The Albert Project
Free Bonus Pack for fStretch*
The complete Albert project with all the scene files** is
exclusively available to fStretch customers for free.
Buy fStretch today and grab your own copy !
exclusively available to fStretch customers for free.
Buy fStretch today and grab your own copy !
|
Table of Contents
|
|
** The rendered frames were much too heavy to include them all (7gb in total). Only the last rendered frame with all its passes has been included in the project. This is not a problem since one can already see everything that's going on with only that, and it is also possible to re-render any other frames if needed.
Finding the Idea
|
Christine |
We had to find an idea according to our goals. Our goal for that project
was to show the deformations that fStretch 2 could do on the model. So
we thought of grimaces, of course. But we didn't want to make the
character do only a series of funny faces without any purpose. So I
thought of storytelling: He could do grimaces because of something
logical. I thought he could have bitten in a lemon, and then he would be
doing grimaces because of the bitter taste of the fruit. Or he could
see a bee around him and be frightened. Or an apple could fall on his
head and hurt him so he's getting angry... But all those ideas didn't
have any link with our character's nature since we have decided the
character would be Albert Einstein! Then we had the same idea at the
same time: use his high intelligence and eccentricity! We linked it to
the fact that this animation had to announce the release of fStretch 2.
So the spot is beginning with Albert yawning, getting bored by the
dullness and silliness of life around him. He's hoping for something interesting to come up.
Then bounces in a 3D text announcing the new version of the plug-in.
Albert turns, reads and reacts happily. He can't wait to check the
deformations he has on his face. Then he has so much fun doing faces
that he can't suppress a good laugh while being caught in the act by the
spectator. Then he catches back his seriousness, but ends by giving us a
smile with a quick seductive kiss. Matthieu and I believed that this
idea was packing all that we wanted to put into this animation : the
link with fStretch 2's release, Albert's personality and the plug-in
deformations.
The Concept
|
Christine
|
When I was in CEGEP, I had an optional class of astronomy. I've always
liked to hear about universe and planets. In that class, I've learned
that Albert Einstein had developed a very controversial theory that says
that the universe is in constant expansion. From that idea, I wanted to
do my student film about a 3D Albert explaining his theory at a black
board. But it would have resulted in a too long short film so I finally
chose another idea. I had done some sketches of Albert that year (2005)
and I've put them away for a while. But the concept of Albert and his
amazing theories stayed in my mind. When Matthieu was looking for a
cartoony character he could use for his project, he borrowed all my
sketchbooks. He started to look at all the pages one by one of all my
books and I was so surprised when he came up with my Albert's sketches. I
was so thrilled and flattered that he wanted to model my character! I'm
still very excited when I think of how lucky I am to witness the rising
of one of my character from a sketchbook to the computer screen and
furthermore, to have the chance to animate it myself!
Modeling
|
Matthieu
|
The modeling process was pretty straight forward. I knew pretty much what I wanted and I had Christine's sketch as a reference. I've also been using some references from Albert Einstein pictures but those were almost not used at all. This was primarily a cartoon project so I didn't want to put too much realistic wrinkles and preferred going for a not too old Albert. I preferred having a neutral pose with soft wrinkles and intensify them when the face deforms. The goal of the project was to show skin deformation with the aid of the plug-in so I didn't consider the hairs that much. I also didn't want them to take too much attention so I though that the best choice was to go for plain geometry. Even though it is not necessary at all, I felt like I would use the most out of the plug-in and drive refined displacement maps that are turned on in stretched and squashed area. This is not essential at all but it add another nice level of details.
Rigging
|
Matthieu
|
For the facial setup, I spent some time looking in the mirror to see how things deforms and how I could make something that work well while not being too complicated. Also, this was only for the marketing shot so I didn't want to waste too much time on details and polishing, as long as what we get in the end works well. Since I wanted to show that the plug-in could run on heavy geometry, I wanted this facial rig to follow this idea. It ended up having 68 000 triangles (or 32 000 quads) which is quite a lot. Since I never rigged a face with that much resolution, a little bit of R&D needed to be done. It wasn't long to realize that it was just impossible to paint the weights of such dense geometry. I then thought of using a lower geometry to paint the weight and make that geometry deform the high res mesh. At first I got a bit discouraged since all the techniques I tried weren't fast enough. I tried using poly and nurbs surfaces as influences in a skinCluster with the option 'use geometry' and played around with the settings but even the fastest result I got was pretty slow to me. It would make the entire rig too laggy and that I something I really wanted to avoid. A wrap deformer also couldn't be used at all because of it's speed. Not to say that both of these two techniques are also pretty bad if the rig need to be scaled in production ! It's just then that this very simple, fast, elegant and obvious idea came to my mind: just do your painting on a low res and copy the whole skinCluster and its weight to a high res ! In the end, the rig was quite simple. A standard bone based rig without much fanciness with the weight being copied from the lowres to the high res. On top of that I applied fStretch and only used it's tension deformation module. I ended up having only two sculpted meshes plugged in it: the Stretch and the Squash meshes. A little bit of tweaking and painting was done in the attributes in order to get everything to work well together. I'm very happy to say that this step was pretty quick - especially for the result it gives. No blendshapes were used at all. This is exactly what I aimed at when I first thought of the idea behind the plug-in so I was very pleased with how things turned out. I also wanted to refine a little bit so I decided to make a little setup for sticky lips - which is very subtle in the end. I also added up some 'average vertices' deformers on the eyes regions to get a smoother result. You definitely can get away without these last steps but it was just to polish it a bit. The nice thing about all this setup is that it's very easy to automate it. In the end you would only have to sculpt the squash and stretch meshes accordingly and do some minor tweaking in order to get rigs of very high quality. After the animation, absolutely no deformation polishing was made. The final results came without any per-shot sculpting or anything similar.
 |
 |
Animating
|
Christine
|
I chose to do it at a 24 frames per second rate because it's a common
rate. So first I started by collecting references: Photos of expressions
that were inspiring me, the closest possible to the ones that I was
planning to do. Of course, I shot myself with my camera because I
couldn't find every poses I wanted on the Web. But not only that; act in
my own video helps me to feel the intentions and the motion, to
understand it better. I asked a friend to do the exact same thing, to
see if there was some expression I would have missed. Then, I was ready
to do my blocking. I would say that the poses are what defines an
animation the most, and the timing. For this project, I chose some
specific poses to fit with Albert's personality. At the beginning when
he's bored, for example, of course I could have put his head and cheeks
down, like we see most of the time to express boredom. But Albert is
eccentric and independent, so I put his head up, like if he was defying
boredom. Then I made him yawn and made him take his time to do it,
because he's a little bit arrogant. Well... not distinguished, I'd say.
The third funny face he makes is a special request from Matthieu. He
asked me to plug it somewhere in my animation. It's a face one of his
ex-colleagues was doing. Notice that Albert doesn't have eye-contact
with the camera until he's caught in his act while laughing. From that
moment, however, he never breaks the eye-contact with us. At the end,
the little kiss he's doing is a comical addition Matthieu asked me to
plug too. It's like if Albert was saying : 'Aww, come on, I was just
having fun, don't judge me!' I had very tight deadlines to deliver that
animation and I had a full-time job at the same time, so I couldn't
refine it like I would have liked to. I am very glad that I have had the
chance to animate a character I once designed and I am very proud of
the result.
Look Development
|
Matthieu |
It was very fun to do the look development on this project. I really had all the freedom to try any different styles and to go with whatever I decide. All the screen shots above show the artistic iterations of the look development process. All the images were rendered using a misss_fast_skin_maya shader in Mental Ray. Stuff like displacement, bump, textures and compositing is on and off depending on each iterations. The research of the look was done at the same time as putting along all the elements together and trying different poses. I think it's something very enjoyable and satisfying to start with something that looks like garbage and to push it the to the best you can. Once you're at the end you come back to where it was when you started and you really see the difference. When you arrive near the end and you have the chance to take some time off the project and "forget" what it look likes, you can come back to it and really see what's missing. I stopped working on it for a couple of days when I arrived at the 9th images you see above. When I came back I had a fresh vision and I was able to push it to where it is in the 10th.
Texturing & Shading
|
Matthieu
|
The texturing was very simple. This is a cartoon and they normally don't have much details. I wanted the texture to reflect that. I flooded the face with a skin color and added some red in the mouth, eyes and ears area. I then added some freckles and altered the overall color with some yellowish and blueish tone here and there to "break" the color. Since all of this remained very subtle and erased I got away by hand painting everything. The shader I used the most is Mental Ray's misss_fast_skin . It's a very decent shader that get the job done without much tweaking. Some mib_color_mix nodes were used in the surfacing process in order to have a more accurate control over each element. Since this project was for a single shot, I didn't feel the need to get the most out of the shaders inside Maya and I decided to make a lot of work in compositing.
Lighting & Rendering
|
Matthieu
|
For the lighting setup, I used 3 lights. A key light, a fill light and a rim light. This is a very standard setup and it always works very well. The render was done using Mental Ray. I used a couple of render layers. The first one was used in order to split the specular from the beauty. The second one was a puzz layer in order to separate each elements of the face into color mattes. The third one was an alpha that highlighted the mouth area, that way it's very easy to adjust the specularity in comp to have a wet look in this area. And finally, the last one was a beard alpha to get a high contrast of the beard from the face. It sometimes is a little bit long to tweak such effects when you only use the shaders. When you have it on a separate layer, it's really fast to tweak it afterward (in compositing) and when it's only for a shot, you don't really have to think of something that have to be solid enough for a whole production. In addition to render layers, I also used passes. Those were very standard: an ambient occlusion, a zdepth and a matte to cut the silhouette. With all those elements, I had everything I needed to control the overall look in comp. I also used an HDRI map for the reflections and to color/light the scene with some final gather.
Compositing
|
Matthieu
|
Doing the compositing of Albert was really fun. Unlike surfacing and rendering it is something that can be done in real-time. I first mixed all the passes accordingly: take the beauty pass, add the specular, multiply the ambient occlusion and create depth of field with the zdepth pass. Once this is done, I added the beard pass in order to get a good contrast on top of the skin and I used the mouth alpha pass to select the mouth area, extract its whitest white and put them even more white. This way, we feel much more that the mouth and tongue is wet. This is so long to tweak inside Maya that I prefer doing it that way. With that technique I get much more control over the final result and it's very fast.
A little glow was added on the image by extracting the whitest whites. I also extracted each channel (RGB) and offseted them to fake a photography phenomenon which add very subtle realism. Some light rays were added here and there. A very low lens flare was added on top to bring some area of the image brigther. On top of that, a little noise was also added. That's pretty much it for the big lines. Much more tiny adjustments were also made. Those can be seen by downloading the scene file.
A little glow was added on the image by extracting the whitest whites. I also extracted each channel (RGB) and offseted them to fake a photography phenomenon which add very subtle realism. Some light rays were added here and there. A very low lens flare was added on top to bring some area of the image brigther. On top of that, a little noise was also added. That's pretty much it for the big lines. Much more tiny adjustments were also made. Those can be seen by downloading the scene file.
(c) CGADDICT - all rights reserved