Major Project - Proposal / Inspiration

For my major project I want to create characters for a game that is a mashup of the lighthearted magical girls genre like Sailor Moon, Card Captor Sakura etc, and the investigative mystery of the x-files ! I'd like it to aim for the amount of detail seen in triple A games. - Focusing more on a more mature art style with procedural texturing. However; at the same time I'd like a contrast between horribly odd aliens and girly dainty characters . Scale between the aliens and the characters will be exaggerated but there'll need to be some experimentation into how big I make the aliens. I think straight away they should be more organic rather than armoured etc.

Games featured: Batman Arkham Knight (DLC), Final Fantasy 14, Final Fantasy 15, Evolve, Bayonetta 2 , Aegis (Persona) Fanart by Blair Armitarge!

I'll need to look into polycounts for these games to aim for a similar amount. I'm looking here on Polycount to which I see FFXV ranging at 100,000 max per character (20,000 assigned to hair alone), with Sunset Overdrive (particularly more stylized) at 40,000 per character. In Bayonetta, there's around 20,000 per character. I'm thinking already I'll need to use 2 texture maps but I can allocate 2k for the main parts and a 1k map for the accessories so we will see.
Finally I'd like to research into a proper game rig rather than a strange hybrid with what I did in my previous project !

As for how the gameplay should impact the design of the characters, the amount of "gravity" the characters can get away with probably needs to be lenient. - Similar to how Kingdom Hearts deals with size difference and gravity etc. A lot of this seems to be reliant on weapons however so part of the design I'll need to consider will be how the characters fight these giant aliens !

What I'll make: 

For this contrast in giant aliens and a character to work I'll need to make:

Concept work & Complete 3D Models (Highpoly+lowpoly/textured/rigged/displayed properly in Marmoset viewer)

1 human character
1 alien

At the level of detail I wish to achieve I think this is reasonable for the timeframe as they'll be vastly different and nothing would be reusable.

soma - reflective statement

Reflective Statement of Minor Project - Soma:


When beginning my minor project, I decided fairly quickly that I would be combining both the minor and major project together into one. The grand outcome would be to make a fake trailer and no time was wasted deciding on an idea. In this way, I was able to begin working immediately and to fruition. I felt I communicated substantially better than I did in projects previously and that because of this I had gained confidence as to where I was with my work at the time. I was also more properly able to gauge my own progress aside others. 

However, one crucial point that became increasingly prevalent to me, was the balance of quality over realistic time constraints. It was extremely important, I felt, to develop a healthy work ethic; one that previously I had never adopted. With this doubts began to form over whether or not I could properly achieve the trailer I wished to create; and I felt my highest aspirations for the success of this project, my idea, come into question. Quality and skill of art is subjective, and so if I am to decide what I deem acceptable to me then I must only look at my own happiness toward it. I began to know fully well that if I were to continue with this trailer into the major project that I would not enjoy the outcome.

Instead I followed through with what I had originally planned, but took a more foresight-induced decision to switch the mysterious “trailer” into what would be a trailer for a game. With the trailer made, and the music coming along, it began to fit more into this targeted ideal all together. Making the character, Mada, I built her around a pipeline to be game ready - rather than for film. All these choices were spired on by the foresight of my career path and what journey will help me the most. A cinematic-style trailer following my aspirations could have been awe-inspiring and capturing, but it isn’t what seemed most logical in the end.

Applied skill & knowledge:

There was a very key part in this project where I understood I needed to take a strong stylistic approach that meant being more simplistic than what I would usually want to do. After this point I did not go back and forth between stylistic choice again and went in one direction towards simplified design. I had qualms about pursuing it in this way; whether or not it would show my design ability well enough, but from where I stand now it is clear to me that more detail is not a substitute for good design. The need to keep saying “go further” was ubiquitous to me, and my deviation towards subtlety was tested. I do still wish I could have gone even further. A good example looking back on my design choices, was with Mada’s boots. I knew I wanted them to stand out: so I made them bright red. “Great!” I thought. Yet I did not consider the silhouette, and so the shape of their design did not stand out and left them in a strange limbo. 

Technical look-back:

I hadn’t attempted anything as fully created as this character before. I had incredulous knowledge of the pipeline but with little experience in following it through. As such it was a slow process that I was determined to get as right as I could. It did feel as if I was taking 2 steps forward and 1 step back constantly. I’d begun Mada’s character model with it being subdivided and smoothed for animation in mind. After I had chosen to do the described in the above, I shifted the pipeline for games. Immediately I realised I could have made her much more low-poly. ZBrush is usually my “go-to” program for speed when making 3D Models. However, for the sake of learning experience, knowing fully well it would be slower, I decided to do all the modelling inside of Maya. I figured also that I had become extremely familiar with normal maps and the process of high to low baking, that now my basic modelling skills needed to be reassessed instead. I’m extremely glad to have done so as I can now highlight how possibly inefficient I am with standard box modelling.

 I went the route of hand-painted textures rather than procedural for the following reasons: For one, with the rise of procedural texturing I felt it very important to gauge my own level at hand-painting textures, and two, it fit the style far better. 

When it came to rigging and skinning, I was met with all the errors that I had not had enough experience to notice when modelling. I had made topology around the elbows to allow for a bend, directly from what I had seen other artists do, but found out after just how it does that. I did not want to use corrective blendshapes, keeping the model light; and for the hypothetical situation where I might export it into a game engine. Thus the minute errors and inconsistencies in the topology maximised time tenfold with skinning. Rendering was something I did not have to consider as the model used just one diffuse texture map at 2048x2048.

Going forward:

Whether or not I will continue with the trailer is something I have to address immediately, and that I must look to the guidance of my tutors for. With such a doubt on my mind toward it I feel it is not the best choice for the Major project.

I know fully well now where I stand with my skills from an objective view on what I can and cannot do. If I had not have made this character with such an attention to detail then perhaps I would have overlooked what I thought I knew I could make without hassle. It is my goal to be strict with myself to achieve an industry standard and highlight what is not up to par. Currently, I believe my design skills are dwindling. Going forward; making sharper, all encompassing decisions early will benefit me greatly. I’d like to next tackle procedural texturing, with rendering suited to triple AAA games in order to see how close I can get. Overall this project has been an extremely informative journey for me, and the next time I make a character I have the knowledge to not make the same mistakes that will save me time in the future. - Time that can be better spent matching the aspirations of my goals.


soma - finished mada model + sound updates !

Mada's model is completely finished and with some further skinning tweaks I can say what's done is done ! She's got blend-shape expressions but the jaw and eyelids are controlled by bones. The model doesn't use any corrective blend-shapes and I'd tried to get the skinning good enough not to use them as a faux limitation. 

Skinning those eyelids was proving very tricky, and I had to go back and fix topology on one bad loop that was throwing the corners of the eyes off. I also deleted an extra loop I realised I didn't need as the texture work was doing enough. What I did learn is that for visualisation, having open eyes is great, but for the practicality of rigging it would be better to lower the eyelids so that they are half open, giving them less space to travel. (and this way the UV map would not stretch as it is doing in the occluded area above the eye) Great learning process !!

and finally you can see the posed model in 3D space here:

ok! Now in terms of the musical collaboration, we've been talking over different tempo changes and whether or not vocals will be included. (There will be!) As such, I can show what is currently the template for the BGM next to the video ! It gives the trailer a slightly different emotion to what I expected , yet it's an unexpectedly pleasant feeling : ^) Composed by the great folks at Meine Meinung . Please check them out !

soma - mada rig updates

A large update in small form !

Alrighty ! Rig is done basically. I need to still add eyelid bones and get the controls for them, but aside from that everything is done and working well. It took me forever because I was determined to get all the LRA's set up properly ! I am in the process of skinning Mada right now. I also went back and gave her teeth and a tongue finally ! Alongside this I decided to add fake speculars onto her face and hair to aide a focal point. I'm gunna add corrective blend-shapes for face expressions & hair to save on a surplus of bones. The spine has a twist rig in it and both the knees and the elbows have pole vectors in them to stop flipping, as well as being double jointed.