted to create this story that carried the puppet characters through ridiculous situations and mirrored the slight irreverence of the movie as a whole.
What you don’t see in the storyboards is the craziness that will be layered on top of every second of this video. We’ll be mixing stop-motion with 2D animation, with 3D animation, with puppet animation and it will all work together amazingly. Ultimately we want the final product to be a perfect blend of the three videos I’ve linked here:
Alas, here are the storyboards:
Want to jump on the phone to talk through any thoughts you have? You can just give me a call whenever!
On Thu, Jul 20, 2017 at 3:37 PM, Andrew House wrote:
I will be tied up for the next hour and a half. But this looks amazing and I will absolutely call you as soon as I can this afternoon.
He called as soon as he could that afternoon.
It wasn’t a long call, he essentially said that everything looked perfect and we should move forward into production. Hurray!
Tuesday July 25th, 2017
There were some emails about check-in dates and getting the original full resolution files of the puppets.
Friday July 28th, 2017
On Fri, Jul 28, 2017 at 4:23 PM, Samuel Cowden wrote:
Here’a a couple updates from out end.
We’re cruising along through some drafts. I’ll include a link below to the most recent. It’s still missing sections, but should give you an idea of where we’re headed. The only thing that’s changed is that we’re probably not going to able to incorporate any papercraft like we had originally planned. Getting models that were good enough quality in the time we have proved impossible. That being said, we’ve just replaced it with 3D stuff, which looks amazing. We are still planning on the cardboard dashboard at the beginning.
This is still missing A LOT. We’ll be adding lots of 2D effects and characters and animations. There will also be real-world photos incorporated as the background.
We’ve settled on an idea that covers the time it takes to say twelve “K is for Kids” lines in the outro. Each time the phrase is said, we’ll switch to another, crazier, rendition of a “K.” Here’s a link to view some of our initial renditions.
We had a phone call and some emails that just confirmed we were on the right track.
Tuesday August 1st, 2017
We received a hard drive with the footage for the puppets. It had arrived a little later than we had hoped and there was still a lot of keying to be done. Because of this, I had a conversation on the phone with Andrew about pushing our final delivery date. He was happy to push it a bit as long as we could still deliver a full-draft on the 4th — done deal. We now had until August 8th for final delivery.
Thursday August 3rd, 2017
Want to know something embarrassing? At some point in the process, I had come to believe that Thursday was the fourth, so was pushing our team for a Thursday delivery. It was not until the day of August third that I realized my mistake. However, it did relieve a lot of pressure from the team when I admitted my mistake.
Friday August 4th, 2017
Aside from a few small additions and tweaks, we were still planning on delivering a mostly final version of the intro on the fourth. We wanted to use the extra Monday and Tuesday to work on the outro and move on to some other pressing projects. This wasn’t a huge deal, since for the majority of the project, we have been planning a final delivery on August third — we just kept our plans mostly the same. However, we ran into quite a few compositing issues and didn’t get it quite a far as we wanted, so it was fortunate we had asked for a few more days.
Additionally, we decided to explore a few new avenues for color and backgrounds with the extra time, so what we actually delivered on Friday ended up being a lot farther from the final than we anticipated.
As for communication, there wasn’t much. Andrew checked in about delivery early in the morning and I told him we’d definitely be getting them something for their rough cut today. When we delivered it was at the very end of our day (hurray for West Coast being two hours behind!), so I told Andrew I would follow up with some planned changes to the cut on Monday.
Monday August 7th, 2017
On Mon, Aug 7, 2017 at 11:06 AM, Samuel Cowden wrote:
Andrew! Wanted to give you an update as to where we are and what’s still changing.
We’re reworking the colors to be more fun and bold. It’ll be awesome.
We’re going to go to a painted background instead of the image background. It gives us a little more flexibility where we need it, but will maintain the variety of styles.
There’s a lot of compositing that isn’t finished in this draft. Things like smoke being in front of trees when it should be behind, puppets not perfectly aligning with the camera movements, characters not disappearing when they should. Point is, we’ll have all this fixed.
This was just a test. We ended up having to move away from the car dashboard because of quality reasons, but we thought it was an even weirder and more interesting way to start. The tape will say “K is for Kids” on it and the screen will blend into the following scene a lot better.
There will be a puppet in the space suit.
Closing Title Scene
There will be many many puppets added to this scene.
Could we chat on the phone real quick today? Just give me a call.
We ended up talking on the phone the next day.
Tuesday August 8th, 2017
Andrew called me and I told him everything was on track for the delivery of the intro, but that we felt the outro would greatly benefit from two extra days. I told him that we could absolutely deliver it today if they needed it, but that I wanted to check just in case they had time to play with. Turns out, they had received some extra notes on their full movie edit and weren’t delivering another cut until Friday, so they did have two extra days! He also told us not to worry about delivering the intro until the next day (Wednesday), which just took some end of day pressure off our team and let Zac get home a little earlier to his 3-week old baby girl.
Side note: If you think a project will benefit from more time and you have the time to give to it, it never hurts to communicate that to your client and see if it’s possible on their end. Everybody involved just wants the thing to be great.
Wednesday August 9th, 2017
We delivered the final draft via a frame.io link. Every client relationship is different and, as you’ve probably noticed, this one is very relaxed, so the email was short.
On Wed, Aug 9, 2017 at 6:04 PM, Samuel Cowden wrote:
Andrew! I’ll get you a prores version tomorrow, but here’s the final draft! We really love it.
And Andrew responded.
On Thu, Aug 10, 2017 at 1:48 AM, Andrew House wrote:
It is incredible, exceptional work.
I cannot thank you enough.
Thursday August 10th, 2017
Not too much to report about the last day of the project. We sent Andrew a ProRes version in the morning and delivered and outro later in the day. There were a few other odds and ends that we talked through, but that’s the end of this story!
To achieve the final ‘vintage’ look we desired in the opening sequence, we treated the footage with Red Giant’s VHS plugin.
All the 3D assets were modeled in Cinema 4D and lit and shaded using Octane. These sequences were then rendered as PNG sequences and brought into After Effects for compositing.
Everything that’s frame-by-frame or illustrated was created in either photoshop or rough animator (on iPad Pros).
The nebula space background is an altered photo of a nebula from Hubblesite and duplicated many times.
We created the hyperspeed space tunnel using Red Giant’s Trapcode Particular plugin.
Here’s links to a bunch of drafts, boards, tests, whatever.
Hope you enjoyed this look behind the scenes. If I missed anything you would love to know, hit me up on Twitter!