Personal Development Reflection

At the beginning of the year, I had been focused on producing highly detailed and well thought out 3D characters. It had been something that I had been wanting to achieve ever since my second year, however I was unable to achieve the standard I had wanted due to my lack of skill in certain areas of character design and creation. I was very good at much of the technical side of modelling in Maya, however, I was always afraid to pursue areas that I was unfamiliar with. This had left we with many detailed and unfinished works which I was unable to finish.
Influenced by the workflow of both industry artists and the CG community, it made me determined to see a highly detailed character from start to end.

2D Digital Art and Design

I observed the workflow of character creation, from the early stages of 2D design, to the final stages of posing and rendering, I was able to identify areas that I had to tackle in order to progress to finished characters. From the early stages of concept development, it was clear that, although I had many interesting designs that I had wanted to explore, it was difficult for me to communicate that visually before the modelling stage. 2D Digital Art was an area that I had to greatly improve on in order to visualize my thoughts. Not only did I have to improve my Photoshop skills, I had to learn the stages of the design process.

Throughout the first semester, I was able to improve on this through the creation of Silhouettes, Thumbnail sketches and concept paintings. Silhouettes were relatively easy for me, but thumbnail sketches were time consuming and difficult; the concept paintings even more so. Much of this was due to my difficulties with anatomy as my proportions were often very wrong. This work ended up accumulating into 2 highly detailed concept paintings for 2 of my initial characters at the end of the first semester. I was able to produce paintings faster, and to a higher quality than when I had started. I was able to visualize characters in more detail, whilst showing colour, which I rarely included in my 2D art before. My main notable change was my change in preference from creating concepts on a physical sketchbook, to creating designs digitally.

Anatomy

The development of characters also called for a high degree of anatomical correctness. I had been out of life drawing classes for a year, and many of my previous characters often had terrible proportions. I was able to improve on this throughout both semesters, initially through an ecorche study, and later on through further sculpts of my characters. I found that feedback from online sources and students helped find areas that seemed out of proportion with my character, and in turn, made it easier for me to identify anatomical problems better. Comparing the proportions with my initial low poly mesh with my final character, I can see a big improvement.

Texturing

The low poly modelling stages of the workflow was an area that I was very familiar with and I was fairly familiar with the high poly sculpting. However, when it came to texturing, and generating maps for my character, it was an area that I have had little experience with. I had stopped many of my projects before even reaching the texturing stage, so I felt as if I was walking into unexplored territory. Baking texture maps seemed relatively easy, though time consuming to get right. Getting the texture right on my first character was time consuming however, but as I moved on to the second, the process was a lot easier as I picked up a lot of techniques from the texturing of the first. Concept painting also helped me improve my painting skill, as well as work outside that i had done for the Samsung challenge making environments in 2D.

Rendering
The final part of the workflow, and a very important one if it came to presenting work, was the rendering stage. I had always been poor at rendering my work, often taking screen-caps and relying on Maya's viewport 2.0 to make models look nice. Finding a method to present my characters properly was important for me if I wanted to show then in my portfolio. I had always been interested in Marmoset toolbag, seeing highly skilled artists use it to generate amazing looking renders. However, it seemed incredibly complicated to me so I never got round to it. Fortunately enough, it was incredibly easy to use, which was why it was so popular. It also provided real time rendering which made the process very convenient which I loved as rendering had always seemed to cause lots of problems for me. In the end, although not learning how to generate nice renders in a traditional method such as rendering in Maya, I found a tool which was easy to learn which did it for me, designed specifically for the game characters I wanted to make.

Hair
The area was the section I had left till last, as every time I had worked on it, problems with alphas always occurred. I had tried to avoid this a lot by adding masks to my character designs, but in the end I decided it was something I needed to tackle. Problems still occurred when I attempted to get my hair rendered properly, but I was able to overcome these problems to finally have fairly decent looking hair in the end. Finally finding out how to overcome these problems, making hair is now a much easier task for me. A stage that I had always feared to tackle is now a more enjoyable task for me.
I am able to work through all areas of the character design process to end up designing, creating and presenting my characters in a very professional manner.

Initial business card designs

One last post before I submit my work. Just a quick mock up for the back of my business card idea. I will need to get the design finished real soon if I want to get them in for the showcase. Once again, I am going for the hexagonal patter as a boarder round the edges, whilst displaying my characters in the center. The composition is terrible, but I will want to display my character in the middle. The hexagon boarder is far too large right now, so I will shrink it down significantly so the character is not crushed. This design will play towards the front too, however it will mostly be black, with grey sections the same tone as the other side breaking through, almost as if it has wrapped itself round. No time right now however to get them in the blog.

Expo prints

I cam thinking of having the patterns on the expo prints as well. They won't be as obtrusive like the examples above, and should hopefully blend in better with the back. This is my backup for if I can't get them to protrude out. I am also labeling them with the team name and logo to make them seem more official.

Expo layout 2.

A slightly simpler design for the showcase. The hexagons can be cut from paper, and the background is just plain white.

Expo layout 1.

I will need some kind of material to make the finished pieces protrude from the wall. I will somehow have to mount it to the wall so it will have to be light. I will probably do the same to the Logo and mount the labels for the characters too. I will also need a lot of dark grey paper to cover the background of the wall. I don't know what I am going to have the hexagons made out of, but I want them to be large and chunky, like in  the models. Shiny if possible. 

Preparing things for the Expo

How do I clearly communicate my ideas?

How can I communicate that my stand is about the creation of a team in a creative manner?
At first, I though of showing this through concept development sketches along with the final pieces. I would be able to show the process of creating a group aesthetic through media tests. However, this may not be as interesting for the audience. I also didn't want to show pictures and images I didn't make myself which I used as influences for my characters as it doesn't display my own ability clearly.
Displaying my characters together may reinforce the idea that they are a team, though the focus is taken away from the characters individually.
I think a more creative way of showing the group culture is by designing my stand around the group aesthetic for my characters.
The Group name would be displayed clearly on the top, with clear names displaying the name and class of each rendered character.
I will try and include Hexagonal patterns and cut out angles to create a feeling of things being technologically advanced.
I will also try and keep these aesthetics consistent on my business cards. This consistency should show that everything is linked together.
I will also try and cover the white of the stand, not only to stand out, but because the characters belong in an environment that is pretty dark in tone.

How do I display my work in a clear manner?

What can I display?

Display -
-Final Character renders.
(I want these to protrude out from the wall slightly to show them as final pieces)
-Description of project
-Character Silhouettes
-Concept artwork from Semester 1
-Clear labels of work

PC
- Turntable videos of Characters

Here is a layout of the pieces I plan to display for the expo. 

1 - Front/Side & Back/Side view of Character 1

2 - Front/Side & Back/Side view of Character 2

3 - Character silhouettes

4 - Thumbnail itterations

5 - Initial character concept

6.- Business Cards

Final Characters

I would have like more dynamic poses but trying to use soft select to move the limbs is tedious. If I have the time, I will try and give them more dynamic poses and weapons for the showcase, but I'm pretty happy with the results. I should be able to get some nice turn tables out of them.

Assault Class "FRACTAL"

 Scout Class "LINKER"

The bodies are low poly, but I didn't have enough time to optimize the face.

Wireframe Models

Finished Characters

Finally getting the entire model rendered in Marmoset in incredibly satisfying. I do notice that the Scout's head is far too large. I'm pretty happy with my texturing job however, and sharpening it up in the render settings really make it stand out. These are my first entirely textured character models.

Fixing the proportions

Fixed up the proportions slightly. I made the proportions on the limbs a lot chunkier on the Assault character to make him look tougher, and changed the length of the legs and size of the head in the Scout.
I'm probably going to remove the hair on the Male character as his hair style doesn't seem to suit the environment very well as it looks very 80s. The goggles make him look like terminator. I added the goggles, mainly because I can't get the eyes looking right as the topology on the head is slightly off, and the feedback I got thought the head needed some more technological features to match with the rest of the body. My anatomy skills do seem to be improving however and the body sculpts for the discarded designs did help me improve.

Polygon Hair Modelling

 I did not look into hair design very much, however I did not want my characters to be bald. I wanted to add a degree of personality to my characters, and hair is a good way of showing that.
In many western shooters, characters often had bald or short hair to fit into that "Military" look. However, my group didn't fall under the conventional "Military" archetype, despite being designed as a shooter. They act more as a group of high tech maintenance workers, rather than soldiers. Militray hairstyles are short, not only because they don't get in the way, but to make a unit look uniform which takes away from a characters individuality.

Main character from Spec Ops : The Line 

Without distinct facial features, it would be difficult to identify a character's head without facial hair or tattoos  I wanted some degree of individuality within my characters so I decided to give them basic hair. On the other side of things, incredibly expressive hair much like a lot of Japanese game characters, in most cases may make my characters seem too individual, giving the vibe that they do not work well in a team.

I aimed for something fairly contemporary to suit the working environment.

I've been using this tutorial to try and learn how to generate low poly hair for a game environment using alpha
maps.
http://www.paultosca.com/varga_hair.html

I've never modelled polygon hair before, and I have always been put off by experimenting with alpha maps, so I ended up leaving this till last when I probably should have tackled this earlier. I attempted to do quick hair models and alphas in Maya which turned out alright, but the problems occurred when I attempted bring the final hair model into Marmoset. It was problematic trying to get the alphas working properly on the character.

After much troubleshooting, unfortunately I found out that alphas don't work too well in Marmoset. I had to use aplha testing as I couldn't get the alpha blend mode to work properly, which lead to choppy aplhas on the hair. 

End result of adding the hair in Marmoset.

Baking and Texturing

Baking and texturing the accessories is taking longer than I expected. I am having to do each section, such as the limbs, accessories, and head separately to keep the textures clean. Though I had to adjust the low poly mesh slightly, it didn't take me long to UV map the torso. However, I needed to remodel the limbs and accessories as I made them in Maya without making a low poly version. I was also having problems with the head model which I sculpted from scratch in Mudbox. 

I used this short youtube tutorial to familiarize myself better with the interface and tools : http://www.youtube.com/watch?v=UGKsbh9vw2Q

Transferring the paint layers

As I painted on top of the Mudbox sculpt, I had to use PTex to paint as the UV's were not set up properly. Transferring the paint layers onto the low poly mesh meant that some sections were not aligned properly so I had to adjust them in Photoshop. Next time, I will paint directly onto the low poly model to avoid this.

Painting a specular map in Mudbox was also quite tedious as I was unable to see the shininess of the textures easily. Applying the specular map to the accessories on the character make the varying material of the objects on the character stand out a lot more. Details such as the buttons and buckles stand out a lot more as the light rotates around the character. Turned out much better than expected. 

Finally, I added smaller decals to the character. One of the traits of the group was that they were to be an organized bunch, which I was supposed to show through numbered and labelled clothing as well as consistent uniform colour. The whiteness of the number and labels added well needed contrast the the character so that the overall tones weren't too dark, and emphasized character role and placement within the group through the labels; The role "Linker" being No.04 of the group in particular.

The model is currently around 9000 tris so far, but it looks much higher than that, which I'm really happy with.

I have also added the hexagonal pattern onto the floor as I picture the dome environment they work in being very similar. The ground looks very advanced, and conforms the the aesthetics of the characters. The reason I modeled the feet in a very fragile way is because the environment itself is very fragile. The hexagonal floor, also being very unusual, should spark curiosity to what the environment is like for the viewer. 

Marmoset Toolbag

I will be using the Marmoset engine to render out my characters. I had been looking at Marmoset ever since year 2, but I have never gotten round to using it. The program makes it easy and simple to render out models in real time and is used by many industry professionals to render out models without having to spend a lot of time setting things up and rendering them out.

 http://www.marmoset.co/

Spending a short time learning the tools for the trial version, I did a test to render out a small portion of my character and got amazing results quickly. I won't need to waste time learning how to get nice looking renders out of Mudbox or Maya, however I will need to set up the character and textures properly before I can import the whole model. If I have time to rig and animate the character after the hand-ins too, I can display the character animating in an idle stance.

As Marmoset can only render one .obj file at a time, I need to set up the character so that it is labelled and split properly into chunks before I export the obj. I also want to bake the ambient occlusion from Mudbox onto the low poly model as it gives the textures a lot more depth, where as Marmoset's ambient occlusion isn't as flexible.
When I finish everything properly and add proper posing and lighting, I'm hoping to expect some amazing results.

High Polygon modelling in Maya

I have been modelling the hard surface sections in Maya. I'm hoping that when I bring it into Mudbox, it will contrast will with the organic sculpted clothing. I've only started working in high poly in Maya last summer for a short while and I have no idea if it will translate well into mudbox. Even then, I never got to the stage of baking the normals of the high poly versions before. 

A little preview of my maya contruction line....

Samsung Challenge

 Been working with my team from last year in the Samsung Challenge. I've been as one of 2 environment artists creating underground environments for a 2D puzzle game on the Samsung Tablet. Though the 2D art I've been producing is very cartoon like, which is not the more realistic, painterly style I would like to improve on, it has helped me improve on my photoshop skills.

Playing around with Alphas to get a glowing look to the environment. 

Making rock variants for different levels has taught me how to use blend layers properly

We ended up getting into the finals with our game, winning 3rd place after being sent to London from Samsung.

Low Poly Characters

I have the majority of the low poly versions modelled. The torsos are all below 10000 tris. I think the proportions look correct, but its difficult to tell without a head model. I stuck a placeholder head model onto the male character and it seems okay though I can't find one for the female one. I'll see any problems soon enough when I start sculpting the head. 

Characters : Additional Influences

The Captain

I have decided to further adapt the hermit character into a high ranking officer like character. He will be some sort of long lost officer who has been living in seclusion. By having a sophisticated hierarchy of rank, it should hopefully result in more believable characters. 

"A more sophisticated hierarchy can lead to more believable characters" Player Character engagement in Computer Games. 

I focused on the use of equipment and mechanical sections as a form of decoration rather than functionality. It emphasizes him more as a decision maker rather than a fighting soldier and makes him seem a lot more important as he is highly decorated. 

I combined the use of religious imagary of comparable religious roles. Things like the General's cap made of rings. Unique coloured uniform to stand out from the rest. Torn clothing to suggest experience, and a scepter that implies power. 

The Aviator

I used the flying character I had designed previously and adapted him to the angel role. I wanted to make the character very impersonal so I continued giving him very inhuman like elements. 

I added influences from both WW2 pilots and angels to adapt him more to his role. 

I gave him more "Angel" shaped wings. He has a very pilor like collar. I added additional rags to make him fit the team aesthetic more.His mechanical elements are a lot more functional. I also like how his large upper torso implies strength. Overall, I think I was able to make him look very inhuman. His role is intended to be more of a Boss like character, hence why he is the largest character of them all. He is not necessarily the highest ranking, but he is the deadliest of all the characters. 

 I also began developing weapons for the character. I wanted the character to be very role orientated if he is to lack individuality. His back heavy design relies on a weapon for balance, further reinforcing his reliance on his role as a character.

The third character uses influence from religious crusaders. Being the lowest ranking of the 3 characters without specializing in a specific role, the aesthetics call for the character to be functional and adaptable in her role. 

 -Cloak/Mask Variations
-Cloak suggests more secretive role such as sniper or spy. This doesn't really suit the character so I may remove this.
-Kinght like decorations.
-Flag for motivation and team identification.

I'll need to do a lot more with this character to make her fit in more with the others as she lacks any mechanical forms.

Basic colour test : 3 blocked out characters together.

Head Sculpting

Attempting to sculpt the female character's head. Its not turning out too great. I've always struggled with faces in both 2D and 3D. I should practice sculpting them more, looking at more reference material.

The facial proportions are definitely all wrong when I compare my sculpt to the image above. 

Pitch Feedback :

For pitch, I presented my recapped the objective for my project as well as showed the new variations of the character designs from the thumbnail and more detailed sketches I had done at the end of last semester. I also went into detail about the journals I have been looking at, looking at group mechanics in Second Life. Feedback stated that it was very informative, however, I didn't explore the visual element of this very well. I would have to return to explore the role of things like Silhouette, Colour, texture and form if I was to apply this to my own characters.

I had also talked about hierarchy in the group in both a rank, and visual style with sub-categories splitting up the group into more specific segments, down to a level of "Role". I then went and talked about the levels of abstraction a player has within how he relates to the group. This however didn't link back to my own character designs very well. When I went on to talk about the designs of my characters from last semester, there was no evidence that there was any form of hierarchy within the characters. Though I was attempting to show varying levels of group participation with The Hermit character, it was unclear how this linked with the project and to why I made these design changes. 

Feedback also stated that characters seemed quite generic which was quite disheartening. It was suggested that I add more of an unusual twist to them to make them more visually interesting. I would also need to take a step back to explore multiple areas, rather than focus on just one theme. 

I was also to be careful not to make the character sculpts, specifically on the male, too flat. The front and sides didn't seem to flow well. 

Main changes to make...

-Clearly show group hierarchy

-Add more of a twist to the characters

-Experiment with other areas. 

Sculpting the Skinny Body

I was sculpting the skinny body for the reclusive character. Because one of the arms were mechanical, I couldn't remove it from the torso in Mudbox as it didn't allow you to delete sections of the mesh. Instead, I took the sculpted mesh into Maya and manually deleted the sections I didn't need. I then imported it back into Mudbox to sculpt the head and texture.

Looking back, I should have imported the lower resolution mesh and transferred the details once I brought it back. It would have been a lot more efficient and I would have managed to keep my sculpt layers. Better remember to do that next time...

Finished Torso

I don't know why this isn't in here yet....

Im finishing up the male torso sculpt and cleaning it up so that it looks like there is skin on top. It was done rather hastily, but it will be a good reference none the less.
I now have a complete male and female torso which I can then proceed to shape into the desired builds for my character designs. I want to make incredibly skinny builds of both torso to model on top of for reference.

I have also been working on some facial modelling as I have not done many realistic faces before and has been something I have been meaning to improve on. I plan on having exposed faces on at least one of my characters so it is important that it looks right. Here is my attempt at a facial model of Patrick Stewart....

Concept Development

CRITICAL DESCRIPTION

A brief summary of the stages and thoughts of my concept development...

-Following the concept development timeline, silhouettes were initially drawn up. These initial silhouettes proved quite powerful.  

-Through studying read hierarchy withing TF2 characters, additional silhouettes were drawn up. 

-Working from the initial silhouettes, which seemed more visually appealing, different thumbnails were drawn up. 

-The thumbnails for the first character proved successful as it managed to maintain the tall elegant look of the original silhouette.

-The second set of thumbnails for the third character were not as successful, showing that the characters look out of place and not very functional. 

-The thumbnail for the middle "flight" character remained strong and little was changed to the design. The overall aesthetic remained appealing and fitted in well with the planned game environment. 

-Additional thumbnail sketches were drawn up for additional roles, including one of an officer, heavy and assault class. 

-These, although suitable for their roles, look fairly generic which is why I did not pursue these designs. 

-Colour tests were done and compared with designs to make sure they work well together.The Paulette used, although dull and muted, it fitted in well, though they were pretty common and forgettable. 

-Weapons and accessories were drawn up for the characters, ensuring that they were suitable for the role of the character. 

-I believe the weapon design was successful as they are fairly unusual for weapons in a shooter. 

-These tests and sketches built up to more detailed final drawings for the 3 proposed characters. However, during the process, a lot of the original appeal was lost for a more generic look with characters 1 & 3. 

-I decided to go through one last iteration for character 3 in the hope to make it more memorable and break the norms. 

-I looked at POWs in WW2 for inspiration which led to an increasingly fragile, ill looking character design. 

-This design proved more successful as it was the complete opposite of what is seen in protagonists of shooters where a strong male lead was common.

-At the end of semester 1, I have 2 unique designs that I am happy with which both show the harsh environment of the game. 

-The first design still needs some work, but overall I am quite happy with it, even if it seems a little generic. 

-The next stage I aim for with the design are more final artwork pieces that show the designs in detail, as well as orthographic images for reference when 3D modelling.