By Bradley Hennessey - N10199756
Development
Fantasy Creature
Chosen - Lamia
For this assessment, we are tasked with developing a fantasy creature. I decided to go with a Zoomorphism Hybrid, as it was more interesting to me to do a half human / half animal creature. I originally wanted to do a mermaid, but a previous student had already done one and I did not want to repeat the same. So, I decided to do a Lamia! A Lamia is a half human female / half snake creature from Greek Mythology. It is part of the “Libyan myth”. Like mermaids, they hunt young men, and hook them in with sexual desires. In where then eat them once they've been dragged away from society.
Synopsis
Genera – Spy Thriller
Theme – Mythpunk / Dieselpunk
In an alternative reality, where electronics never progressed beyond the transistor. In a world where Frankenstein’s inhuman methods where standardized. Humans have made themselves obsolete, through there experiments. Bringing rise to new species, of human / animal Zoomorphism Hybrids. These beings have skills and abilities far beyond that a human can poses. Because of this, the world is now dividing between multiple factions. Leading to hard divides between groups.
Here in this dystopian, we meet the Lamia named Ava. She’s a spy that works for the U.T.I.B (United Tactical Investigation Bureau). Their job is to bring peace between the factions. A task that’s far harder said than done.
Her current mission is to delve into an illegal smuggling ring and gain intel. This ring is run by the water dwellers. Intel shows they have set up a ring near the water, in a poor location in town. It’s her job to blend in and preform her job as efficacy as possible.
Character Profile
Ava is a Lamia part of the Laminass Faction. She’s a new bread, compared to her parents. Who where mutilated souls from past human experiments. Once the world was turned upside down and the new factions’ lines were drawn in the sand. There location put them inside another group’s lands. Leading to them having to escape. In the process, both of them where killed but they managed to in their final sections get Ava across the border. From this, she’s decide to dedicate her life to unifying the broken world and bring peace to all.
Ava, while being flawed from her past, always tries to do the right thing. Even if that’s not what is seen in the eyes of the law. but when not in the heat of a mission, if a childish individual finding enjoyment in things many never find.
Personality: Dedicated, troubled, childish, sneaky
Physical: Lamia (half-human/half-snake), long ponytail, vest
Design Process
Where the fun begins... again
Stylistic References
To help in the styling process, I developed a stylistic reference sheet that explains the visual mood and style that the character art will follow. These will help get an understanding of the visual shape for thumbnail sketching.
Thumbnails process
To begin the development of the character, I produced a varity of silhouette thumbnails that explore different types of character designs.
When I went into this process, I wanted to explore different ways I could add detail to the characters silhouette. That helps to separate them from other similar characters. The snake tail… well not much I could change there, so the bulk of design work went into the human top half / different clothing outlines. You can see on the left of the image; I was exploring different types of hair and body types. While towards the right, its focused more on tail and final details. The 2 far right images are the final images from this process. When I started to imagine what these outlines mean.
Next, I sketched out some appeal sketches in order to get a feel for the characters design. This will help me develo my understanding of how skin, hair and body parts should move within the model. Allowing for a more fluent modelling process later.
These drawings also help me decide on what sections of the body will be rigged to add extra animation. By experimenting with appeal, I found that the ears and mouths (tooth) add a great deal of appeal and emotion. So from this I built a basic mock-up sketch to show the sections that will need to be rigged.
Colour pallet sketches
To gain a better understanding of what colour pallet would be most suitable for the Lamia. I decided to try out and experiment with multiple colour pallets. I looked at different types of snakes from across the world as to gain colours and styles that would look suitable for a snake character. In the end, i decide on the blue pallet. As i felt it best represented the character and felt the most consistent of the four.
Turn-around Sketches
With the design finalised and colours selected. Now all that remined was the Turn-around sketches. For this part of the process, I decided to do 4 faces turn around sketch. But saved time by only having to sketch 2 faces, and then editing the other 2.
To again help with appeal, I drew 2 expression faces. I will use these later on as a refence, when modelling the face to make sure it can allow for these kinds of expressions. This final turn around spreads the tail out in 1 straight line, this is too help with the modelling process. But under normal use, the model would be far shorter, as the tail would be scrunched up. Therefor I needed to account of this in the length of the tail. The final length allows for great move ability, while not being ridiculously long. But also still will look big when it’s being animated.
let the fun begin!
The first step in the modelling process was to create a low polly base model for Blocking in Mud box. This low polly model acts as a guide for mudbox, in where I’ll be able to model the high detailed character.
To help in the modeling process, I looked for a similer model I could use as a refrence. I found not that many models compared to other fantasy creatures. But i did find this Jashin-Chan model that would work just fine for my modeling process. I used this to see how they did the vectores and wireframe in the model.
While this process was fairly simple, I was running into two issues. These where TIME and EFFORT! Time wise I only started modelling the character in week 10, already far being schedule by 3 weeks adding a lot of stress. And EFFORT wise I was going and adding far too much detail to the character.
For example, I went along with adding details for the eyes so they could be animated and conducted a range of research into how to do this. This reserch was based around how and in what ways you do an anime eye on a 3D model. But this is only meant to be a basic model used for the modelling in mudbox. So it was really a waist of time trying to get this model looking amazing. When really it should just be a quick and dirty model. Regardless this all means I would have a lot less time for the incoming steps
That being said, some of the infomation I found turned out to be really useful, for example:
I found a Deviant Art user by the name of Nsio, that has a large and extensive libary of anime modeling and design sheets. These where really useful and i used to help me in the rest of the modeling process. You can find the rest of the sheets here: Nsio!
Overall the Turn around model came out pretty good, while there defiantly was some areas I wished to improve. Overall it should be more than enough for me to do the sculpting.
Mudbox! So, it’s finally time for Mudboxing! Going into this, I was excited to get into making the Hi-Res model. Unfortunately, I had no drawing tablet for my main pc… so I would have to do all of the Mudboxing with a mouse. So, going in I knew this was going to be a “FUN TIME”.
I began by firstly smoothing the model up and ironing out the rough edges to get it ready for “Mudding”. After that I mirrored the X axis, as to mirror all my actions to speed up production. With this I went onto grabbing and pulling the mesh around to construct the high resolution model.
From here I used a verity of tools to bend, move and form the high-resolution model. I used a system of grapping and pulling on the mesh. then smoothing it out. I repeated this process until I formed the basic shape that I wanted.
I decided early on that I would just do the hair inside of mud box, given I believed I could just use the grab and pull tools to form a wig. I gave this method a try, but given the complexity of the hair, this simply wasn’t going to cut it. While it does work, it only looks good for small changes in the geometry. Also add the fact it would make the next steps a pain. I decided to look into other methods.
I found out after a little digging, that mudbox allows you to add new base shapes into the workflow. After some testing I found that a base sphere would allow me to easy form it into hair like shapes. I conducted a few more tests just to make sure, but the hole process was very easy. Given that I decided to use this on the main model.
By the end, I was able to create all the shapes needed to form an appealing hair design. I followed the turnaround sketches as closely as I could. But I did need to deviate a little under the ponytail. This was to add more appeal to the back of the hair design with the hair dropping down a little.
With the hair done, i was almost done, all that was left to do was some small edits to the model. Just to make sure everything flowed and looked good before i moved on. Below is the video of the workflow.
Unfortunately, as I was almost done, I found another issue. As you might be able to see, the split in the lower tail is no symmetrical. This was because Mudbox would glitch out, whenever I tried to use model down there. I Looked around but could not find out why, this was happening. My best guess was it had to do with the made model not being symmetrical and that confusing Mudbox. Unfortunately, I was out of time, as to make any changes. So I would just have to fix it when doing the Retopology.
Mudboxing for the most part was straight forward. While there was some hiccups. That naturally comes with trying something for the first time.
Time for retopology. Retopology is where you go over a high-resolution model with a smaller number of vectors. This leads to a far more optimised model. That keeps the detail, but can run on lower end hardware.
This is where the pain began, retopology is not fun. At least as a learner! I decided to begin the process at the head, As that is the most complex area of the model. To aide in this process I looked at a range of models on Sketch-fab and google. To get a better understanding of how they formatted there retopology. I looked at how they work around the face, and the layout of triangles and vectors.
The area I found issues in was once again that tail. I knew I needed to make it symmetrical. But I had no idea how difficult what was going to be. I firstly tried to just do it normally, by just doing the retopology on one side of the model. Then let Maya handle the mirror on the other side. This sort of worked… but maya keep bugging out. I think it was detecting the difference and then would get confused as to where it should be placing the next vector. This also meant smoothing out the Mesh was impossible, as I had nothing to base it on. In the end, the best solution was to wait until I was done, then delete and duplicate one half of the model to the other side. Not the best… but it works.
The other area that I found I needed to change was to do with the bra. During this whole process it had be apart of the body. But this led to a range of issues during retopology. Mostly around I could not get the vectors to effectively display and separate the bra from the body. In the end, I found that the best way was to make a copy of the body, and just deleting everything apart form the bra. Then edit it to have volume.
While it would have been more effective to have added the bra since the base mesh model. I simply do not have the time to go back and redo all those steps. It is something to keep in mind for next time. As I now have a better understand of how these tools operate
Overall, for my first time, not the worst. But It has made me learn a lot about the process of Retopology. It’s something I definitely need more practice in, to improve my workflow.
Time to do the UV Unwrapping. With the experience I had from unwrapping the previous model in assessment 1. This time the process was far easier. The main ideology I used for this unwrapping process was to do with texture resolution. In the previous model I had run into issues with the size ration of the UVs.
Now this was very difficult because of the shape of my model. Her long snake tail was difficult to position correctly. I didn’t want to cut it in half because then the scales I was planning to add would not blend smoothly with one another. While I wanted everything to be as highest resolution as possible, in the end I found I would just have to use a quarter of the canvas for the tail, and religate other objects to there own layers. While it is resolution will suffer, I can mask that in how I texture it.
While using more Tiles would have probably been more appropriate here, to help set out all the UVs. I was unsure of what issues that might create in substance painter. With time running out, I decided to just keep it to one Tile. But utilize separate materials to help divide it up into layers. As so texturing would be easier in substance painter.
I realise now, that by dividing them up, like I have done. Will dramatically increase the files size; I should have not made all the UVs be so big. A minor error, that I will pay for later. It will not affect the final model, so I kept it. But I realise the why and how the method I used was not the best.
Once I was happy with the UVs, I moved onto Substance painter!
Texturing! Now having used Substance painter, a considerable amount of times. The process here was a piece of cake.
I began by doing the skin and tail, to test the UVs resolution and quality. While the tails quality was lower compared to rest of the body. The difference in resolutions was very reasonable and could be easily hidden with specific texturing. After some messing around with different brushes and textures. I found the best way to mask this contrast was to use the interlace between the belt as a blend point. Here I could transition between the different resolutions easily without having to worry about people seeing it.
But I still needed a texture that would not be noticeable. I played around with some custom snake scale material from the substance SHARE Store. But unfortunately, some were too old and did not work with the current version of the program. While others were to realistic and didn’t match the cartoony art style I was aiming for.
In the end, I found a diamond tile material, that with the correct adjustments work perfectly. It’s cartoony design helped the tail flow with the rest of the body and hide the lost detail.
From this, I moved to texturing the rest of the character. I kept as closely as I could to the colour elements from the turnaround sketches. To match that cartoony art style, I used 1 base colour for each object, based off the sketches. Then painted in a darker shade in darker areas, then a lighter one to highlight lighter areas. This process adds weight and contrast to the model.
Overall, Substance painter was very straight forward. With having used the program many times before. I had a good understanding of what to do. Leading to a very easy and quick process. I'm pretty happy with how the model turned out. It is leaning a little towards the uncanny valley, but not so much i am worried.
Rigging, oh man… if there was one step that gave me a massive headache. this would be it. Using advance skeleton was for the most part, easy. When it came to the basic body rigging, it was very simple to understand on what I needed to do, and the program helped step me through those steps.
Once I moved to building the skeleton… that was there I ran into issues. Advance skeleton is designed for use on mostly human based characters. Therefor having a half snake-based character… this led to a lot of bugs and issues. When I would build it, it would usually end in crashes or duplicate tails. I eventually found out that I needed to add an attribute to the tail rig called “no mirror”. This stopped the rig making duplicates
The area I faced the greatest issue was around the race rigging. I spent 3 days none stop working on it. But I could not get it to work, I was just facing a wide range of issues and bugs. There little online information out there, that I could use to help fix these issues. The main area I believe these bugs came from, was 2 areas. firstly – it was to do with the anime art style, the big eyes and small mouth design did not lend itself well to what the program was expecting. Secondly, the Retopology could be improved, the layout of the face and the vectors. could be the main reason the program was stuffing up. Regardless in the end I simply ran out of time, but I tried by hardest to get it to work.
The final rig works well, I added some bendy joins to the tail to help make its movement feels more realistic. That being said, I need to get a better understanding of skin weights. Even knowing I used them to great effect around the head and body. I struggled to get the arms and elbows to bend correctly. While it’s not too bad, the weight needs to be distributed more evenly over the arm.
In the end I had to redo the rigs a total of 12 times. but I finally got it done and moved on
After a lot of work, Finally I got the model compete and uploaded it. I found on SketchFab that there was a setting to put the model into an unshaded state. Where it looks similar to cell shading. I though this helped bring the model closer inline with the anime art style that the concept art was designed around. In doing so, helps prevent the feeling of the uncanny valley.
Overall, I am very happy with the final product. While there is defiantly a lot of room for improvement. Mostly around rigging. For my first time, it has a hard effort that I feel has payed off well. I think the area I will need a lot more practice in, is definitely to do with skin weights. While the model got close, there is a few areas that are not exactly perfect. While I attempted to improve the final weights, I was running into a range of bugs and crashes and was simply out of time. So when I plan to make another model, this will be something I focus on.
I managed to develop 7 test faces, based on the universal expressions. With the issues with rigging talked about above. I moved the vectors around myself to show off the expressive potential of the model. if I had more time, having different type of eyes would be a good inclusion. As to give even more expressive potential to the model.
Regardless, I’m happy with the product. and with the wide range of information I have learn from doing this model. The next one I do will defiantly be a lot better, as I have a better understanding of the process and the steps needed to get to the end product.
GHPurple. (2020, April 12). Zero Two - (fully rigged). Retrieved 2020, from https://sketchfab.com/3d-models/zero-two-fully-rigged-cf94036a0b634637a4feacceb1f5db46
NSIO. (n.d.). Nsio - Tutorials. Retrieved May 7, 2020, from https://www.deviantart.com/nsio/gallery/45894768/tutorials
Mark Escuadro. (2018, July 21). Jashin-Chan. Retrieved May 7, 2020, from https://sketchfab.com/3d-models/jashin-chan-29a79c5467074174a38ee44504e4d781
goblin165cm@gmail.com. (2018). Girls Frontline _ PKP. Retrieved 2020, from https://sketchfab.com/3d-models/girls-frontline---pkp-3480a1f36b464134a86418020643001d
darmru (2015). Hestia 3d Figure fanart. Retrieved 2020, from https://sketchfab.com/3d-models/hestia-3d-figure-fanart-6abb95d56bca486793df6f99c9c73788
nlongchamps. (2016). Creature Scales Retrieved 2020, from https://share.substance3d.com/libraries/986
Alec Chalmers. (2016).Reptile_Scales. Retrieved 2020, from https://share.substance3d.com/libraries/733
silentecho. (2016). fishscales. Retrieved 2020, from https://share.substance3d.com/libraries/851