Finished on the Six Armed Nutter

With the deadline in a few hours and me only just completing my door, I didn't manage to do nearly as much to my Armed and Dangerous Project as I would have liked. After taking a look at what really needed doing, I focused on quickly sculpting in the hair. This turned out to be fairly easy, although with more time I would have added much more detail. Still I think the lower level of detail works with his cartoony appearance. However, after working on my door project, I did figure out that the reason the paint looked a little bit off was because of bad UVs. I now know how to correctly unwrap UVs and next time I will be more than capable.
After sculpting the hair I decided to colour it green, the pure reason being that it gave the character a less human and slightly more grimey look, which I think works on him.

One big problem I had towards the end was that Mudbox 2013 tend to like unloading UVs when it thinks the object is hidden. This happened with the eyes since they where inside a sphere I was using as the eye lens. I tried taking the model back into Mudbox 2012 but this didn't work, possibly because of a backward compatability issue with Mudbox 2012 and 2013.

Failed to finish Armed & Dangerous on time.

I failed to complete my original Armed and Dangerous challenge in time. This is more or less due to me completley misjudging how much time it would take to sculpt an entire character. I have, nevertheless decided to try to finish the project for the final hand-in. I will also, however, be entering myself in ConceptArt.org's Character of the Week competition, which I will be talking about in another post. I also realise that I have not been posting my progress on the project nearly as much as I should have, I will be attempting to make a much more regular thing of it from now on.

Here is what I have done to the 6 Armed Nutter since I last posted. I managed to finished up the arms with a decent level of symmetry. I did continue to forget to put symmetry on several more times afterwards. My solution to this was to just open up the last save and redo any work that was not done symmetrically.
I then began work on the hands, using my own hands as reference far more than the anatomy books I had been using before for other body parts. I found that once I had spent a large amount of time on one hand, working out all of the nooks and crannys and trying to develop a technique, the next hand looked far too smooth. This was because the first hand had gained texture and, I suppose, character from all of the erasing and retrying in certain areas. The best way I could emulate this one the other hands was to build up or shave away the shape slowly with a weak brush.

I then decided to move onto the jeans. My initial plan was to sculpt the feet into the bottom of the jeans. However, the distortion from this was far too extreme and difficult to work with without moving up to much higher levels of subdivision, which wouldn't have been pratical. I thought, for the time being, I should sculpt the jeans as is they cover the feet and then make the feet as seperate objects in Maya and bring them in later.
I found that the actual fold of the jeans could not be made realistically with just the Wax tool and found that the sculpt tool was much better for creating the inital shape of the creases and folds and the pinch tool gave them the appearance of a sharp fold.

I used this image as reference for the general shape and location of the jean creases.

 

 

Considering the anatomy was bent out of shape by the grin, with the additional problem of the out of proportion cartoon style I found that I need to have a play with the shape of the head before commiting to any particular technique. It was as I was colouring in the teeth, I found that paint would jump to random areas of the mesh or only colour in chunks. After looking at the UV map I realised that this was because of crossing polygons in the UV. I returned to Maya and re-created the UVs using the "Optimize for: Less Distortion" setting on the Automatic Mapping. I then re-imported the UVs into Mudbox. The paint still ocassionally blots out from the area I paint. I think this may be where the polygons of the UVs are spread to wide. I will re-create the UVs soon, this time manually so that I can ensure the paint does as I wish it to.

When I got back to the face I managed to figure out the shape using a mirror, see check how the face distorts, and Classic Human Anatomy by Valerie L. Winslow. However, I did decide it would be better to create the eyes seperately in Maya out of two spheres, one with indentations for the iris and pupil, the other acts and the transparant gloss that causes the specular highlight on eyes. I found it easier to sculpt the eyelids around the spheres than I did when attempting to sculpt the eyes out of the base mesh during my play around. I painted the eye, including the irish and pupil and then gave it a mid grey specular layer, with a slightly lighter colour around the iris for extra gloss. The outer sphere I then painted with an almost black opacity layer, a completly white diffuse layer and an almost white specular layer. I then discover that the outer sphere would not give the desired specular highlight unless I sculpt a bulge over the iris. Although I am happy with how the eye originally looked, the transparent outer sphere is a little foggy since I loaded it up again. This is a problem that needs fixing.

There are still a few things that need to be done:

• After having a play with sculpting the hair out of the base mesh, I would much prefer to make a seperate mesh that holds the rough shape of the hair. I can then sculpt in texture. 
• The UVs need correcting manually.
• I need to work out a definite colour scheme before the final painting. 
• I need to add smaller details to things like pores, fabric, teeth etc. 
• I haven't yet even started to look at how I am going to pose the model. Although after a quick go on Mudbox's pose tools, it may be better to use Maya for the final pose and then either reimport into Mudbox or use a normal map to render in Maya. 

Above is the turntable showing where I am at at the moment.

The Sculpt has Begun

With only 15 days left to create the Six Armed Nutter, I have to pick up the pace a little. I have finished the tone on the main torso. I had very few problems with this area apart from figuring out how pronounced I should make the definition. I solved this just through experimentation and figuring out what works best. I found that, when I did not know the muscle groups by heart, as I do with the neck, chest and deltoid, I needed to use Classic Human Anatomy; The Artist's Guide to Form, Function and Movement by Valerie L Winslow as anatomical reference.

The only real problem is that I forgot to turn mirroring on when I was doing the second arm down. I also forgot to set up incremental saves. This means I have to make sure that the other side is absoloutly identical, which is extremly difficult. It is more likely I will treat the two symmetrical arms as seperate arms, I will attempt to keep incremental saves from now so that I do not create any more work for myself.

Armed & Dangerous

Our next uni assignment is to find an online competition that requires you to use skills used in game art. This includes 2D artwork, 3D Sculpting and Modelling and animation. I chose the "Armed & Dangerous" Mudbox challenge. The aim of the competition is to sculpt a character that is "armed and dangerous". I have spent the last week creating orthographic concept art of a character I call (working title) "The Six Armed Nutter".
I started with an initial sketch of the sort of character I wanted. and then worked out each element of the character as needed. Looking back, I should have created several initial sketches and picked my favourite.

The initial, very rough sketch.

Since the competition allowed a lot of leeway for personal tastes, I picked the same sort of mad, surreal kind of character I enjoy drawing. This was a bad idea. Since this competition is just a way of increasing my abilities, I should have tried something novel and learnt how to draw a new type of character.
Once I had my initial sketch I began looking for images online that worked with the sort of character, and the style, I wanted to create.

A collection of images used for reference and influence for the concept art.

I wanted the character to be insane and, much like the joker or the cheshire cat (American Mcgee's Cheshire Cat inparticular), the character needed a nice wide, evil grin. Already having a cartoon style, I wanted to make the bigger than it proportionally should be. After playing around with proportions I decided to move the top of the mouth to where I would usually put the nose (halfway between the brow and chin) and make the chin slightly smaller. I also made the skull slightly elongated to make room for the mouth to reach right back without making the jaw or cheekbones seem small or non-existant.

Studies of mouths from American Mcgee's Cheshire Cat, Jamie Hewlett's Murdoch and Marvel's Venom. On the right I have tested out a few different mouths and head shapes, finding out what works best.

Once I had the mouth, figuring out the rest of the face was fairly simple. I already knew that, if a character is grinning whilst the eyes are frowning, the character looks more than a touch sinister. I found that by keeping this expression and raising one eyebrow, I could make the character look even more mad and unnatural.

On the left are the final workings out for the head shape. The profile and front view on the right side of the page are the final designs.

The final thing to do was to take a quick look at his anatomy. I already knew that I wanted him to be skinny, with tone. I also know enough about anatomy that I only occasionally need reference for a more cartoony character. The only real problem I had was figuring out how the extra 4 arms where going to attach. All drawings of Hindu Gods have the arms one on top of the other. This only works when working in 2D from a front view. I figured my best option was to have them running down the side of the torso. One on the side of the ribcage, the other one the external oblique just before the character's jeans begin.
Once I had worked out what each part of him would look like, I put the parts together by drawing him in photoshop in front and side orthographic views. The only problem I had here was that he seemed too human. His face, and the fact he had an extra 4 arms, were the only things that made him seem off, I want to really throw people off. I looked into how the hands of Yoda and the District 9 "Prawns" where design and decided to do a similar thing, making him feel more alien.

Final Orthographic Concept Art

I plan on making one more piece of concept art before I start sculpting. A 3/4 posed, roughly valued and coloured digital painting of the 6 armed nutter. In the meanwhile, I plan on creating the base mesh myself in Maya using Digital Tutor's "Creating Digital Humans" series. Although, I only intend on creating a rough block out with good enough topology to sculpt with. I want as much of it to be sculpted as possible.
So far I have blocked out the torso.

Progress with the torso base mesh.

Finished the Sculpt.

I found that, once all the detail and basic shape is there, the rest is just making tweaks to features that are already there. As the sculpt got more and more accurate, I found that the tweaks gradually became smaller and smaller. However, I also occasionally noticed a big tweak that needed to be made, such as one eye being too small, or the jaw being too narrow and not low enough, both problems I had. I found that it was extremly difficult to make some of these changes without distorting a lot of the smaller tweaks made previously. This meant I often found myself redoing things like the eyes and cheekbones. I also found that the tweaking could go on for an age. Being fairly happy with how the sculpture was at that particular point, I chose to leave it, but there are a few changes I would have liked to have made, given the time.
The eyes look scuffed and unrealistic and I would have like to have spent a few hours just focusing on them. However, because I put the eyes through a lot of changes and small tweaks, the polygons in the eyes ended up stretched to the point where it was difficult to add detail at levels where it should have been possible to do so (around 100,000 polygons). This is a good lesson to learn for next time. To make sure, when creating the original shape of the features, to leave enough resolution in that area to add detail where it is needed.
Another thing I would have liked more time to improve upon is the hair. I managed to etch in a feeling of the flow and direction of the hair, but didn't have enough time left to give it more shape and detail. The hair also covers the ears, which would have been a tricky, but enjoyable bit to perfect, for now I have only left a rough impression.
The only other Improvement I would like to make is just, more detail. Going from the creases in the lips to the pores and moles. I enjoy sculpting. I can get lost in it and I feel like it gives me a feeling of form which I can apply to other diciplines like drawing and 3D modelling.

Almost there.

After adding hair and eyes to the sculpt and playing with the existing features to get a closer match to the front and side reference pictures, I went to a crit with Mark Wickham, a university tutor. Mark told me that the likeness was definitely there and pointed out a lot of small details that could be tweaked to get a better result. He then pointed out I could get further by using the photos as references on the side, instead of laying them over the model to try to find an exact fit. So, using some 3/4 photos of Jack, I was able to pick out loads of small problems with the model I couldn't see with just the front and side views. One such problem was the cheekbones and shape of the cheeks.

Using 3/4 view photographs, I was able to really find the shape of the cheeks and cheekbones much easier, as well as correct some of the smaller features.

 The cheeks lacked a lot of flesh, making him look much skinnier than he actually is. Although for some of this "fleshing out" I found that I could use the Wax tool or, in some more extreme case the sculpt tool on a low strength, the Grab tool was sometimes necessary to make major changes. Unfortunatly this often meant distorted a lot of features I had already working on, slowing me down by making me return to parts I had already worked on. On the positive side, this often gave me the opportunity to find small details I had missed before, such as my misjudgment of how suddenly and how far the mouth sticks out and how the cheeks bulge slightly around that area, causing creases. Also, whilst using the perspective photographs as rough reference I managed to find the planes of the nose to such a point that I could un ravel the shape of it in my head and make the nose much more convincing in the process.
One more important point Mark made was that the model should still be recognisable on the lower Sub-D levels. I believe I am very close to this goal, as well as completion after perhaps a few more hours work.

I believe that the face is still recognisable at lower levels as well as the higher ones.

Begining Sculpt

Our brief for this project is to use Autodesk Mudbox or a similar piece of sculpting software to sculpt an accurate likeness of a person of our choosing. Be it ourselves, a classmate or a friend. The model I chose was a classmate, Jack Eastoe.
Before starting the project I had already performed drawn studies of the shape of the skull and the musculature of the face. Although my knowledge and skill in drawing anatomy still have a long way to go, this knowledge helps a great deal in doing simple things like finding cheekbones and the location and proportion of facial features where reference photos fail me.

I already have a pretty good knowledge of the structure of the head and face.

After playing around in photoshop for a while, I discovered that it was near impossible to get the front and side reference photos to match up completly. I found that the best way to do things in Mudbox was to find a safe point in-between what the two photos suggest.
I have not yet made an enormous amount of progress on this project, but I have managed to come close to finishing the nose and mouth.

I haven't made a lot of progress, but have come close to completing the nose and mouth.

I have also found that it is possible to spend too long on one thing. The nose and mouth alone have taken hours to get close and I feel I should leave them for a short time in order to ensure I get the rest of the face done.
Jack, my model, has a very unsymmetrical face (no offence Jack). This means that the model will need a lot of tweaking when the right side of Jack's face has been sculpted.

I have found that one side of Jack's face is fairly unsymmetrical to the other. Major adjustments will need to be made to the left side later on.