Tuesday 16 September 2014

A shade of watercolor:- crafting Underground part 1




Greetings September!

I am not a designer so i never had much hopes on creating glamorous and perfected pictures, doing it in some style with an sleek and clean approach.  I`ve graduated in fine arts, took painting classes and was a child of a sculptor (my mother) and a painter (my father).  Perhaps that is why since i began working with 3d in those early computers were subdividing a model`s geometry could be cpu torture.

This was my first contact to a 3d graphics package, the first 3d Studio (dos)


 There was no Ambient Occlusion or  Global Illumination (both tools today provide a more accurate light scattering=more real cg). So i ended up making clean, expressive shapes with a very low polygon look. I just wanted to be able to tell a story without worrying  with the realistic aspect of it. Back then many of my interests revolved about Japanese anime and that was when i realized a great part of their expression is simplification, stylization and investing on storytelling. That was the comfortable side i was trying to work on.  So when having to revisit Alice (my old college comic book obsession), i knew that some decisions had to be taken. It should look and feel real to my drawing expression but also be faithful to its computer graphics nature.

Rendering. courtesy of Bill Altenburg

The Alice comic was more different in many aspects, it revolved around the later years, when the character becomes a bad ass, it is more like a space opera  with a  Jean "Moebius"Giraud comic book vibe in it.
So i had here a huge challenge, re adapting the story of 19th century little girl, introduce an alien world and have some visual cohesion between the two. The easiest guess is to paint a  Steampunk laye in the nooks and crannies of the Vorpal Spaceship. That could work but it isn't a way to make cg look interesting (and i knew realism was out of the question). One guy building models, sets, texturing them and overall trying to pull some realism, that would be like building a castle on my own. But i knew that i could build a small house, that was a less impossible quest-Stylization.

The Tenniel  shader test

The second try was of quoting directly the "Wonderland" illustrator, John Tenniel by creating shaders that resemble the inking patterns from the book. The result was quite interesting but that would be a visual burden for the film. Imagine a cross hatched look, flickering in your eyes for 20, 30 minutes. Also it would be really difficult to have cross hatched characters look good while in shadowy spots, they would blend with the dark and everything around. Came to mind being invited to the Portuguese premiere of the Marc Miance film Renaissance(2006), and how the darkened theater became a burden over that extreme bright/dark of a film, for 2 hours,when every glimpse of a silhouette and contour becomes a tiring visual exercise (also a headache). Loved that film  and recommend  watching it, but at that theater in pitch black it`s hard to focus on the subjects, sets, action...

The answer i was looking for came suddenly and curiously by researching the other (and controversial) passion of Lewis Carroll, photography.  Well, lets get this elephant out of the room, most of his collection involves naked or semi-naked children.  There was a naive, but somewhat dreamy quality of his water colored photographs that felt right for this movie. Much of his tone and vibrant colors are referenced in the film, you can´t say what is painted over (or recreated from scratch) in his photos, the same for these shots. Many composite shots go all the way as having 3d graphics retouched and color toned just for this purpose. That was my little unexpected reference to Lewis Carroll.  As for that elephant, i get often asked what i think of Carroll, was he a pedophile? First of all  i never met the guy. Also the 1800 were somewhat weird, specially in the tendencies of early photography, there were no boundaries and culturally it was no tabu to explore some dodgy subjects like post morten photography, ghost pictures, cryptozoology sightings and in the midst of what could be made, there was this niche where Innocent pure children were inserted on fantastic and romantic scenery. I think its safe to suppose that many adults  identify themselves with the child they once were, when everything seemed magical, possible! And naked children apart, i feel that Carroll was talented in many ways, his Alice in wonderland is like a Dracula or a Frankenstein, an immortal icon evolving in our collective imaginary. Concluding, what his author did or did not in life is now a blurred mess of guessing games that does not invalidate a masterpiece.

Lewis Carroll above and Underground influences below

To make the image look more stylized  a combination of two source composites was used, first one that had a more flat painted look is created using filters like median blur, sharpen edges and some rotomaks. This is blended with a normal processed image from that shot that preserves the texture. The final result accentuates shadows, edges and highlights while keeping the details of the second image. Overall i was trying to make 3d graphics sound hand-retouched without using cheap toon filters or  brush stroke simulations. The visual needed to evoke some painterly qualities in a fluid cgi while avoiding more gimmicky results.

The final "merging"process  
The overall shot

As this is an isolated chain of this huge process of making the film, it was a very important step that gave me visual cohesion and a tone to keep going.  Researching more about the author (Carroll) was useful for a  more three dimensional scope of his work and imaginary. That saved me from probably using a more drastic visual treatment that would n`t add  anything special on the project as a whole.
I will,  in further posts, be documenting other challenges i came across.


PS:Once again i thank you all for the support, and because at the moment i am juggling the film, the masters thesis and my daily job there hasn't been much time to post as often as i wanted.

Tata
   Jules Raimes

Next: Gardens, hallways and corridors  crafting Underground part 2





Thursday 26 June 2014

A year ago Alice Liddell-Character Featurette


A year ago i made a drawing of the child Alice, since then her world grew and grew...Almost forgot how things changed in one year. Here is a sneek peak on some of the development stages i went while creating the Character of Alice. Many ideas didn`t worked, abandoned was the concept of using a NPR sketch shader and  2 models were scrapped before one that looked interesting.  hope to explain it all in detail on a future diary. 
Jules Raimes

Tuesday 17 June 2014

Alice teaser trailer 1

And now i present you the teaser trailer:




June 2014 status

Since we have been more active on our Facebook page, it comes the time when as an archiving solution all the development gets stored on the Blog.

Alice and Leopold meet the Queen of Hearts

The film project is coming along well, with many changes on the script, mostly making the scale of the film bigger with the tools we have at hand. It lays now on me on the direction, animation and modeling side. But also on the work being made by the Sound Designer Christian Tanto. As a collaborator the talented Bill Altenburg is giving me a hand on some hard surface modeling, like robots, set pieces or props. His work has helped me spend some time on the animation or designing the shots and lighting. What i have to say after a rough 20 minute unfinished film is that "Alice`s Adventures Underground" clearly owes more to an indie live action guerilla film than to an animation project. The mocap part adds to that immersion in the virtual Camera work. It is a solution that gives the editor (Mafalda Relvas) all the camera angles (in animatics) I figured out for each sequence, after she puts the scene together I spend some time refining those shots for rendering.
Meanwhile at this time Christian is mixing sound effects over the XML scene and some of the trimmed shots get rendered. This is a tricky solution, but because we couldn't get sponsors for a renderfarm our personal laptops and desktops have an extra nocturnal life, just like cats! This also needs another crazy solution of having the shots travel back and forward trough Dropbox or We transfer (used more for the Blend files). The usual routine:  I prepare the sequence  and upload the Blend file, Christian downloads it and prepares the batch rendering. The rendered sequences go directly on Dropbox for me to assemble the final compositing shots.
These composited shots vary from 2 layers to 30 or 40 depending of what I want the image to reflect. In the desert scenes, when we meet the family I wanted to give the film a very warm technicolor vibe like we see in Lawrence of Arabia but when Alice is lost on the Vorpal spaceship everything is mysterious, cold and foggy. But when Alice becomes aware of her goal on this world she is not scared anymore, the way the world is portrayed changes drastically.

The desert sequences where we meet the Liddell Family

Then the film is not being made in chronological order but by the sets. Each one of these places have different requirements, some require heavy compositing work, others work better when rendered directly with  a low depth of field and more closed shots, angles, lighting etc. Every camera trick I use today on digital animation are somehow derived from my experiments on an old sony handycam back on the 90`s.

Then come some disadvantages of  working on many different sides of the film. The one I really miss is not having much time to draw.  Because every shot is composed  directly on "camera" that excludes the need for a very complex storyboard. What I usually do is a quick sketch of the direction I want to have in that specific area of the film. It then becomes 3d, which tends to change the way characters and sets are displayed.


sketch for the doll house scene
3d animation is more out of control in terms of stylization than in 2d . Because I chose this style, to get a fairly quick reading of a rendered shot, everything needs to be mediated in terms of luminosity, framing and focus and we already know how 3d tends to love the eye candy of detail. So without compromising style or detail hidden positive and negative light sources are located on areas that take advantages of  more expressive qualities in the shapes. Sometimes a character gets lamps glued on him just to keep him "there".  


The shot before compositing, when lights get positioned. 

Some live action principles do carry to 3d animation, but I also think that every form of representation was once a drawing, even one that never left our mind. One made of brain cells instead of charcoal, painted with creativity instead of pigment. That is why the solutions need to be an extension of the problem. In this case getting expression trough textured surfaces and luminosity while keeping the freshness of a drawing. 

Alice Underground may not reached all the goals I had imagined, surely there were some supports that lacked and could speed up render time, provide an unique soundtrack or even finance our long work nights .  But sincerely the project becomes stronger by some of its limitations and while we want it to be a good film, I wouldn't´t rush it by any means. Underground is now targeted to be completed on the winter just to give us that margin. I would like to thank everyone who supported the project, the sponsors IpiSoft, Faceshift, Blender Foundation (for truly believing that dreams are better open source) and the fans who curiously await news of this strange Alice in Wonderland adaptation. 
Jules Spaniard Raimes (Julio Ramos)

Thursday 28 November 2013

Leopold the White Rabbit

Leopold`s name was an homage to the involvement of prince Leopold (Queen Victoria oldest son) in Alice Liddell`s life, curiously is the name of one of her future sons. The "White Rabbit" is actually a Plush toy who becomes alive through Vorpal technology, so the Rabbit needed some background research on the dolls of the late 1800.
This first toys depicting animals were kinda freaky, some had real hair and some appear clothed. Also que problem with design was making it cute while respecting the period. You can check out this Pinterest link for some Retro/Vintage/Primitive Bunny rabbits

A Rabbit plush toy in early 1900
So this challenge was not to make just a cute plush toy, but to make a real character that needed to convey sympathy, be true to its time and also needed to look like a very loved toy. The body needed to look used and over cared, full of patches with different patterns as if he passed from brother to brother until he came to Alice. I also wanted to use the less amount of plush possible, to avoid looking like an "living, furry animal". He should look artificial, man made and then, progressively become a living character in the viewer`s mind.
 Sketch/Concept used for the final Rabbit
The next step was visualization, a real doll was needed to use in our capture sessions, sometimes as a reference, Alice spends about 60% of the story interacting or looking at Leopold. So came the need to contact Ana Santos from the Store "Coração de Pássaro" who provided the skills  and inventively created the life size Leopold for us to use. Then, to have the real deal meant that i could copy the textures using Ana`s fabric patterns (or even the seams).

Leopold prop by Coração de Pássaro
Because the digital Rabbit model was already being used in animatics, a small update needed to be made, in this one i blended the good features of the real Leopold with the first concept,so the details went to the model but i preserved the volume from the original,  at some point in time there will be a robot inside.



The "Wunderland rig" also needed to capture a different type of anthropomorphic body, Leopold was the first body in which i had to modify the actual armature to accommodate the character structure (large torso, small and wider hips). 


The next upgrades will require some cloth dynamics for the ears and in some spots of the body. The goal is to achieve a more sacky motion on the limbs, but at the same time i will be continuing with the capturatics (yes, a mixture of motion capture and animatics). 
I am very proud of this character because he is the marriage of two worlds and various crafts.  Thank you Coração de Pássaro for helping with Leopold`s birth.


More news as soon as possible.

Jules





Saturday 6 April 2013

The Wunderland Rig 00.1

What is the Wunderland Rig?


The input mocap Rig in yellow and the the keyframe rig in black


The Wunderland Rig is my journey to develop a tool that can be useful to blend motion capture files with a keyframe animation.  Working as an animator for about a decade, my ambition is to gradually work with huge, or even massive projects. I can understand the frustrations of having to much to do in one scene since in Alice I have about 5 rigs/actors moving at the same time in only one set.

On the other hand, I want to leave the animator's position for a while and use the actors movement to sketch the action. The finessing would occur later  in Keyframe animation. Even if I have scenes where a character runs in motion capture and then flies in the keyframe. To "Blend" those actions would be very useful.

The capture process in Ipi


Because Alice "Adventures Underground" is almost a "no budget" project, the affordable solution for mocap was to use both kinects (depth sensors) and pseye cameras. As capturing software, I used Ipisoft Studio for body capture and Faceshift for... the face. Both need to be captured in separate and then brought to Blender. The Wunderland Rig will need to interpret those "signs" and translate them as a motion sketch for me to continue my work.


Then, and since I'm not a coder, I had to come up with alternative solutions and some constrains and use some "MacGyver" skills in order to make a working rig ready for the film. The rig's name was Hybrid rig, but, as an homage to it's main project, I decided to call it "The Wunderland Rig". Why Wunderland is written like that? Because my Alice story arc (my expanded universe) features World War II and Hitler's involvement in the Vorpal's technology. Hence Wunderland.

On The Muscle system
My muscle system is nothing much than another detour in the Blender developer's wish list. It works by using the shrinkwrap modifier (only Blender geeks can handle this part). But the Shrinkwrap can sometimes perform with not so perfect surface snap movements. Ex: when the skin slides over the muscle, the shrinkwrap is not ready for a smooth sliding. The function detects that the skin is going into another vertex of the muscle and it snaps when and if you're using a lower poly mesh. I realized then that using a smooth modifier on the muscle affected zones could ease those movements. The muscles can also be simulated by using softbodies on them, that will have cool effects on the skin, specially if you`re animating a larger creature.

Inspirations
My experiments in this way of creating muscles in Blender came from two sources: first, I loved the Blenrig that J.P Bouza made and shared, and second, Weta digital tissue solution gave me some ideas about having ways to simulate skin and muscle wiggling.

Final toughts and the future of the Wunderland Rig
Well, at this point I will be cautious since the Wunderland rig is still an half achieved dream. I am an animator, a  3d artist and a storyteller. I have built this system mainly out of necessity, to shorten the gaps in motion capture and ik animation.
I think that there's nothing wrong in using both systems. The purists may say that motion capture is a cheat, and it's not animation. In my mind (and in my heart) they may be half right and I think it's not a bad thing to use actors to input models. Even in today's hollywood's standards, most of the captures are tweaked by animators, fingers and faces are corrected, but for marketing reasons most of the times only the actor/performer gets the higher credit. But I am very proud of using a mix of both techniques. I know it's the best part from both worlds.

As for the future of  Wunderland rig, an interface needs to be coded into an addon to allow the switching on and off of the ik joints from the mocap input rig. If I want to have an upper torso mocaped and the legs keyframed, I have to select the joints, go to the constrains tab and switch from there.

Also, I can see huge improvements on the muscle system, but those still depend on the future of Blender. The improvement of both the shrinkwrap modifier and the implementation of a gpu subsurf would make it more accurate and faster to work. This way of deforming muscles depends on mesh detail and softbody simulation, the higher your skin, the best it can interact with your muscles.  For now, many characters in the film are heavy clothed and have most of the  muscles turned off.