Rhys Saraceni Spring 2012 – Spellcasting Capstone – Draft

This semester I participated as a member of the spellcasting capstone team and to a lesser degree as a member of the design team. For an overview of what the capstone team did as a whole, see Ozzie’s blogpost. In general, my role as a member of the capstone team was the primary generator of non-code content. More specifically, I did these things:

• Created design team wiki, google doc, midterm presentation, and final presentation
• completed some of the required senior design “paperwork” (use cases, etc)
• researched various games and software for inspiration
• created UI mockups
• worked with other members of the capstone team to design the final implementation of the spellcasting system
• helped create the list of spells currently available
• designed and created the spell glyphs
• co-designed wands and created all wand models
• created the first person perspective arm w/ wand
• represented the capstone team with Carson and Ozzie and Design Team meetings
• represented the capstone team with Ozzie at the first end of semester presentation day and with the rest of the team at the second end of semester presentation.

Spell glyphs were initially created on paper, then translated into Photoshop. After the initial design phase, the glyphs were passed by Celia and some were redesigned. After finalizing the list, the glyphs were given to Carson who added them into the gesture system.

Wands were co-designed by Elyse Watson and myself, then I modeled all of the wands in Maya.

Initial fairy-tale/generic wand model.

The spellcasting arm was modeled in Maya.

Mockup of UI for dowsing spell. Each bubble/image represents something that the player can dowse for. See Ozzie’s post for more UI mockups.

Ryan Simmons, Spring 2012 – Island Modeling, Plant and Fish Rigging/Animation

Floating Island Modeling

My first main task this semester was to model the floating islands that Adam wanted to place within his Ecosystem playground. Rachel created concept art for the islands, and I stuck to modeling them much like the illustrations.

To create these two islands, I used the general modeling method of rectangle extrusion and manipulation. I started with a single cube for each, then added edge loops, extruded faces, and moved the vertices to create the general shapes presented in the concept art. The smaller island was made mostly from dividing the original rectangle and extruding downwards. The larger island consists of two fairly distinct pieces, so I extruded horizontally first to get the top profile and then extruded and shaped the bottom from there.

Vacuum Plant Remodel and Animation

In the fall, I was tasked with reworking the vacuum plant that is used to attract the vacuum fish. The original model was a filler model that consisted of a simple pole-like plant growing from a dirt mound. Rachel drew up new concept art for the plant and handed it off to me to model.

I took the art in to Maya and created a model that followed the art directly. I created it in three pieces: the main stem, the big leaf, and the small leaf. I wanted to capture the organic, smooth feel of the plant shown in the drawings, so I created the pieces from cv curves. For the stem, I outlined half of the plant from the side and the revolved the curve around the center to create the vase-like profile. From there, I sculpted it to be asymmetrical and then used a bend deformer to introduce a slight curve in the plant. For the two leaves, I started by creating cv curves from the side profile. I then duplicated the curve twice, scaled the two copies down and lofted the three curves together. This gave a nice organic-feeling plane in the leaf shape. To make them usable in Unity, they needed to be polygons, rather than planes (because planes only have normals on one side, the leaves were see-through in certain views within Unity). To fix this, I duplicated each leaf, scaled the duplicates down slightly, reversed the normals, the merged the edge vertices of the leaves together to make it one cohesive polygon model. The texture on the stem was created using a procedural color ramp adjusted to run down the length of the plant.

The next step for the plant was to give it a nice animation. To accomplish this, I had to rig the plant and animate using the skeleton. Having never rigged in maya before, I chose the simple (yet painful) option: Use a lot of bones and hand animate each of them. In rigging the plant, I made three separate bone chains; one for the stem and one for each leaf. The stem is a chain of four bones, the small leaf is a chain of sixteen, and the large leaf is a chain of eighteen bones. I created a large number of bones because I wanted the leaves to have a tremendous amount of flexibility to bend and wrap around the stem.

After skinning the plant pieces to the bones, all that was left was to animate it. The idea of the plant is that it grows out of the patches of dirt in the mermaid gardens. To accomplish this effect, I wanted to make the vacuum plant unwrap itself and sprout the middle stem. I created a beginning pose that had the stem shrunk down to a minuscule size, with the two leaves wrapped around and concealing it. From there, I experimented with different movements of the leaves as the stem grew out from the middle. I wanted to give each leaf a smooth, organic movement that felt like it was underwater. I eventually settled on having the leaves unfurl in a clockwise motion, overswing their end poses and then curl back into the final pose. The result can be seen here:

Vacuum Plant Growth Animation

Fish Rigging and Animation

After I completed the vacuum plant and floating island tasks, the rest of my semester was devoted to the rigging and animation of the new wildlife that we were incorporating into Adam’s ecosystem. The fish that came to fruition were the Tusk Fish, the Little Gardener, the Kelp Dragon, and the Lamp Fish. The tusk fish, little gardener, and kelp dragon were modeled by Bobby Blackburn, and the lamp fish was initially modeled by Carlos Duralde (with poly-reduction done by Bobby). The textures were created by Ashu Gupta. My job was to take the completed models, rig them with skeletons, and animate them.

Tusk Fish

The tusk fish is a unique creature that has a stone headpiece that it uses to ram other creatures. Its model was made by Bobby Blackburn and textured by Ashu Gupta.

Its stout little body is very much eclipsed by the large facial mask. In animating it, I took the headpiece into consideration and crafted a skeleton that would allow me to move everything behind the mask. This was my second foray into the workd of rigging, so it has an actual skeleton that follows the general shape of the fish. It is all a simple bone chain, starting with its root joint in its face, following its spine down through the tail. It has limbs for the two fins and the top spikes (which are rather unnecessary and only used to avoid mesh problems). This construction allowed for easy skinning of the fish, but also meant that all of the animating hand to be keyed per bone once again.

The animation of the tusk fish was created based on his shape and proportions. Because the head is heavier, the back has to compensate for the weight by working harder. The two fins constantly give downward force while the tail works harder to give propulsion and keep him straight. All of this animation was done by key frames on each individual bone. My biggest goal for this little guy was to create a fluid fish animation that is believable but exaggerated. The finished swim cycle can be seen below:

Tusk Fish Swim Cycle

Little Gardener

The Little Gardener is a cute, tadpole-like creature that roams the land carrying and dropping seeds. It plays a big role in the spread of wild plants in the ecosystem. The model was created by Bobby Blackburn, and the textures were drawn by Ashu Gupta.

The little gardener is a smaller, more versatile creature that needed movement much different from that of the tusk fish. It heavily resembles a tadpole, so I studied the movement of tadpoles in the later stages when they have developed arms. I based much of my rigging and animation work off of that research. The biggest factors I wanted to address were the use of its arms and support of its elongated body. In creating the skeleton for the gardener, I was able to build a much more complex rig that made animating much easier. I had done a lot of studying and reading into the crafting of rigs for another project, so I was now familiar with kinematics and clusters. So, I created the base skeleton, with the root joint at the head, flowing down to the end of the tail. There were limbs added for the arms and the fins. I placed IK constraints on the arms and used spline handles on the fins and the tail. This allowed for a much easier animation experience.

I created two different swim animations for the Little Gardener: One swimming using his arms and one with his arms holding a seed out in front of him. The aim for the gardener animation was to focus on using his arms and his two fins to keep him afloat, with the tail used as more of a balancing element and a guide. To achieve this, I made the arms push down with some force while the two fins work hard to keep him moving. The tail moves much slower, acting as an auxiliary floatation and guidance element. The second animation is pretty much the same, just with his arms outstretched in front of him.

Little Gardener Swim Cycle

Little Gardener Swim Cycle (Arms Up)

Kelp Dragon

The Kelp Dragon is a large, majestic creature. It is based very much on the whale in shape and behavior, with the added characteristic that is can be settled upon. They use echo location to communicate with other dragons and move about very slowly. The model was done by Bobby Blackburn, with texture by Ashu Gupta.

The main characteristic of the kelp dragon is its size. It is a king of the ocean, and it roams about at a slow pace. In being both majestic and unhurried, the main goal for rigging and animation was to give really broad, fluid strokes to the tail and large wing fins. I wanted to be able to comunicate that it is moving with powerful motions, but even in that power, the dragon is relaxed. In making this skeleton, I followed much the same process as the other fish, starting at the head and moving to the end of the tail. For the tail, I made sure to give it five joints close together so that I could get more minute movement in the wave motion I envisioned.  Then I added the bones for the two fins. The same principle as the tail applied, for the idea is to have a very fluid wave motion.

The animation for this one turned out great. Both the wings and the tail have a real heft to them, yet they move very freely. The wings especially feel nice, with a strong snap as they pull in towards the body to push the water. The whole animation has a calm, regal feel to it.

Kelp Dragon Swim Cycle

Lamp Fish

The lamp fish was the last fish I took care of this semester. It is a scary little creature based loosely on the anglerfish. It dwells in the darkness of the volcano and slowly chases the player with its dimly lit lantern. Carlos Duralde had modeled it the previous semester and I had a rough animation for it, but the whole thing was pretty incomplete. Bobby took the model and reduced the polys significantly and I created a new skeleton and animation for it. Ashu created the texture. It is a very dark purple right now, but it will be lightened.

The rig that I created for this one was thrown together really quickly, but it is effective. It starts at the nose and goes through the tail with two single bones for the tiny fins. The lantern arm has many tiny bones, which allowed me to sway it back and for with grace.

The animation for this scary little guy turned out really well for being so rushed. It has a wide, slow swish in the tail movement, accompanied by quick flaps of the two little fins. The mouth opens and closes menacingly as it swims. The best part is the slightly asynchronous bounce of the lantern tentacle, which rocks the lantern gently back and forth as the fish bobs along. The whole idea was to make him move strangely to add a sense of disturbance to its presence.

Lamp Fish Swim Cycle

Adam Le Doux – Spring 2012, Ecosystem Design

This semester my primary responsibility was designing and implementing the Ecosystem for mermaids. I divided this into two main stages:

Research & Design Document

As part of my research I consulted with two marine biologists that work at Georgia Tech: Dr. Hay and Dr. Kubanek. The purpose of this was to ground the workings of my system in reality, and also to get inspiration. Though Mermaids is a fantasy game we want the ecosystem to behave as naturally as possible, to allow opportunities for players to interact with it and generate emergent behavior. Furthermore, I found inspiration from ideas I found in ecology papers and textbooks that were recommended to me by the biologists: for instance, sunlight as a natural resource, currents that carry marine organism far from their homes, and deeply interconnected, cooperative relationships between organisms were all things I learned about in my research and tried to implement in the design of the ecosystem.

fig 1. The ecosystem design and research document

While I conducted my research I started to build a new, more coherent design document for the ecosystem. We had done some work in previous semesters on the design of the ecosystem (especially fish designs), but it was not very systematic or organized, so I took these old ideas, and incorporated them with my new ones to create an “Ecosystem Design Bible.” By the end of the semester it was 45 pages long, and there is still work to be done…

Coding the “Aquarium”

The Aquarium is our name for the test level and ecosystem I created over the course of the semester. The goal was to have a test bed where I could iteratively try new designs and ideas as I went along. It was also supposed to be a microcosm for the eventual game-wide ecosystem, which just a few fish and plants, but ones that would represent the diverse range of behaviors and organism we hope to include in the final game. These are the elements I successfully added over the course of the semester:

Environmental Features
*Sunlight: beams of light move through the environment, increasing the growth rates of plants and spawning plankton, both of which are required to sustain other life forms
*Soil: zones of soil keep track of their own health. Healthy soil nourishes plants, but if it is overused it becomes unhealthy and damages plants that grow in it
*Currents: currents push objects (such as seeds) around the level, moving things from one part of the ecosystem to another. This is a signature feature of marine ecosystems.
*Floating Islands: the floating islands are meant to serve as mini-ecosystems that hover above the main terrain. Plants can grow on them.

fig 2. Sunbeams

Plants
*Balloon Kelp: These kelp plants come in two varieties (blue and purple) which compete with each other for nutrients and space. Their seeds float upwards, serving as food for fish.
*Kelp seeds: These float for a while, then fall back down and spawn kelp when they hit the terrain. They can be carried by currents and can grow on islands and kelp dragons.
*Plankton: Plankton spawn under sunlight. They serve as the main food for kelp dragons.

fig 3. plants growing, seeds floating, fish in the background

Animals
*Kelp Dragon: Kelp dragons are big lumbering herbivores. They mainly feed on plankton and can detect them from miles away. When not hungry they seek companionship, singing to alert each other to their locations. Plants can grow on their backs.
*Tusk Fish: Omnivores that eat kelp seeds and little gardeners.
*Little Gardeners: Fast herbivores that go after seeds. They sometimes eat them and sometimes pick them up to plant them elsewhere.

fig 4. kelp dragon with plants growing on its back

Future Goals
I have a lot of goals for the future of the ecosystem project: For one I want to work on documentation that will allow future students on the project to expand my current system. Much of the code should be reusable or could be re-worked to be so. I would also like to expand the number of fish and plant species: right now a big problem is that there are no real predators, and so the herbivores quickly overrun the plants. Another large project would be to introduce more player interaction with the ecosystem, which right now basically runs along on its own, without player intervention.

Rachel Keslensky – An Underwater World Of Glass

Much of my concept work from last semester continued — and as a result most of the work I have to show for myself involves providing profile sheets for 3D modelers to use, with the occasional piece of concept art to put it in context.

In addition, I began working on my Master’s Project — developing the interface design for the Mermaids MMO. A few iterations and working with folks in class later, and I ended up refining the “Field Guide” interface for players to use.

The idea of the field guide is simple — as much of the game is up to player interpretation, the Field Guide serves as inventory container, map, and notetaking system all in one to allow players to guide their gameplay and behavior.

Also, a major breakthrough in the available technology for Mermaids — glass, and lots of it! (What else happens when you have sand and underwater volcanoes, really?) Glass technology allows for an interesting mechanic within Mermaids, as demonstrated with their magic wands, and can also be spread to jewelry, buildings, and other interesting artifacts — after all, it’s waterproof, malleable, and recyclable!

Capstone Design Blog

This semester I worked on the capstone team to help create a gesture-recognition system for spell-casting in Mermaids. While we all helped out in different aspects of the project, I focused a lot on the player’s experience with the spell-casting by focusing on the UI and controls of spell-casting.

Early in the semester we spent a lot of time researching different control-schemes and camera techniques for inspiration. At the time it was not clear exactly how the spell-casting would work outside of having different symbols for each spell and requiring users to draw those symbols with their mouse before being able to cast each spell. There were 3 main components to the design of the system: finalizing the nature of the symbols, designing how the spell-casting would look like in the game, and designing the controls for it all.

There was a lot of discussion about how the symbols should look like. In Mermaids there is an entire fictional language comprised of different glyphs that are displayed throughout the world. Spells would be a certain type of glyph. Each spell needed to look very distinct from one another, and each of the spell glyphs would need to be clearly a spell and not just a “word” glyph. To add to this, spells would need to be simple enough to be drawn in somewhat of a hurry, and easy to remember. Ultimately we decided to make all spells a single-stroke, while all other glyphs would be multiple strokes. Single stroke symbols are easier to draw, will stand out from the other glyphs, and as an added bonus would “feel” more like casting a spell with a wand instead of simply feeling like drawing a symbol on paper.

The look and feel of the spell-casting was a completely different dilemma. While there are a ton of different games out there that use gesture-recognition in different ways, we could not find an example of a game that used it in the same way we were. We experimented with a ton of mock-ups for how the system could work: over-the-shoulder, first-person view, having a “canvas” appear to draw on, even implementing it into the notebook inventory system. Besides worrying about how to trigger the “draw spell” mode, we weren’t sure as to how to allow players to “submit” their gesture, notifying the game that they were done drawing the gesture and it needed to recognized.

over-the-shoulder mock-up

notebook spell-casting mock-up

“canvas area” mock-up

Eventually after a lot of prototyping and feedback from the team, we decided that the first-person viewpoint was the most desirable as it felt like the least contrived solution. Players would go into a first-person view, hold down their mouse and draw. Carson was able to add in a nice blue particle effect at each location the player drew on, with essentially creates the lines in the game. This was a very intuitive way to add feedback to the player and also make the spell-casting become part of the world, as the particles are in the real world for a few seconds and other players will be able to see them (which takes care of another aspect of the spell-casting Celia wanted: players teaching glyphs to each other in the game).

final implementation of spell-casting

The final aspect of the spell-casting design is the design of the controls for the player. Originally we had a simple tacked-on, “proof of concept” sort of control scheme where players pressed a button to enter “draw spell” mode (which brought the camera into a first-person perspective), held down the left-mouse button to start drawing with their mouse, and then pressed the draw-spell button again exit this mode and check to see which symbol had been drawn. We quickly realized that this would need to be more elegantly integrated into the rest of the controls, and so we overhauled the camera system in the game. Now players can press the “q” button at anytime to toggle first-person view off and on In first-person view they can move the camera around (by holding down the middle mouse button), select objects just as they could in normal view, and draw spells by holding down the right mouse button.Whenever the player lets go of the right-mouse button the game automatically checks to see which gesture was drawn. This allows the player to elegantly draw a spell, look around to find a target, and quickly cast the spell on a selected target without changing camera perspective.

Finally we added in some alternate button controls for players that wish to play the game without a mouse, so players can use a keyboard button that corresponds to certain mouse buttons (e for right click, alt for middle mouse button).

Overall it was a great experience working on Mermaids this semester and I am glad that we not only achieved gesture-recognition for spells. but that it also feels very fun and intuitive to players.

Erica Penk: Mermaid Mocap and Research

Hello all! It’s been another exciting semester working in the Emergent Game Group. As a registered graduate student for this Project Studio course, I coordinated the motion capture research and implementation for the mermaid models.

Research:

I first sectioned the mermaid model into smaller sections in preparation for a swimming motion simulation created by Jie Tan, a fellow Georgia Tech student. Jie Tan uses fluid simulation and motion path optimization to procedurally generate articulated swimming motions.

You can read more about articulated swimming motions here: http://www.cc.gatech.edu/~jtan34/project/articulatedSwimmingCreatures.pdf

These are the final simulations released in April of 2012:

http://vimeo.com/40986196

http://vimeo.com/40987432

In addition to creating procedural reference material, I will conclude this semester by using an underwater camera to film swimming sessions with a mermaid monofin. Embodied interaction with the mermaid monofin is crucial to create a believable motion captured mermaid.

Implementation:

Here is a link to the first iteration of mermaid motion capture:

http://vimeo.com/39845874

Things I’ve learned this semester:

-Use Maya mesh>”extract” to create a new object  and then use edit mesh>”append to polygon tool” to fill the holes (when sectioning the mermaid)

-Vicon Blade Software

-Importing .fbx motion data into Maya

Alissa Hartenbaum, Spring 2012- Windmill Fish Model

This was my first semester with the Mermaids team. In addition to brainstorming during the design meetings, I was tasked with creating the 3D model of the windmill fish. The windmill fish is a predator that uses its long windmill-like fins to propel itself forward. The fish needed to look both streamlined and powerful. I spent a lot of time tweaking the vertices of the fins and tail. If nothing else, making this model made me very aware of the slight difference between blocky and organic shapes. I also constantly needed to keep in mind the type of animation that would be applied to the model, and allow for a lot of flow-y movements.

The concept art used as a reference can be seen here: http://mermaidsgame.net/wp-content/uploads/2012/04/windmill-all.jpg

Carson Britt Spring 2012 – Spellcasting Capstone

This semester I worked with a team of four on designing and implementing a spell casting  system as part of my Computational Media capstone project. This spell casting system is based on gestures that the player drasw with the mouse. Each gesture corresponds to a spell that will be cast upon the successful completion of the gesture. My role on this project involved the initial design of the system and then the programming of the core gesture recognition logic and the drawing mechanic. The following is a brief explanation of how I implemented some of these features.

Drawing:

When the player holds down the mouse button they will see a stroke appear in the world that follows the mouse [fig 2]. The 2d coordinates of the mouse are converted to 3d world coordinates with a set distance away from the player, this makes a more tangible drawing interface and also allows other mermaids to see what other mermaids are drawing. These drawings will fade after a few seconds.

fig 2.

 Gesture Recognition:

Essentially the spell casting works by comparing the gesture that  the player draws to a set of predefined templates. When the player holds down the mouse the system will store a list of x,y points for the cursor, this list will then be compared to each template list [fig 2] and assigned a score as to how close the gesture matches each template, the template with the lowest match will correspond to the gesture, if no score is within a given range there will be no match.

fig 2.

The system also accounts for users drawing a rotated or small gestures by optimizing the gesture (i.e. rotating, moving, scaling the list of points to match the templates).

Ashutosh Gupta – Spring 2012 – Mermaids Concept and Texturing Artist

As one of the primary concept artist for Mermaids this semester, I designed many of the fish, plants, and structures within the game.  Much of my inspiration came from fish images found on Google and various other underwater influences.  It took me a couple of tries before I finally understood the flowing, watercolor-like art direction for designing for Mermaids, but once I got it, I took it and ran.  My skills in designing were pushed further as I learned to texture models and edit my images on the computer.  By learning software such as Maya, Photoshop, and Unity, I was able to literally paint my ideas onto the models and animations built by other members of the Mermaids Team.  Here are all of the designs and textures I have worked on this semester:

The following are images of models implementing my concept art and textures.

Richard Shepardson – Mermaids Spring 2012 – 3d Modeling

This semester, my job was to model static 3d objects for use in the game world.  The models that I made were glass arches and pillars for a new temple that has been implemented in the level, several different colored glyph blocks that use normal mapping to look like high-poly models, a jelly seed that was used to replace the original seeds in one of the levels, and a building block that was designed to look like a d-8, that has not been implemented.

During the semester, I learned how to build low-poly models and make them appear to be high-poly models.  I also learned of several tools in Maya, such as Soften Edge, Insert Edge Loop, and various deformation tools, to accomplish the task.  I also learned how to use a high-poly version of a model to create normal maps for low-poly models to also create the appearance.  Finally, I learned several ways to model simple objects, such as modeling one half of the object and then mirroring the geometry to save time.

I am also working on remodeling the mermaid for the game and assisting Erika Penk with the motion capture for the mermaid to incorporate it into the game.