This month I dabbled in the second course of Visual Effects in Full Sail. The focus was mostly on Fluids both in Maya and Houdini. The past month was my first time really working with fluids and more complex dynamics in Houdini.
The third project for the month was to make a liquid simulation with FLIP Fluids in Houdini.
The reference used was of a bottle pouring water into a glass. While observing the motion, the liquid can be seen hitting the bottom of the cup and wrapping across the surface until it reconvenes on the other side. By then, it curls onto itself towards the center, causing a short and stocky wave.
I tried doing the same motion observed in the reference, and had to model the glass with a curved bottom as to direct the particles smoothly instead of the harsh displacement a right angle was causing with a previous model I had been working with. Soon after starting to test different settings, it became obvious that the cup was not being filled; a search had me going back into Maya and thickening the model by selecting the outer faces and scaling appropriately. Later, for rendering, I used the original thin cup and hid the collision mesh so the effect was seamless. I also used an emitter with initial velocity on -X for that curve characteristic of fluids being poured from a container and most movements in nature.
A particle separation of 0.02 and the use of only the basicliquid shader with refractions and reflections enabled allowed the fluid to gain the visual appeal of water. An environment light with a map provided the lighting on Mantra - PBR.
This project was based on a chain reaction machine with a ball following the track projected. Done in Houdini with the purpose of conducting a research about RBD Objects and their different types of solvers; and how to set up the scene and node chain for the simulation to work smoothly.
The projected path is the ball falling down the platform and rolling down the helix to fall onto the ground.
First thing to do was import the assets into the scene, make them static and RBD Objects respectively and go on the AutoDopNetwork to adjust the solver type to Bullet with Concave Collision. Because of the complex shape of the helix, I had to go into the sphere's physical attributes and adjust them to make it stay on track without using any constraints.
Voila! Running the simulation in real time finally gave me the results I wanted and that seemed like a natural reaction in real life.
This month in Character Animation has been very informative and eye opening in regards to constructing a scene optimized for animation and creating your own little story to have your character act out. I consider my work to suffice based on the experience I have with animation, though I wish it was better. I had a lot of fun with my key poses, especially making my character express emotion with their facial features which I think translated well in the end result. However, I recognize that my anticipations, breakdowns and timing need quite some work. I made the mistake of not roughly timing when I had my key poses and main moving holds already keyed and had to do so when I was almost at the end of the pipeline for the video. Overall I like the direction my animation was heading, but need to put more work into it.
The images shown above are final renders from the last project in Visual Development. Spent over half of the month working on the different aspects of it. The blue gear was textured using real-life pictures altered in photoshop and applied using Maya after laying out the UVs. The file texture for the box was given in class, along with the blood's alpha. However, the material was done with the guidance of our instructor and used layered textures for diffuse and reflectivity. For the glass bottle and liquid, we used layered textures for diffuse, transparency and reflectivity. Photons were implemented to simulate caustics and the renders were done with multiple passes to make the final composition look different than the original render. I did not really explore a lot on my own, but I did end up using the techniques mentioned on my other projects for Motion Capture (transparency layered textures). Also looked around the utilities nodes and tried applying a few without much success.
The video above was my submission for the first project in VEF I. It was my first time experimenting with rigid bodies and fields. We were to simulate a Rube Goldberg Machine using a combination of active and passive bodies throughout the project to make the simulation flow smoothly and accurate to how realistic objects would behave.
Three types of constraints were implemented also: nail, hinge and spring. The nail can be found at the beginning of the video with the hammer. Hinges are in every object that rotates in one direction, such as the wheel (also active) that starts rotating when an active body (ball) falls on it. The spring is applied to the board that shoots another ball towards the casino wheel at the end of the video. In addition to this, I used several gravity fields and manipulated their magnitude and direction to get the desired outcome from the motion of the objects.
All of this was introduced in class, along with an explanation and examples of the attributes found in rigid bodies. I researched further into the constraint and field types to understand them better.
In this project we were supposed to simulate three effects with particles and fields from a given list. I chose rain, fire and smoke. For the rain, I used three different surface emitters with particle type tubes and placed one right after the others to control the appearance and motion better. The first set of particles only had a uniform field applied to them while the other two had both uniform and turbulence in phaseY.
For the fire, I used cloud particles on a directional emitter affected by a uniform, turbulence and a newton field. To get the motion, I used an expression in phaseY of the turbulence field that utilized sind and cosd, along with time.
A similar expression was also used in the smoke, which had a turbulence, newton, uniform and radial field applied to it after the cloud particles were emitted from a surface emitter and had their rate randomized with an expression. The other wispy smoke effect was created with a directional emitter and tube particles with the same forces applied to it and the PhaseY expression, but with lower numbers.
The textures for the effects were procedural and combined ramps with marble on the life color, life incandescence, life transparency and blob map on the attributes of a volumetric cloud material.
The last project was a simulation using soft bodies. The cape is a dynamic chain composed of a simple rig with a joint chain in the middle, left side and right side of the geometry; which was a simple thin cube with quite the number of subdivisions to make the deformations look better. I bound the geometry to the joints and made a simple blend shape to simulate the cape being draped over the statue.
Afterwards, I did IK Spline Handles on the 3 joint chains, leaving the first joint right at the top out of the equation so I could key the rig properly. I turned the curves created with the ikHandles into soft bodies and used the copies as the goals for the original soft body. The goal weight per particle was changed in a decreasing amount ending with 0.4 on the component editor and the object goal weight was set to a number between 0.45 and 0.6 depending on the location of soft body.
For the motion, I used gravity and turbulence and keyed them on different directions, phases and magnitudes. To make it fall I simply changed to fields to make it move to the side, then forward at the same time gravity was pulling it down. The rig was also keyed to reach the floor at an specific frame matching the timing of the fields. When the cape reached the ground, the goal weight was keyed to 0.
The statue was easier as the geometry of the goals were provided for us. However, I took the liberty of altering the last goal mesh to make a few more. The original mesh was turned to a soft body and the following meshes for the goals were also. I set them up as goals for the start particles and keyed the goal weights in order at 48 frames intervals from 0 to 1, then keyed them all to 0 on the last frame except for the ending goal which was at a value of 1.
This month I went through the Character Animation Course at Full Sail University. I learnt a lot about walk mechanics and staging for characters and their emotion. On this project, I felt all the skills acquired come together to complete a walk, starting from a relaxed pose up to a button push and reaction. I struggled with the timing and some of the breakdowns and poses for the beginning and reaction for the button push. I will work harder on my animation from now on so I can improve on those aspects as well as grow as a strong animator, using this course as a base for the techniques.
Sunday, March 29, 2015
Rabbit World - Rig by Katie German
During the month of March I worked on this rig under the guidance of several instructors from Full Sail University. A joint system was developed and improved on, starting form basic joint chains up to reverse foot lock and a more efficient back system. Using a diversity of constraints in Maya, as well as introductory scripting; the tasks became easier the further we advanced in class.
