YOU'VE READ THE INTERLUDE, so you know: simply put, workflow is the step-by-step process that you follow every time you create animation. It is your go-to guide. It is your savior when things go wrong, it is the ever-evolving and improving rubric for success that you depend on.
No? Still not sold on workflow? Fine, maybe this chapter will change your mind. We're going to walk through a shot and demonstrate all of the great cheats that repeatedly make an appearance in a professional's workflow. If by the end you are still not convinced, then consider this: before I started committing myself to improving my workflow, this shot would have taken me over a week to complete. With workflow? This shot took three afternoons without a single headache!
THE BEST WORKFLOWS START with a strong foundation. This means thoroughly planning out your scene. There are a lot of planning methods but the most common and beneficial planning tools are definitely thumbnails and reference video.
With thumbnails, it's important to your workflow that the drawings are very strong. Focus primarily on the body positioning and pose, and less on the staging and camera. You want these drawings to be your guide through the entire shot, so getting too caught up in the staging and direction of the shot at the thumbnail stage will get in the way.
Reference video is easy to gather on YouTube. Spend some good time finding as many related clips as you can. The more the merrier. You should also be creating your own reference footage. For this shot, the animation needed to be very cartoony. I've found the best way to make a reference video for really cartoony actions is to “puppet” your hands around and create a “sound effects track” by making noises with your voice. It may seem a little silly to be yelling and screeching at your desk, walking your fingers around like legs, but when you see how closely the animation was timed to my reference video I think you'll see the benefits outweigh the embarrassment.
We're going to go through a few cheats that are great for making sure your drawings and reference video are at your fingertips. We're then going to import your reference video as well and keep it handy.
01 - Cartoony_Start.ma
If you ever accidentally close this panel, get it back by going into any panel, switching to the thumbPanel you created, and tearing off the panel for easy use. Your thumbs are never far away…
ONE OF THE BIGGEST CHEATS of all involves timing your scene. In the old days, animators would use stopwatches to time their animation. By repeating an action over and over in their head and timing it with the stopwatch, they could write down the specific frame numbers and always have this timing reference close by. I find it alarming that now that we've moved into the digital age, this cheat hasn't been updated as well! We are going to use a piece of video reference created specifically to give us really high-energy timing to our scene.
Maya 2013 imports all kinds of movie files now, making video reference accessible and valuable deep into the production of your shot. With multiple codecs and formats supported, you will always have your video reference at your fingertips.
The movie file we're going to load was created to give a very specific sense of timing that could not be acted out. Hence the reason why I use my hands and sound effects (whistles and screams) to get a feel for the energy. “Acting out” your scenes like this, especially when they are cartoony-styled shots, is an invaluable piece of reference. Don't be shy, the more energy you put into this “timing reference,” the better the shot will turn out!
02 - Cartoony_Timing_Start.ma
If you go into the Panels > Saved Layouts > Edit Layouts you can save this panel layout as well, and get even more time-saving customization in Maya.
WE ARE ALWAYS LOOKING FOR ways to cheat our timing. Any visual cues we can give ourselves to help the timing should be employed whenever we can use them. We've already checked out how the timing can be helped by using an audio file. Let's see what we can get out of a little trick using FCheck.
If you aren't aware, FCheck is the tool that ships with Maya for playing back.iff streams (or any supported file format). Commonly, you watch your playblasts in FCheck if you haven't changed the output file format.
One little-known thing about FCheck is that you can draw on any frame using the right mouse button. And, if you hit play, it will actually ‘capture’ your mouse gesture for some visual cues. We'll use this to get even more helpful reference for the energy of this scene.
You can save this playblast with your gestures on it by going to File > Save Animation. Save it to a new folder and then load it into your Maya scene on a new camera for even MORE visual reference right at your fingertips!
WE KNOW FROM THE GRAPH EDITOR CHAPTER that stepped keys give us instantaneous transitions between values. When you key all of the body's controls in stepped mode, you could say that the resulting key frames are almost like “images” that you can retime at will to give yourself the best result.
Workflow-wise, blocking is the most efficient stage to be doing this retiming. Maya 2013 has awesome tools for retiming keys. We'll practice this by using both the brand new Retime Tool in the Graph Editor, as well as the Dope Sheet.
With the Retime Tool, our goal is to retime whole sections so that they line up with the energy of the beats much better. This is achieved easily because the retime tool is intuitive and rock solid.
With the Dope Sheet, a very underutilized tool in Maya, we get a top-level view of the distribution of the keys in our scene. It's very important to make sure you are hitting your beats exactly at the blocking stage. We'll use the Dope Sheet to make sure the impacts of the feet are right on the same frames as the sound, and the impact of the body on the ground also matches the audio we've recorded.
Many novices think that the best animators create perfect timing on their first try. Quite the contrary! The best animators are normally really good at quickly RETIMING and adjusting their animation. If you can go through a dozen iterations in the time it takes another animator to do just one version, you will be far ahead of the pack. Some might say it looks like you're cheating…
03 - cartoony_Blocking_Start.ma
03 - cartoony_Blocking_Finish.ma
If you right click on a retime handle, you can insert a key on all channels of the selected objects. This is a great, quick way to create a hold.
03 - cartoony_Blocking_Start.ma
03 - cartoony_Blocking_Finish.ma
If you notice in the top right corner of the Dope Sheet, there are buttons for the Graph Editor and the Track Editor. Maya makes it easy to switch between these common editors. Using these quick buttons will save lots of time!
AS WE MOVE THROUGH BLOCKING, it's imperative we are also sticking close to our workflow. Deviations at this stage can mean we get so off track that we can struggle to regain control of our scene. Indeed, as the scene gets more and more complex, it will be our workflow that keeps things manageable. In blocking, what tends to happen is that, as we add breakdowns, the silhouettes get muddy. Posing tends to sacrifice itself to timing as a scene comes together.
It's a great idea to continuously check your silhouette as you work. Everyone knows the good old “seven” trick (we'll go over that one too just in case), but this cheat gives you an even better way to check silhouette. Using render layers, we can set up a silhouette check that still allows you to add lights to your scene (which negates the “seven” trick), and also allows you to compare the silhouette with and without the environment geometry.
Keep a constant eye on silhouette through the blocking phase and your poses will not go astray.
04 - cartoony_Silhouette_start.ma
04 - cartoony_Silhouette_finish.ma
You can even unhide the set_GEO_layer and it won't show up in your silhouette checker. You always have to manually add objects and groups to render layers in Maya. You can change the sets, characters, lights… everything, and this layer will remain the same.
THE SPLINING STAGE IS SO TERRIFYING: it's very hard to know what you are going to get! When you are done adding essential breakdowns in stepped mode, most animators hold their breath, close their eyes, and click Spline in the Graph Editor, only to fret when the resulting mush doesn't carry any of the old appeal of the stepped version. Transitions seem slow and floaty, the character cruises “through” poses without any of the snap that you imagined in stepped, and overall it seems that all of your amazing timing choices have all changed.
This is fine and normal, but it won't do for production. Pros know that when you are going through your workflow, you can't have a step that basically leaves the success of your shot up to chance! And one of the most striking differences between stepped animation and keys that have been freshly splined is the absence of any moving holds.
In the last edition of How To Cheat in Maya we showed you how to do copied pairs. This cheat is an expansion on that thought; we'll start with a copied pair but this time in stepped mode. Then we'll use a fancy cheat I've devised to give just the right amount of movement on a character in a moving hold. It's the best kind of cheat – one that you can and should use over and over in your workflow, but more importantly one that takes all the guesswork out of a scary stage in your shot!
05 - cartoony_Moving_Holds_start.ma
05 - cartoony_Moving_Holds_finish.ma
Remember from Chapter 2 that clamped tangents give us flat tangents when there's not much change in value, and spline when there is a big change in value. We can use this knowledge to predict the results of going from stepped to spline.
If you go halfway between two keys, you get roughly 50% change, but for more fine-tuned adjustment, you can copy a key only a frame or two before your last frame and get a 5% adjustment. It's common to do a 50% copy and then a couple smaller copies to get a finely tuned result.
Converting your stepped keys to clamped when splining gives the most predictable results if you create your holds using this method. But change your default tangent to AutoTangent after converting everything to clamped to get the most help from Maya as you move on.
THE BRAND NEW RETIME TOOL OFFERS a brand new paradigm for manipulating and creating keys in the Graph Editor. We can make broad adjustments to timing, or in this case repeat the copy-paste method of moving hold creation in a snap.
Most professionals know that the best way to cheat in Maya is to find quick, effective methods of doing tasks that you must do repeatedly throughout the workday. And while Maya offers very powerful scripting capabilities, rarely will a script come in handy when it comes to something like a moving hold.
Instead we must use all of the new tools Maya has to offer to refine and perfect our workflow. In the end, your workflow should take full advantage of the aggregate improvements added to Maya every new version. To ignore these awesome new tools is to ignore cheats that could save your shot!
06 - cartoony_retime_Holds_start.ma
06 - cartoony_retime_Holds_finish.ma
Remember Retime Handles are not saved into a scene. You need to make your timing adjustments before you close your file or you will lose the handles you've created.
BY THE TIME YOU HAVE WORKED THROUGH your entire workflow and have made it to the polish phase, you are most likely dealing with an amount of keyframes that is unwieldy, to put it lightly. To make even the slightest adjustments can take hours of deleting, redoing, tweaking, and frustration.
Arcs are such an important fundamental that we need to make sure at this point in our workflow that we double-check our arcs look great. However, like I mentioned before, finalizing your arcs in the polish phase means navigating a rat's nest of curves in the Graph Editor.
Fortunately for us, the Editable Motion Trail tool is good not just for creating and defining motion, but also for keeping arcs manageable in this time of super-dense keys. We took a look at how to use Editable Motion Trails in Chapter 1 and Chapter 4 so you should be pretty familiar with creating them by now. But unlike in those chapters in which we created smooth motion, we will now be using them to finalize arcs among very dense keys. Maya 2013's Editable Motion Trails are rock solid and very stable now.
07 - cartoony_Arcs_start.ma
07 - cartoony_Arcs_finish.ma
Remember that an Editable Motion Trail is most useful on controllers that have translation keys on them. You get much less predictable results if you add a motion trail to a controller that is in an FK Chain.
Our final check of any arc is looking through the main camera. An arc may look great in perspective but until you see what it looks like to camera, you are not finished finessing.
THE VERY LAST THING YOU SHOULD HAVE in your workflow is adding final details like little bits of texture to your scene. This is called non-performance texture: these little details aren't contributing to the performance choices you've made. Instead they are making the scene feel full of detail and real.
With Maya 2013's animation layers, adding texture to your polished scene means that you can non-destructively experiment with ideas. It is crucial that you understand the importance of workflow at this point; animation layers need to be planned for very carefully. If, halfway through working on your shot, you suddenly decide to do some of the animation on a layer, chances are you will get horribly off track. Instead, you must either plan on your animation being done in layers (a walk cycle on the legs on one layer, for example, and adding torso movement on a new layer), or you must wait to add very fine details at the end.
By adhering to our strict workflow, we are safeguarding our scene against unforeseen problems later on. Animation layers give us an immense amount of control over the polished scene, without any destructive effects on the keys we already have.
08 - cartoony_Texture_start.ma
08 - cartoony_Texture_finish.ma
You want to create all of your textural “moments” on different animation layers, just in case you want to adjust the weight of each separately.
OZGUR AYDOGDU IS A 3D GENERALIST at 1000 VOLT Post Production, Istanbul, where he has worked on a variety of commercials for Coca Cola, Vodafone, Turkish Airlines, Danone and feature films The Ottoman Cowboys and The Breath. |
TELL US A LITTLE ABOUT YOUR CENK RIG. WHERE DID YOUR INSPIRATION COME FROM?
As you could imagine, the daily work in a studio might easily fall into a routine, especially when the works are mostly commercials. At this point, I believe in the importance of personal projects as a tool to try different things and develop yourself. I was looking for a challenging task to improve my rigging and character animation skills; where I would carry out each step by myself, starting from design, model, texture to rig and animation. I was already sketching some characters; and what I had in my mind for my project was kind of a fat character instead of a stereotypical cute one. I imagined him as a goofy character who would get into a series of funny situations due to his physical appearance. My aim was to try some new techniques and I experienced a lot more than I imagined throughout the project.
WE SEE ON YOUR INTRO VIDEO FOR CENK THAT YOU INITIALLY SKETCHED HIM OUT BEFORE PROCEEDING WITH THE MODELING AND RIGGING. DO YOU FEEL THIS IS IMPORTANT FOR CREATING ANY GOOD RIG?
I think it's an essential step, not only for creating a good rig but also for modeling and animation as well. It provides a quick end vision in your mind before you actually sit in front of the computer. It helps you better explore the character, its volume and expressions so that you can easily see the limits of the design or any other pitfalls beforehand. So, you start playing with the character to see how much you can push the design. But, apart from all these benefits I should admit that I enjoy sketching itself. I also enjoy looking at other people's drawings. Most of the time, I like the character exploration sketches in the Art of books more than the final 3D versions of the characters on the film.
DID YOU GO TO SCHOOL TO RECEIVE YOUR ANIMATION AND RIGGING EDUCATION, OR DID YOU TEACH YOURSELF THOSE ASPECTS OF THE PIPELINE?
I received my bachelors degree in computer science and did not receive an animation or rigging education. However, I was growing up in a family where I was very much exposed to various art disciplines – mostly due to my father's being a painter – since my childhood. I always had an interest in drawing and technology as long as I can remember. When I was a child, I used to fill my school notebooks with drawings and flip book animations like many other animators. Later, I started to try basic scripts with my first computer, a Commodore 64. Then I met 3D with early versions of 3D Studio on PC and started modeling any stuff that I see around. Throughout my university years, I pursued my interest in 3D modeling via various freelance projects. Upon graduation, I started to work in a company developing simulation software. But soon I understood that the work was not really for me and started to work on some personal projects mostly on modeling and rigging, at home. With one of these projects I applied to a post production company and since then I have been working as a 3D generalist where I mostly do character rigging and animation.
But, apart from these, I am kind of obsessed to find solutions when I come across something which I do not have much clue about. That's why I read, watch and search a lot. Whenever I see a fascinating piece of character animation, I dig into it and watch over and over again to understand timing, acting and the deformations of the character. And I enjoy trying new things that would be challenging for me. Maybe these contributed to me developing my animation and rigging skills and teaching myself throughout the pipeline.
WHAT ARE THREE QUALITIES YOU FEEL ARE NECESSARY FOR A GOOD CHARACTER ANIMATION RIG?
The first thing that comes to my mind is stability. The rig must be robust enough in terms of its technical structure. The IF/FK and dynamic calculations, transferring animation keys, caches, reference updates, new features, and component additions during the production should all work properly and as expected so that the animators could focus on their animation without having to deal with technical problems which they are not responsible for. A robust rig would maximize the efficiency.
Flexibility is another important quality. The rig should be flexible enough so that the animators would work freely and reach poses that they have in their mind easily. Even, the flexibility of the rig might give them an opportunity to try a different pose that they did not imagine at first. By this way, the animators would not be limited to the rig but the rig itself would push the animator. For instance, the flat and soulless look of CG characters due to their symmetric structure seems to be a problem for me so I tend to solve it by adding secondary deformation points on the face which let animators make finer tweaks on the pose and get more natural expressions. Besides, the ability to squash and stretch adds a lot. There would always be some other things to try.
To me simplicity is another important aspect of a good rig. It should not take a lot of time for an animator to achieve a nice pose after digging into numerous controllers. In fact, putting too many controllers may sometimes freak the animators out, and it can even become cumbersome to animate. The rigger should take the story and concept drawings into account and should not define controllers more than needed. I also prefer making my controllers visually appealing and colorful, which make them shout like “come play with me!”
AS AN ANIMATOR, YOU ARE ALWAYS THINKING OF THE MOTION YOU WILL GET OUT OF A RIG. HOW DID YOU DECIDE ON THE CONTROLS YOU PUT INTO CENK REGARDING ANIMATION YOU WANTED TO DO WITH HIM?
As he does not have a body, he had to give out his mood and feelings via using his head only. I was looking for a way to be able to easily define line of action on the head, that's where the IK controls come in. The free deformations like he has in the eyes and nose also adds to the candid expressions I guess. The shape of his head when he bounces was also important as he would move from somewhere to another by bouncing up and down, so I thought to keep this motion as realistic as it can be. Since the character has a very round design, I added the ground collision feature, making his bottom flat and bulge when he sits on the ground, which makes the audience feel his weight.
DID YOU ALWAYS PLAN ON MAKING HIM A FLOATING HEAD, OR WAS HE ORIGINALLY A FULL BIPED?
To tell the truth, I thought him a fully biped at the beginning. However, it seemed to me that it would take a lot of time for me to incorporate the level of details I imagined into the full body. So, at some point, the project kind of stopped for me. Then while I was watching Miyazaki's Spirited Away for the hundredth time, when I bumped into the three floating heads in the Yu-Baaba's room, I thought why can he not be a floating head at all? I would still be able to tell a short with a head and it might even be more interesting to watch. But I made some changes in the design, for instance he did not have a helmet when he was a full body but the sketches which I made with the head looked too neat for my taste so I added the helmet to corrupt the symmetry and tried to make the design more appealing.
I imagined a series of funny events that he would experience thanks to his physical shape. Like, he was trying to get rid of a fly on his head, but he cannot as he does not have hands. He bounces up down, gets eager and angry, his head falls down and the fly seems to be getting all the pleasure that it can out of this; or I imagined him while listening to an iPod, then one of the earphones falls out and he tries all his best to put it again into his ear. And he cannot actually. I'm looking forward to see him in these hilarious situations.
ANY ADVICE FOR COMPLETING A PERSONAL PROJECT OF THIS SIZE AND QUALITY?
For rigs, the quality of the result is directly proportional to how well you organize the project from start to end. I can recommend the philosophy of “divide and conquer” as a technique for handling a large, complex problem by breaking down it into smaller, solvable problems which in fact can be applied to any other issue in real life. In my case I layered all of the deformations, which helped me focus on each feature individually. I had my rig hierarchy and layers clearly sketched on a sheet of paper all the time, because the deeper you go into the details (which in fact make the difference) the more likely you get out of line and lose control on the project. This might also happen if you are too much focused on the project for long days. I find keeping my mind always busy with the project in daily course (in the bed, on the subway, etc.) useful but sometimes you would need to move away from the project for a couple of days to refresh your mind. This would help to see things differently and you might even spot a problem in a second which did not take your attention before. Eventually it's really a challenge to balance things as you don't actually have a deadline in personal projects. I tried to overcome this by setting deadlines to myself and trying hard not to be too self-critical.
Last but not the least, is that you should really enjoy what you are doing. It takes a lot of work and real passion to achieve in this field. The pace of development and change is incredible and new things to learn always fly around. One should be eager and passionate enough to keep up to date and then be able to make a difference. This could only happen if you really love it I guess.
HOW DID YOU DECIDE ON WHAT FEATURES TO PUT INTO CENK'S FINAL DESIGN? DID YOU LEAVE ANYTHING OUT?
Although Cenk has been taking attention mostly due to its rigging features, I always had the idea of animating him in a series of vignettes to tell different stories. That's why the features I had are the ones that I would like to use when I animate him by myself.
To fulfill the needs of an exaggerated style of animation, I added IK and FK controls with volume, preserving squash and stretch. Since there are a lot of dynamics going on around the face and on the helmet, I decided to add a geometry cache importer/exporter into the rig. For the mouth, I added a sticky lips feature as I had a couple of stories where he was supposed to be eating. It helped me get some nice smiling expressions, too. Other than these, I was planning to add a camera-based lattice deformer on the overall mesh, which would let me enhance the silhouette of a particular pose by moving the lattice points. I hope I can do it next time!
WHAT ARE SOME OF YOUR HOBBIES OUTSIDE OF ANIMATION AND RIGGING?
Photography is kind of a passion to me. I take photos of artworks and portraits of artists to be used in their exhibition catalogues and art books. But actually I love street photography and taking photos for myself, reflecting my confusions and documenting lives around me. I enjoy sharing them on my blog, which keeps growing as a visual diary of my personal life. I am interested in black-and-white film photography. Hanging out in the dark room developing and printing photos literally takes me out of this world. Besides, I believe that the art of photography is a very influential medium for anyone dealing with CG, in terms of cinematography and lighting.
Apart from that, I am interested in watching exhibitions and seeing what other artists are doing. I might say that cinema is another passion for me.
WHAT IS YOUR FAVORITE FEATURE THAT YOU INCORPORATED INTO THE FINAL RIG?
I love the IK controls best. I think that the IK spline applied on a full mesh with a combination of convincing squash and stretch really adds value to the character. Think of a head shaped with a curve, when you make it concave towards the audience, the character has a timid expression who is not willing to communicate. And he turns out to be an extravert person when you make the curve convex. I can give an instant clue about the mood of the character very easily by using this feature.