[Contextual Studies] Representation and Space Task

We were given the task (in groups) to one low pixel character and one HD character to compare in relation to technology, creativity and culture. We were also asked to find one silent or basic sound environment and one fully immersive 3D environment.

The two characters we decided on were the Pokemon series of games and also Altair from Assassin's Creed. I think this shows a broad change in a relatively short space of time to the technology used to create characters in games. Although the culture varies vastly between the two characters, there are a few common traits that can be noticed. The first is that both games contain a form a combat or violence. Even though they marketed at completely seperate age groups and set in different times, in different worlds, they share this underlying trait.

The two environments were also quite a difficult choice. We decided to go for a comparison between Lemmings and the citadel from Mass Effect (2). Yet again there is a stagering leap in technology, going from a 2D sprite based design to an advanced polygonal landscape, and this shows in the level of immersion that the game presents. Although entertaining, Lemmings doesn't absorb you into the world anywhere near as much as Mass Effect does. Arriving on the citadel in Mass Effect is a defining moment that will stick in your memory for years, if not indefinitely. Even though I played Lemmings substancially when i was younger, I still couldn't give you a clear picture of any of the environments other than they are brown rocky soil with grass on. On the other hand I could give you an entire monologue on all the attractions in the Mass Effect citadel. For that reason I think that environments have definitely got larger and more extensive in modern games, showing that perhaps as games modernise, the 'Space' aspect of them increases to give that feeling that the player can go anywhere, and do anything.

[Rigging] Weight Painting

To make my rigged legs animate more realistically, I tried to use weight painting to represent the body’s natural deformation with movement. I found that painting weights didn’t make a huge difference with this mesh, possibly because of the low polycount. I accessed the Weight Painting menu by right clicking on my leg mesh followed by selecting “Paint Skin Weights Tool”.

I imagine when used in the future this could put the finishing touches on a mesh to get the smoothest possible movement.

[Rigging] How to Rig in Maya

After Chris's lectures, I made several notes based on the Rigging process. These were the rough notes that I made which just detail the entire process for easy reference:

Rigging
--------
Select Leg Mesh -> Layer New Selected [Puts Leg Mesh on its own layer]
4 screen perspective, work on orthographics or joints will become misaligned.
Switch to Animation tab, Go to Side view.
JOINT TOOL (Or Skeleton -> Joint Tool)
Shading -> X-Ray Joints [to see joints in perspective]
Move joints sideways over leg.

Panels -> Saved Layouts -> Persp/Outliner

O Hip_L
         O Knee_L
                  O Ankle_L
                           O Ball_L
                                    O Toe_L


OPTIONAL: Create Locator to plot Joints.
                      CTRL + D -> Duplicate
                     JOINT TOOL, hold down V, middle mouse button -> Snap to points


Function F8 -> Joint orientation - Goes Blue
                        Press the [?] icon at the top to show orientation
                        'x' is going down


Skeleton -> Orient Joints [ ] -> untick "orient child joints"
                                                second axis : [+y] to [-y]


ALT+B -> Change background colour


Skeleton -> Mirror Joint [ ] -> Mirror across: YZ       -           Behavioural
                                               Search for: _L
                                               Replace with: _R             // this changes the naming convention in the Outliner


IK Handle Tool -> Options: Needs to be set to "ikRPsolver" for the first joint connection.
                             Select Hip -> Select Ankle (joints)


Create -> NURB Primitives -> Circle
Middle Mouse, Hold V, Drag to Ankle IK Handle
Modify -> Freeze Transformations
Select the NURB Circle - then shift select the IK Handle. Contrain -> Point     // this is done so you can use the NURB circle as a controller, and  the contrain means moving the controller will always move the IK Handle.


Front View -> Joint Tool -> Draw one for base of the spine, one slightly above. P to Parent.

At this point the lesson ended, so I researched the rest of the process myself and followed the tutorial that Chris gave from Rigging101. This included finishing the Reverse Foot Lock so that the foot could be controller easier.

[Door] Mudbox and Normal Mapping

My next step was to add finishing details in Mudbox. I wanted my door to have a cobbled appearance, with the center parts having a more metallic feel. The most efficient way of doing this is to sculpt the desired detail in Mudbox followed by exporting the detail as a Normal Map into Maya. This would retain the detail of the original sculpt but without the performance issue of having to render a high number of polygons.

Before I worked on my final model, I tested the Normal Map process on an early version of my door with fewer parts. I went up to subdivion 6 or 7 in Mudbox (totalling around 2.5million polys), and was happy with the look of my door.

At this stage when I tried to export the Normal Map through the 'New Operation' panel, I ran into several problems. The first being that due to the high number of polygons in my model, the exporting time was horendous. After 11hours it still hadn't exported on Subdivion 7 and Mudbox went into 'Non-responding' mode. I then tried to export on Subdivion 3, which took about 90 minutes to complete. When I opened in Maya, I realised that the Normal Map obviously won't add and subtract geometry, which is what I had done in Mudbox to get that cobbled appearance. For this reason there was clipping on the edges of the model that I had pulled out slightly.

I then began work in Mudbox on my final mesh, with much more knowledge of how the entire procedure would work.

This was what my final sculpt looked like - retaining much more of the original shape so that no clipping would occur when it was imported into Maya. Unfortunately this again was done at Subdivion 6 and wouldn't export after 8hours, so I decided to export it at Subdivion 2 into Maya, then simply do the turntable from Mudbox. I may do a turntable of my original sculpt as well as an addition.

[Door] UV Mapping

My next stage was to begin UV Mapping. UV Mapping is a way of unwrapping the mesh so that it can be skinned at a later stage. More often than not it is quite a time consuming process, and took me several attempts to get it correct.

The UVs should be layed out in such a way that you can easily see what each 'piece' is. This will make the entire skinning/texturing process that much easier. I followed a UV mapping tutorial on YouTube based on the 'Planar Mapping' technique. [http://www.youtube.com/watch?v=TnTn6gjPCVA&feature=relmfu]

This tutorial was very useful in understanding the overall theory of UV mapping. I also tried to experiment with the different mapping options including Cynlinder and Spherical mapping. The Automatic mapping setting seemed the most efficient and quickest, however some parts were layed out quite unusually.

This is my finished UV map, with all pieces of my mesh clearly visible and easy to work with.

[Door] My Second Attempt at Modelling a Door in May

Learning from past mistakes I made sure to test my mesh in Mudbox after every alteration. I also tried to keep the subdivison slightly lower. Finally, I wanted to alter my design somewhat, as I had planned to just leave the center of my model blank.

This design looked much more sci-fi compared to original medieval design, but I thought that a synergy of the two aspects might make quite an interesting new concept, so I was happy with the outcome.

The concept I imagined originally was my door would be more of an archway with a portal type mechanism. So instead of having to physically open a door to get somewhere, the user would be teleported. In my revised design, I tried to cross the new sci-fi design with my previous medieval arch theme. I did this by changing the center of the doorway and how it worked overall. In my new design, the door had a ‘sentry’ system to omit people who could use the portal. My idea was to add metal chimes in the center, with rope dangling slightly below them. The user would grip the rope and bang the chimes together, allowing them to teleport only when the correct frequency is used to open the portal. Almost like something from Myst.

The best thing was the model subdivided correctly in Mudbox.

At this stage I had learned to use Maya a bit more proficiently, and was trying to get the same error with my new mesh as I got with my old one to stop me repeating the same mistake sometime in the future. As in turned out, I discovered the error was caused by me merging vertices incorrectly. When designing my original mesh there was an area that I wanted fused, so I simply used the ‘Merge Vertex Tool’ to merge the parts together. I didn’t realise at the time that Maya doesn’t automatically merge edges when you use the tool, so it turned out I had a double edge somewhere in my mesh. I’m yet to discover exactly where it is, however I will continue with my latest mesh as I believe the design as evolved at this stage anyway.

[Door] My First Attempt at Modelling a Door in Maya and subsequent issues

From the lectures and demonstrations on Monday and Tuesday I began modelling my door in Maya. I followed the method that Mark suggested, using the extrusion tool to build up my model from a square polygon, followed by realigning the vertices to my original design by using an image plane. Before going into Maya I had done a (to scale) sketch of what I wanted my door to look like. This was the result:

                                             (Front)                                                                           (Side)

After a short time I had managed to develop a rough outline of my door. Since my next step was going to be to export my door to Mudbox, I thought it would be a good idea at this stage to check to make sure my door would export correctly before I continued work.

After I exported my door into Mudbox, it appeared to work fine. This was until I began to try to subdivide my model, where I got the error message; “Cannot subdivide meshes with edges shared by more than two polygons!” Through research I found out that the most likely cause of this problem was that there were triangles somewhere in my mesh, and Mudbox can only divide meshes with 4-sided quads. After finding several triangles in my model, I managed to fix what I thought was the problem, but unfortunately I received the same error in Mudbox. I then tried using the Cleanup utility in Maya, but it couldn’t find any problems either.

 

At this point my stripped down model looked like this:

With very little information about this problem on the Maya help files or support forums, I tried several generic computer-associated fixes that often work with many applications. This included selecting all the visible parts of the mesh and duplicating followed by deletion of the original. I thought that there might be a hidden vertex; edge or face somewhere in my model and this might fix the problem. I also tried pulling vertices about using the move tool to try to see if there were double vertices somewhere on my mesh (which it said could be the problem in the Maya help files). Unfortunately neither of these solutions fixed my problem. By backtracking through my save files, not even the .003 or .004 saves subdivided in Mudbox correctly, and by talking to people I heard that the problem was very hard to fix due to a tiny discrepancy somewhere in my mesh.

At this point due to my limited knowledge of Maya, I believed that I tried to make the model too complex and rounded, and that is why Mudbox refused to subdivide it. For this reason, I decided to begin again but try to make my model more square-edged with less subdivision.