FMP - Finale

 

With all minor adjustments complete, I think it’s safe to call my FMP finished!

This has been a long project, a lot more demanding than I anticipated, but I learnt so much along the way and I’m quite pleased with how everything has turned out. I wanted this project to not only be high quality but also showcase the type of 3D work that I want to do in the industry, and that I think I (currently) do best. I think having that at the forefront of my mind when deciding what to do was what kept my motivation high continuously to keep making high quality work – it’s always more fun to do what you like.

 Overall I feel like my understanding of textures has improved a lot from when I started to when I ended. From the feedback I received plus experimenting myself and pushing how I worked on my textures, I think it shows that there is now a more solid understanding of showing contrast and tonal range within my textures which I was lacking in before which made my models feel flat at times. Really enhancing the shadows for example to push the overarching values and understanding how to compensate for that in colour has really made my textures pop. I feel like they’ve just taken a step up in quality.

 I also think my time management has improved quite a bit with this project. Having the Gantt chart really helped me- having an actual visual of my time meant I was more inclined to stick to it rather than be half dedicated to a loose schedule in my head. Admittedly I didn’t follow the chart exactly because as things do, plans change. But having the big milestones blocked out meant I gave myself mini deadlines which I needed to stick to so the in between didn’t really matter. And I did meet all those milestones! I’ll definitely be planning ahead like this for future projects because it not only helped me but it helped others understand my workflow better.

For more self analysis however, I do wish I was able to add some animated portions to my models. The VFX I couldn’t figure out how to achieve the look I wanted to match with the hand painted textures of my models and, in short, I’m not an animator so creating an animation for my characters felt way out of my comfort zone to the point where it would downgrade the overall look (also the rigs of each model were very specific so even premade animations wouldn’t translate well to them). If I had more time I would want to allow myself the chance to get a proper understanding on how to give that animated flare to my models. I think having them posed like they are is fine but to have that moving aspect would’ve really made them feel alive and magical. 

But I’m not unhappy, I’m very happy with the progress I’ve made. I think this has been a good, big project that I’ve needed to really hammer it into me how I work and what I enjoy doing.

It’s been a fun ride, but I’m more than ready to hand it in!

FMP - Easter Break #2

 I’ve made very specific creative decisions with my FMP from the get go which I’m glad I have because it’s given me a chance to enhance and refine skills in ways I hadn’t thought about before. Mainly this being about me choosing to create diffuse only hand painted models as opposed to PBR models. There are understandable limitations to diffuse only models like them not being as procedural as PBR so editing them can be hard or them not being able to carry as much depth to them texture wise, etc. But considering I had to ensure I was creating a project that could do within 4 months and produce work of a professional standard, choosing diffuse only was the better option for me.  I’m not opposed to PBR nor do I consider myself unable to effectively make a PBR model, it was just I felt that my hand painted skills at the time of deciding what to do for my FMP were stronger. I didn’t want to end up fumbling a PBR based project when I knew I could do a hand painted project much more efficiently.

That being said, I still needed to include an aspect of PBR to my project to showcase that I do have these skills that are sought after in the games industry.

From previous feedback and some thought of my own, I’d decided to make an additional PBR prop that I would have as another focal point for my overall scene. I’d planned to have my characters displayed as a sort of mini diorama in Unreal Engine 4 just for aesthetics (I’ll do individual renders of the characters as well since they are the main priority though).

The prop I’m creating is a gemstone, similar to the ones that appear on the all the characters, with some gold and chains attached to it. What I knew I had to be careful with though was keeping the style uniform with the characters so the PBR material didn’t make the prop look out of place against the characters. I’d looked into heavily stylised PBR and saw a character by Tim Paauwe on Artstation where he’d taken to hand painting it completely but then created a PBR material also. Had I seen this post 4 months ago I might have been inspired to try this for all my characters but what’s done is done so I’ll save that for a personal project. But I could apply what I learnt from looking at his posts to my prop.

It was very basic design wise- I created the low poly version in 3DSMax first before going into ZBrush and applying my levels of detail. Nothing too complex, just standard scratches and cracks in the gemstone and gold. The unwrap was all done in Max and I baked a normal, AO and curvature map in Marmoset. Typically in the past though whenever I’ve done PBR stylization, I’ve used Substance Painter to create my textures. But this time around I used 3DCoat, mainly because 3DCoat was what I was more comfortable getting the painted style I’d achieved for all my characters with. It could handle PBR just as well as Substance though.

I completed all my painting just as I did with my characters. But with my final albedo, I went into Photoshop and rendered it out in grayscale to create a roughness map. In 3DCoat there was a means to paint in roughness right then and there but I thought converting the albedo would be the best way to enhance the hand painted textures without the PBR material making the whole texture feel misplaced. I packed my AO map and metallic map into the RGB channels alongside the roughness map straightaway when I’d gotten it adjusted to my liking.

 

With this though, I had to immediately test out how the textures would look as a material in engine so that I could make adjustments if necessary.

At this point, I implemented my whole scene into Unreal, characters and all. I’d posed my characters previously so staged them around the gemstone and assembled my lights and post processing. My PBR assets worked well against my diffuse only assets- I’d tweaked my post processing so that the strength of the diffuse maps were more impactful and it didn’t look overcast against the PBR materials. I found this tip from an Artstation post by Linus Asplund.

I created smaller PBR assets alongside the gemstone like rocks just to fill up the scene some more and it really turned out to be a miniature diorama, much more to the scene than I’d initially thought. But it works well, it brings it all together. Having made these assets though, I thought I would have time to create some VFX on aspects like Judgment’s tail and whatnot. But it was honestly a different ballpark that I wasn’t familiar with and having to spend extra time making these props meant I had less time to research and figure out how to make the VFX I imagined. I tried to do something simple with just a panning texture in the node editor but it didn’t look good to me so I scrapped it. And with the deadline approaching, I don’t think I’ll be able to make what I want. Which is a shame but I don’t want to add something that looks cheap or half-done to my models as that could detract from it as a whole. I’m fine with them just being stationary.

FMP - Easter Break #1

 

With it being the Easter holidays, I do want to give myself some down time since I have been working on my FMP for near 4 months now. But I still have some portions of the project that are important to complete so I’ll keeping working for the time being.

The main situation now: rigging and skinning.

In hindsight, I think it’s better that I’ve done all the rigging and skinning at once rather than individually during the whole model creation process like I’d planned. It just means I’m in the right mindset to easily hop between the models if I need to make tweaks on encounter any issues.

Speaking of issues, my original thought process with setting up my rigs was already an issue. My knowledge of rigging in 3DSMax had been to use the biped as my rig, adjust according to the shape of my model and then skin from there. This works for basic figure models- a standard human with 4 limbs, a torso and a head. But that wasn’t the case with all my models. 2 of my characters didn’t have legs and the one that did had goat legs. That was how I discovered the limitation within 3DSMax where you can’t edit the premade biped, deleting a bone would delete the whole skeleton. So instead of using the biped, I had to make the rigs for each character from scratch using the CATRig tool.

This was a new method for me but it was a necessity to figure it out for the sake of having poseable models. I created Justice’s rig first as I thought that would be the most complicated since his legs were specifically shaped. Making sure the joints bent at the right angles was a run of trial and error to be honest, ensuring the parenting of the bones was correct and that it would function correctly when in a pose. After slowly building the rig for Justice though (with a few restarts along the way) I got to grips with how the CATRig system worked. It wasn’t overtly complicated, it just seemed so because it was new to me. Having to resort to this method made rigging take longer than I thought.

After completing the first rig for Justice, I decided to go to skinning him straight away in case I met another bump in the road. Skinning however functioned the same as always- I applied a skin modifier to the model, attached the bones, and then started adjusting the weight paints accordingly. This was much more straightforward and I was able to gradually skin Justice as I was used to.

Posed Model

I followed this process for the other 2 characters: creating a CATRig for the High Priestess then immediately skinning, then the same for Judgement. Naturally I was faster the second time round creating the rig and even faster the third and after having done it 3 times, I think I prefer to use the CATRig rather than the biped. I think the only benefit to the biped is that it’s quickest if you just have a very standard character. But even if all my characters were just 4 limbed humans, the CATRig is so much more customizable and easy to manipulate, it really shows the limitations of the basic biped. I’m glad I discovered this method as I’ll probably use it now going forward if I’m to rig in 3DSMax again.

FMP - Week 13

 

This week for my FMP I tackled texturing Judgement’s tail and did some minor adjustments to the textures overall.

I created a very basic shape for the tail by using a cylinder and adjusting the scale of it, editing the mesh towards the bottom to give a little flick to the end of the tail. I felt that this would also be easiest for when I comes to rigging and skinning. I unwrapped it as a flat singular island then took it into 3D Coat.

For any other texturing I would have gone into Photoshop first to lay out the flat colour but instead I thought being able to lay out the flat colour by painting on the 3D form would be easier for getting a better sense of the overall shape rather than kind of having to guess in Photoshop before going to 3D Coat and correcting it.

 

I painted on the forms in a very carefree way, going through and erasing parts to look like natural forming holes and leaving some remnants of erased parts in various opacity to give it a more mystical look to it. With this texture being somewhat transparent, I had to be considerate no to go too harsh colour wise because that would show through from all sides of the model. If I made the tail incredibly saturated, the overlapping transparencies would blow that saturation way out of proportion and it would look bad. I spent quite a lot of my time adjusting the opacity of the main shape’s colour, adding elements of opaque details sparingly throughout the tail as to not make it look cluttered. This texturing process was more challenging than I anticipated but I think what I ended up with would work well for the overall monster.

This week I also went back and completed the texturing for Justice’s weapons. This was much more straightforward in comparison to the tail as I just approached the same way as I do the main bodies of the characters. I found the scales however to be quite interesting to texture because the overall shape of the asset had quite a lot going on with it since there were the chains, plus the base and handle, and the actual plates. Keeping everything uniform whilst not making it visual overwhelming took me going over and simplifying some texture work so that from afar the textures didn’t look jarring.

 

I was able to complete the tail and weapons though within this week so I’m pleased! All 3 characters have officially been fully made. Although I’ve deviated from my original schedule in some areas, I’m satisfied with where I’m up to and I’m really happy with how they all look!

Pay no mind to 3DSMax struggling to preview the alphas

FMP - Week 12

 

As is the workflow now, I headed into Marmoset this week to bake my high poly mesh onto my low poly for Judgement before going into Photoshop to lay down the base colours.

Having completed 2 models now, I was faster at getting a sense for what colours I should use for the base and how I can push and pull these colours (especially on the gold areas). I had to be a bit more considerate on the skin because I wanted it to transition between a light, flesh tone on the front to a darker blue tone on the back. The transition between these 2 colours on the model could get a bit muddy and awkward when painting so I had to thing of how to combat that. I did so by taking the red that I’d introduced after feedback and using that as the bridging colour between the two. I think this worked best since it was a warm colour like the flesh tone but was strong in saturation like the blue tone so it felt unifying and not out of place or overwhelming.

The overall painting process went smoothly though- I began with the torso as that was the biggest area to handle then did all the gold pieces at once before going over everything with some more detail. I spent most of my time getting the back to look correct, really trying to get the muscles to read well but not look too detailed to the point it was blatantly obvious that the model was symmetrical. That was probably the biggest challenge when it came to painting this time around. Since the High Priestess and Justice had asymmetrical textures, I could be as adventurous as I wanted (within reason) when it came to texturing details. But with Judgement I had to make details much more subtle or else the whole thing would look odd.

There were some minor tweaks made along the way as I textured though that differed from my original concept. I don’t think these changes detach anything from the character, I think they just weren’t really necessary after looking at it in 3D as a whole and/or they made Judgement look strange in some way. For example, on the concept there are holes in the helmet that look like eyes. It may have been a case of poor execution but after texturing them into the model, it didn’t read as well as I’d hoped, especially from the front. It just didn’t look good so I decided to retcon it. Maybe in future I could have experimented more with making some kind of eye socket but since my time left is limited on the project, getting rid of it was simpler than spending more time fiddling around.

I’d also done this texture pass without the tail on the model. I’d decided that I would go for the alpha approach of making what is effectively a tube as the tail and making it translucent with a misty texture to it. I’d started the textures on the main body before coming to that conclusion so just decided to carry on as such and I’d texture the tail separately which isn’t an issue.

Overall though I like the textures on Judgement so far, I feel like I’ve really refined how I approach hand painting textures and am much more confident with understanding how to approach texturing now than when I first started.

FMP - Week 11

 

Retopologizing!

A trial as standard as ever.

As all the metal portions to Judgement were already low poly, I only had to retopologize the torso + arms and hands. This wasn’t particularly a hard task, it was just a means of taking major shapes into consideration since this character has a lot of exaggerated forms. I also considered what forms I could paint in rather than build into the topology.

Swiftly moving on from retopo, I unwrapped the model as per usual. I decided that I could afford to make Judgement symmetrical texture wise and I’d add asymmetry with some minor additional assets on top (like the gold on his back and the chains). This would also save me time when texturing the model since I would really only have to paint half of it.

Within this week though, I went back a few steps and made the weapons for Justice – a sword and scales. I box modelled this straight in 3DSMax since that was the fastest option for me and then unwrapped them at the same time as Judgement. The overall design and shape of them weren’t too complex so I wasn’t worried that texturing them would take up a huge portion of my time. Regardless though next week I’ll begin texturing on Judgement before all else so he’s at least up to par with the other 2 characters body wise.

FMP - Week 10

 

Starting this week off in ZBrush, I was back sculpting away on Judgement, my monster character. Having started the previous week on the torso, I continued refining some of the shapes and trying to get a well balanced sense for the proportions. I knew I had to push the proportions to make his body more exaggerated and stylised but it was finding that perfect midway point between good and bad exaggeration that was the tricky part.

His torso was the only ‘fleshy’ portion to the monster since his head was essentially armour and his tail is just a magical wisp. Getting this section of the model right was crucial to unify the whole thing together. It would be painfully obvious if I looked off.

I played around with scale mainly when it came to the proportions. Things like the size of the shoulder muscles, the skinniness to the waist, the length of the arms were the areas I felt I could alter best to look more interesting. I also took to paying close attention to the reference I’d gathered to see how others had created shapes that read well in a monster form.

Aside from that though, once I had the base shape I knew I wanted, I exported the torso to 3DSMax and started working on the hard surface features. The head and the spine were the main features that I had to carefully figure out how to model to get the shape I wanted whilst keeping in mind topology. The head in particular I constructed quite carefully since I wanted to have the mouth open so I had to be considerate of the teeth. The process as a whole wasn’t necessarily challenging but it took me approaching I slowly in order to get the shapes right and not distort them accidentally.

 

Once I’d completed all the hard surface assets though I imported them all back into ZBrush where I gave another detail pass on top of them (and the torso). Additionally I added the tail. However, when it came to the tail I wasn’t fully sure how to approach creating it because I hadn’t yet decided how I wanted to present it. Whether it should be a whole, fully textured portion to the monster or something like a translucent alpha, etc. I also had to keep in mind rigging in the future. For the time being though I implemented a tail that was there mainly for me to get a sense of the scale of the character as a whole and potentially give me better judgement as to how to approach making the tail in future.