Author: Sam Payne

Animation!

Finally, animation, most likely the hardest part of this project, considering my animation skills are incredibly limited, and I have not been taught anything regarding animation in my time. I started off by approaching what I wanted to do, and I settled on the idea of having O+ throw one of its teslas, as that would make sense for a sort of long attack motion. As a reference I used a video that was a tutorial about how to pitch a baseball shot, as I feel that type of action does convey a sense of strength and its also a bit more complex than just a simple animation like an idle or a run.

I animated in blender for this, and luckily my previous decisions in rigs made this whole process a million times easier for me, so im glad that I made the rigging choices that I did make. Especially with my decisions to parent bones on an offset, considering those can be moved freely and are not explicitly bound to the body, giving me maximum flexibility to work with, even including the ability to work with smear frames.

Here you can see my workflow for animation! My main consensus was to work on the main body for O+ first, before going and animating the suit as that needed more manual tweaking due to the design (armholes aligning with action, legs matching up with slot, etc.), this is quite noticeable from my action editor at the bottom. Interpolation works quite well for an instance like this, especially considering I have a complete lack of fundamentals when it comes to animation, so having the software interpret features such as inbetweens was a massive help for me to make this at a good speed.

However, I do wish that I had took snippets from the video itself and imported those into blender, that way I could possible have a much better and more accurate representation of what I should’ve done instead of just eyeballing it. I will attempt in my free time to recreate this action by using the video that I reference, but more directly using it. I think the main thing that bugged me the most about animation is how it didn’t look smooth, it was merely awkward. Of course that would be something that is fizzled out with time and experience but it does make me feel unnerved that this is meant to be AAA style work & workflow with choppy animation.

Render!!

With animation out of the way, it was time for me to render it out and call an end to this project! I did try making use of unreal for the animation, but I was having a lot of issues with assets and it was quite confusing. For some reason I’d import in my skeleton and mesh as an FBX, but when I would close unreal or restart unreal, even after multiple times saving, my model and skeleton would dissapear from my content browser, but all my textures and materials stayed put. It was really confusing and made it quite difficult to figure out what was going on with my render, especially considering that at some point in time the animation just stopped playing in the sequence editor. Even when I managed to get it to work and tried rendering it, there was an issue where my model would bounce up into the air, alongside with the suit, however that may just be due to the models not being anchored, but I can’t say for certain.

As a result, I had to abandon the premise of using Unreal for this project, and make use of blender. Now blender’s rendering engines is the only thing I will ever critique about blender. They can be used definitely, but it is a lot of meticulous tweaking and adjustment to get exactly what you need out of a render, I even struggled with it in Games Art Pipeline when I was rendering a character in Year 1. However, this time things went much smoother! I’m not exactly sure what changed, however, I felt the lighting this time round turned out much more dramatic, even if it was just a basic showcase of the character or its animaion.

Final Rendered Animation!

Last but not least, is the animation that I ended up rendering out. Whilst you can see my keyframing process above, you are unable to see the actual turnout of the animation in render. So here it is!

I am really happy with the lighting, however my concern is if it fits with the briefing of overwatch. Whilst the stylisation of eveything definitely does in my opinion, maybe the lighting is too harsh and lacks vibrancy. I am not sure at this moment, but I am still incredibly happy with the render, especially compared to last years result. Alongside this, the glass visor I made actually renders properly as glass, and not anything else, which is something I have been trying to figure out for years, and now I finally have!

In the end, I am quite happy with how O+ turned out, however, this specificially applies to only the main body for O+. I definitely feel like I could’ve done so much more work on its suit to make it match with that Overwatch level of detail. Whilst I’m happy I used height mapping in an efficient manner, hard modelling would’ve been so much more satisfying to construct. However I do also blame my simplistic design for the suit at the beginning of the project, as a lot more work should’ve been done on those thumbnails to make designs stand out more than others rather than subtle changes.

Alongside this, I was disappointed that I didn’t get to add the wiring in the final model. Whilst I had it all setup and layed out, I didn’t work on it as I should’ve, and that caused issues when it came to rigging and animation, for future reference however, any wires or coils will be modelled flat and then rigged into shape, thus making it easier for me to work with those types of models and features, as supposed to this project where I faced a lot of deformaties that I was unable to fix the weights for in the time that I had provided for this. Animation was also a reallly rough time for me, as mentioned I didn’t want to go for something simplistic, so I decided in the end to go for a more emotive and impactful motion, but it didnt feel that way. A lot more research needs to be done for animation or I will drop the skillset, considering its not a job I wish to apply for in the future, as my focus is in creation, not moving.

Thank you for reading! As of this point Year 2 is officially finished for me, I will be enjoying my summer but also working on preparing for 3rd Year, especially with my Final Major Project and Thesis coming up, its important to rest but I also need to lock in, happy summer!

Texture Time!!

With all the harder stuff out of the way, its time to take a small break by texturing the models! First count of buisness is to texture the base body of the octopus. Also at this time I have just come up with a name for my character: O+ or O-Positive. I thought it would be a fun name for it, considering its a medic, so O+ for blood type, and its an octopus so it kinda fits in with having O for it.

When it comes to texturing, I started looking at Hammond/Wrecking Ball’s texturing for Overwatch again, as I was having difficulty figuring out the approach I wanted to go for. Whilst I did have an exact colour palette based on an image I found for my moodboard, I wanted to make sure I painted it right. Taking a look at Hammond, we see a lot of flat colours, but all of those blend or are of a very close variation to each other, or simply it is affected by the lighting.

Something I came to realise whilst studying Overwatch’s character textures, is that a majority of them are flat colours, and there doesn’t seem to be much use of height/normal mapping. Whilst that may be an exception in Character SKINS, its much more noticable in the main versions of the roster once you take a look. However, they still manage to create really detailed and in-depth textures for their characters.

Starting with the Octopus body! I had my colour palette as previously mentioned, so I started getting to work on applying the base colour, which was orange, and adding some close values with a rough brush to give it a feel of texture in its colour.

Following on from this, I added the head dots as and I started to look at using gradients for the arms and legs, as that would be the best way to display that range of colours in a really pretty way, but also in a way that fits the style of overwatch. I had some struggles with originally planning and putting the gradients down, as I tried making use of position based gradients, but I was unable to mask parts of the model out of it or move the gradient to my pleasing.

As a result, I did a bit of research into it, and found out that there was a way to create gradients and position them by hand, and it was to make use of a fill layer using a type of projection with a white mask.

Having a bunch of these fill layers in folders resulted in this gradient look! As the hardness being lower means that the colour fades out towards the boundary anyways instead of being cropped exactly at it. This was a much easier way to about things and was incredibly satisfying to see come to fruition.

Also I decided to paint on the head dots, I was thinking about doing it for the body also but I decided against it, as it would mess with the look of the gradient, and I was looking into smart materials I could use as a sort of skin to make it feel more octopus-like, as it being non reflective or wet looking was really thwarting with my senses. I would argue it also adds to the stylization of overwatch, and the roughness is very easily tweakable if it is too little, as I wouldn’t want to defer from the style.

Note for the skin overlay texture, I put it at 10 opacity as it did have a colour property, now whilst I could just remove this, the colour fits quite well with the general tone of the texture so I decided to keep all the properties as this looked the best and was easy to tweak if needed. This concludes the texturing of O+ itself, now its time for its suit!

Suit Texturing!!!

With colour and life brought to O+ itself, it was time to do the same for its suit, and make it a whole finished character model! I made heavy use and reference of Wrecking Ball from Overwatch 2 to create my suit texture. A majority of it from what I have noticed is just flat colours with metalic and roughness adjusted, even features such as polyfill would reach the same look. The only other things is that the metal appears stained, so I can just use a brush and smudge to get that stained painted effect.

Now before I do go any further, I do have 3 seperate versions of the final suit that I made that I really didn’t like that much. I think this texture was the hardest for me to face, considering there isn’t much topology or showing detail on this suit, overall I think its quite bland, so I had trouble applying different colours and making it look detailed. I made extensive use of height layers later to compensate.

As you can see, the most substantial change was between v1 and v2. Mainly I realised that the design I had created for the suit at the beginning of this project wouldn’t really fit Overwatch’s design, unless I was going for a character akined to that of Junkrat or Junker Queen, however I wasn’t really going for that, I was exploring the space animals in suits type of thing.

Generally v1 was just not working and I felt it was generally really lacking of a texture, again I didn’t have much topology to work with as I didn’t sculpt the suit for details (silly me) so I had to change my design on the fly essentially to fit and look more like Hammond’s Suit. Luckily for me, Wrecking Ball was the perfect reference so it was quite easy to come up with something to make that change, and you can see it further develop across v2 to v3, especially with the middle panel. I made use of height painting to make it look like wires on the surface, and also layers in the panel area.

There is one more iteration and thats the current one I’m using, the only major difference is the levels of metallic and roughness, as I felt that the previous iterations were too shiny, and Wrecking Ball isn’t overly shiny due to the stained nature of the metal, as a result I turned the roughness up and it seemed to do the trick just fine! Also there is a texture on the visor, its just a glass one that I alpha clipped so I could continue to see O+ for testing.

Something that does really urk me, is that the textures look infinitely better in Substance Painter than blender. This will always be a criticism of mine, but I really dislike how blender processes specific textures in viewport, typically metallic ones. However! This is all I have for the week, its time for the dreaded animation! Until week 10!

Rigging, Continued…

Now with our armatures in place, its time to weight paint my character! For the most part, I made heavy use of blenders Automatic Weights function when you go to parent a model to an armature, as it typically provides quite accurate and strong results. This was the case for my base mesh of my character, as it moved all the appropriate limbs very smoothly and in a way that fit its character.

However, the suction cups seemed to prove a bit challenging, considering each tentacles suction cups are in very close proximity so its quite easy for the weights to get confused. For refenence, I did not make use of automatic weights for the suction cups, I used Weight Transfer as that typically is more accurate and if there are any issues, the weights are easy to correct. I will provide the thread that I used to learn this process, as it may be useful if you are ever rigging! (Check the comments)

For now, take enjoyment in what I ended up having to fix, as it took a sec to try and manage all the suction cups in their places, as well as the fact that sometimes I would just have a specfic vertice or two that decided that they didnt want to be weight painted.

On the bright side, the finished result looks great and functions amazingly! I was really happy that everything turned out so well for it. Honestly at this point if I had just designed the squid I would’ve run with it for the rest of the project. Aside from this it does has its suit that also needs to be weighted, and luckily it was a lot simpler.

Considering the robotic nature of the suit, it needs to be heavily weighted to react to the bones, meaning no geometry deforming, so all red based on its bone. I made sure to specifically create bones for each panel or part of the suit, so that I could have full exact control of each component of the suit without any issue.

This same structure was applied to each and every bone/component of the suit and proved to work quite well, as the HipControllerBones (the Middle Ones) did a great job of splitting the suit into 2 as a lot of the bones are parented by offset to those central bones. Having the squid show in there also shows that its a seperate rig, otherwise the squid would also be a dark blue like the suit is, but its just displaying face orientation in a light blue.

I also weight painted the small armbands that go around my characters arm tentacles, it was also quite an easy process as I just had to transfer weights from the main body over, and make some small adjustments as it was using too many vertex groups/bones to rig when it wasn’t necessarily needed. Same case with extra bolts I added for detail purposes, except those were all red to be mounted to the exact part of the suit they were placed at.

But for now! Everything is properly rigged and working nicely! So I can now move onto two of the most significant parts of this project, texturing and animation! See you in week 9 for texturing!

UV Mapping!

Time to UV my models to get ready for the texturing process! I decided to move onto this first as it was the quickest task for me to do, considering I just have to mark my models seams, before Smart UV Unwrapping and marking the boundaries for the UV to be even. Starting with the squid itself, I had 2 iterations, as the first had some inconsistancies in the bottom left of the UV. Whilst it would’ve worked fine in any case, I still want to make sure I create efficient UV’s for the future work and portfolio reasons, so I remade it.

Following from this, I UV’ed the suit, smaller details such as the bolts on the suit and the suction cups I didn’t UV, as going into Substance Painter, it would automatically adjust. However for future reference, I will UV the singular one before I start copying and pasting it around, this also applies to the bolts. Helmet also had 2 Versions, on before all the extras and details.

Rigging!

Following the UV unwrapping, I decided to create the base armature for my model! Weighting it will be an issue for next week as there are a lot more technicalities I need to consider when it comes to weights so I will work on that next week instead of this week.

As you can see, the rig itself is quite basic, this is due to the motion of mind that it is an octopus/squid type character, so as a result I have no need for functions such as Inverse Kinetics and CR, CL and Controller bones considering I can rotate these bones on individual origins when I select the whole chain, resulting in the same type of squid/octopus flow that you could expect in real life. It is interesting however having to rig a creature that has no bones in real life.

Alongside this, I wanted to create a seperate rig for the suit, as I feel like weighting it to this would be a nightmare, as well as the fact that the multiple bones would split the geometry of the suit, unless I was rigging one part to one bone specifically, which has its own set of problems from experience. It also allows me to have much more control over what i’m doing when it comes to animation later down the line.

For now this is all! I will delve into weight painting next week, followed by texturing, animation and rendering processes! See you in week 8!

Suction Cups…. Continued…

As I mentioned last week, I spent a lot of the time this week working on adding suction cups to the arms tentacles of my character. I definitely feel these turned out much better, as I did have a feel for how the array structure went, so I sorta optimised it to be a bit quicker. Also, unlike the legs, there wasn’t any sections with overlapping verts or deformed shapes, unlike the legs that had a plethera of messes that I couldn’t really fix.

The polycount it self was fine for what it was as the mesh was still sharp and detailed enough to resemble what I wanted, whilst 54k is not ideal, there is a lot of suction cups that need to be accounted for, so it was already impossible to make this really low poly considering the round shape of a suction cup. I will provide images below;

Retopology!!

Now before coming into retopology, I was already intimidated and frustrated due to the time it took for me to create and apply suction cups to the body of my character. I had heard from a friend that they used a mirror and welding strategy (in blender) to essentially weld new polygons, to the old shape whilst staying in different objects. Unfortunately I couldn’t figure out how to work that method efficiently, so I went and researched into some addons that act like Maya’s Retopology Mode. It was at this point I was graced by the heavens: Retopoflow 3.

I had previously downloaded this addon to try and make use of it in a unit from last year (1st 3D character Unit) however, I couldn’t really get the grasp of how to use it, and therefore left it. But upon rediscovering it and watching some youtube tutorials on the basics for it; I was hooked. It managed to provide me with such an efficient working environment and workflow that retopology was honestly the easiest thing I had to do in this entire project, and was actually something I turned out to enjoy, as maybe listening to music whilst retopologising a model isn’t that bad after all. It still took a while, so this was done throughout Week 6 from here on out.

There were a few tools that I made extensive use out of, such as Strokes, Contours, Polypen, Knife and Relax. All of these have an essential purpose and areas of which they were more effective than others in the project, in particular Contours was incredible for the tentacles, as it creates an edgeloop around the model which then can be linked to another edge-loop at the click of a button. Polypen and Knife were useful for fixing topology issues, in which other cases I made use of the Loops tool to add an edge loop if something was off, before relaxing it with Relax and fixing it further.

It didn’t take nowhere near as long as I anticipated, and to some extent I am happy about that, but I also enjoyed it so much so I feel conflicted. I will make sure to do more projects like this one to keep retopologising, or just offering to retopologise my friends work. Alas, here is the final result compared to the sculpted one!

As you can notice, there is essentially no difference, the only thing being how blenders lighting affects the top part of the head, but it still has the same ridge that seperates it, and the same indent on the skin between the legs. Tentacles and Eyes look identical.

Here is the complete current pipeline of my character! From start to finish at the moment, but now is onto the other main parts of this project, such as UV unwrapping, texturing, rigging and animation. Until Week 7!

Zbrush time!

Its been a while since I last used Zbrush properly for a project, I believe it was during Games Art Pipeline last year, when I was working on the monster for Peter’s Unit. So lest to say there is much catching up that I need to do. However, I remember the essentials thankfully, so I dont feel as scared as I would be if it was the first time I had to use the program.

I started by increasing the polycount of the model by making use of the DynaMesh function and turning the subdivision resolution up high, to about 500, resulting in a model with around 130k polys, compared to a 10k base mannequin. Now whilst I could go higher with this to add even more detail, the body of the squid itself doesn’t need to inheritly be detailed, as texturing solves a lot of the issues with it, on top of its suction cups on the tentacles being a whole seperate entity as it would be practically impossible to extrude those out of the base mesh, let alone sculpt all of them.

Whilst I’m not particularly proud of the eyes, its something that can be easily fixed, i will just have to refer back to my moodboard to look at how a squids eyes work properly as supposed to this look. A case to be made in Overwatch, Hammond whilst incredibly stylized has eyes that make sense, so its important I replicate this feature in my sculpt/model.

Suction Cups…

Now for the suction cups, I made a base singular cup in blender, and instead of just planting them around and mirroring it as any normal person would, I decided that it would be a great idea to import it into Zbrush, then use array mesh, and manually split the new merged suction cups as seperate subtool before rinsing and repeating. I did this for 8 limbs, and it took infinitely more time than if I had just done it in blender, so from here on its just going to be a lot of wasted time and small progress sadly. I will show how it looked in Zbrush and blender, but I was not satisfied with my pipeline at all.

I repeated this same process until I had completed all 2 legs, as I could make use of the mirror and weld tool (thank god) to minimize the amount of work I had to do in terms of manual array work. At this point, why didn’t I just turn to blender to place the suction cups more efficiently; sheer will to get it all done in ZBrush. Also I was at the time thinking of retopologizing the suction cups into the base mesh, however, it looked absolutely terrible and destroyed all solid shape and foundation of the suction cup, and sorta just squished it and made it mushy.

This process was generally, quite painful, and I haven’t sadly got many screenshots of my process here, but if you know anything about array mesh, it was slow and monotomous, so I’m very much glad its over. I continued working on this process over Week 5 for the arms as well, so I will mention it slightly and provide blender screenshots, as I dont have any Zbrush ones, till week 5!