Even though I wasn’t able spend as much time on these pieces as I would have liked due to health issues this term, I’m happy with the results and how much I learnt.
For 3D animation, I was able to get more familiar with body mechanics and figure out what I want to improve on going forward. I also learnt a lot with the acting shot, including a lot of nuance with how the face should move when expressing emotion and talking.
For the experimental project, I was able to learn so much of Marvelous Designer and make a head start on creating the clothing for my FMP characters. While I would have also liked to have experimented with some visual styles, I’m glad the project allowed me to experiment with a new software that I might have felt too time stressed to do otherwise. I also really enjoy using Marvelous Designer despite the difficulties I had with it, and seeing the similarities between the interface and my experience physically making clothes and clothing patterns.
Took some time to get body suit not intersecting with itself and the model, but by adjusting both the fabric thickness and the skin offset of the model I was able to get stable simulations every time. However when I went to start simulating over the animation cache, something happened just between switching tabs that brought back some of the intersecting issues. I have a feeling its something to do with the tack on avatar tool, as that was the last thing I did before moving onto the next part that might have needed more testing. Removing all of the tacks seemed to fix the issues again, but I couldn’t figure out why the tacks caused these issues in the first place.
While trying different recording settings for the simulation helped improve the output a lot, I could still see a lot of issues that would have been fixed with the tack on avatar tool. I decided to just try and put the tacks back in place for the simulation to see if it still caused issues.
There were some of the same issues as before with the tool, but not as bad once the simulation settled a little. A lot of improvements could be made overall, but I think for future FMP work any tight fitting clothes might be better to model and skin/wrap deform to the rig or add in texture, rather than simulate. While I have seen some amazing examples of tight fitting clothing in Marvelous Designer, I don’t want to spend too long learning how to properly do what I’m realising is a tricky task when I could save time by using other techniques.
Once the simulation was complete and I was satisfied, I exported the garment to Maya so I could match up the clothes with the rig and set up the materials for streamlined UE5 import. This also gave me chance to fix the UVs (which didn’t export out of Marvelous Designer as I expected) and take a closer look at the pros and cons of simulating a tight garment.
Importing to UE5:
When importing to UE5, I realised I had forgotten to make the changes needed to the rig for smooth/streamlined import to UE5. I initially thought to use the editing rig scene I used in Term 1 with these changes already in place, but the rig I used for this project was an updated version and when I tried to reference in my old rig scene edits the animation messed up a little. The changes took a while to figure out but in themselves were minor, so I decided to just redo these with the current rig scene. Since I already knew what changes needed to be done, the process didn’t take as long as it did for my Term 1 project, but was still a lot more time consuming than I would have liked.
I did save a little time by reusing the materials from my Term 1 UE5 project, but I still had to repath and tweak them a fair bit (plus make new ones as the clothing/shoes are different for this project).
I also at this point set up a sequencer track with cameras, shots, and the alembic caches. During import I separated out the garment from the model so that I could have better control over the two components.
Making trousers and dress
While I originally wanted to make trousers based on the concepts I drew up, I ran out of time and decided instead to create simple baggy trousers that would work with the bodysuit. I plan to go back later and create trousers based on one of my concepts when I have the time to do so.
I had a little time towards the end that I debated spending on working on the set design of the window scene. However because the focus of this project for me was experimenting with and learning Marvelous Designer, I decided instead to make another simple garment to further improve my skills with the software. I decided to make a dress, and tried to incorporate some elements from my clothing concepts so I could also test these without needing to make the full detailed garment.
I didn’t notice until I brought the first dress cache into Maya, but the finger clips through the fabric early on and causes the fabric to bug out a fair bit. I went back and forth with the simulation settings trying to fix this, but the best I was able to do was to double the simulation quality (from what I used for the other clothing) and the issue is still present unfortunately. I tried caching out the walk animation from Maya at 48FPS, and I tried adjusting the fabric settings for the skirt, but neither of those (or both combined) seemed to make a difference. Overall I am still happy with the result as I learnt a lot from just this one garment, but it is disappointing that I couldn’t get it to simulate correctly.
Putting everything together
From there, it was the same process of setting up the Marvelous Designer exports in Maya, then exporting them out of Maya to UE5, putting basic materials on the asset and adding it to the sequencer along with the alembic cache.
While I’m still working on the bodysuit in Marvelous Designer, I wanted to animate a walk cycle to use for simulating the clothes in Maya. I added preroll frames at the beginning with the default pose keyed in. This is so the clothes have time to settle and move into the initial pose before the main animation begins during the simulation. I exported this out of Maya as an alembic cache, and tested quickly that the animation will import to Marvelous Designer which it does.
The next steps will be to finish the bodysuit and try simulating it on top of the walk cycle.
I also modeled a basic window scene in Maya and laid out the UVs so everything is ready for export to UE5, where I will add materials, lighting, and render from.
This week I managed to do a lot more work on the clothing concepts, and while I haven’t decided on a final look for the FMP character I’m happy with the designs for progression onto learning Marvelous Designer.
To start, I exported the Minerva M-Rig out from Maya in a neutral pose, and imported that into Marvelous Designer. I decided to try making the bodysuit first, as I assumed this would be the easiest garment to make. While the basic pattern was pretty easy to mock up, refining the cuts of the cloth along with getting such a tight fitting garment to simulate correctly proved very tricky, especially as I was completely new to the software.
In the end, I was able to cut the pattern up in a way that resembles my concept for the item, but I am still struggling to get the garment to simulate on the model. I am already wondering if tight fitting items would be more suitable as skinned meshes within the rig rather than simulated on top of the animation, but I want to keep pushing forward in order to learn as much as I can since that is my overall goal with this project rather than the visual output.
This week we were shown 360 videos. After creating animation for AR in term 2, I have wanted to improve on traditional 3D character/creature animation implementation with 360 output, as I wasn’t satisfied with the results due to the differences in how the animation is viewed not being considered as part of the process. However overall it isn’t one of my current primary interests, so this will most likely be a passion project I do in the future outside of showreel work.
Project Progress
I’ve still been struggling to make much progress, and I am currently debating the schedule for both this project and my FMP with the limitations I am now realising more.
While I want to create cool visuals alongside 3D Animation showreel pieces, I lack a lot of the technical skills needed to understand even what is possible for my project, let alone how to do it. It is a bit overwhelming with the time constraints, and I’m starting to think it might be best to focus more on learning Marvelous Designer for this project so I don’t overload myself.
As for the wider picture, including FMP development, it would be worth sitting down with friends of mine who work in industry in (what I would consider compared to myself) much more technical roles. I feel that sharing my ideas and thoughts with them and asking for feedback on the concepts will help me better plan what is feasible for me to do for my FMP.
Ultimately with my current health issues in the mix, I feel the need to cut back my current projects and push some of my FMP concerns to the backburner until after this term’s submission. While I’m disappointed that I won’t be able to achieve as much as I wanted to this semester, I would rather cut back and still have enough space to learn Marvelous Designer and create some clothing concepts than overload myself and only manage to do the bare minimum work and learn very little.
In regard to the clothing concepts, I have drawn out two more tops and I’m trying to get the shape down for the trousers. I’m finding it quite difficult to design the techwear pants, with the shape being very difficult for me to capture in a way I like.
I wasn’t in class this week unfortunately, but I was able to catch up with some of the content which for this week was around projection mapping. I recently went to see a projection mapping on the Battersea Power Station with some friends, so learning how this could be done was fascinating. As I’ve said previously for other techniques, this could be a really cool way to have a unique showreel piece using 3D animated characters or creatures, but I would need to put some thought into how to do so in a way that would work in a showreel. Maybe having a character interacting with the space they are projected onto could be an interesting exploration of the medium and demonstration of my skills in a different way to the standard showreel piece I usually see in 3D Animator reels.
Project Progression
I managed to make a small start on the concepts for Rahela’s clothing in photoshop, using the rig as a base for the drawings and the previously shown visual research as reference.
I unfortunately missed last weeks lesson/feedback session due to health issues. I’m still dealing with that, so the project hasn’t progressed since my last post unfortunately.
Class Content
This week a few of us did mocap. I was able to capture various pieces of data, which I’m hoping to use for 3D animation pieces over the summer. Since mocap data cleanup would be good to have in my showreel if I apply to any VFX houses after graduation I have wanted to capture some of my own to have for a while now but I haven’t been able to get round to it, so this was a great opportunity.
This week we learnt about NDisplays, which are used to render videos for irregularly shaped screens such as those seen wrapped around buildings. It was interesting to learn about the process within UE5 as beforehand I had only seen the process done in other programs and in different ways. I’m unlikely to use the technique for my current project, but I have considered if I could use the technique with some 3D animated characters as a cool and unique shot for my showreel, while also looking into the differences in how 3D animation is done for such projects.
Project Progress
I haven’t made a huge amount of progress on the practical side of things, but I have gathered a lot of visual research to reference when I start designing the clothing. Focusing on the main character, I have taken inspiration from both traditional Romanian (and other similar eastern European traditional wear) clothing alongside techwear/cyberpunk clothing. Reflecting on the various aspects from both clothing types, I think having the clothes shaped like techwear/cyberpunk clothing but with colours, patterns, textures, and detailing from traditional Romanian wear will work best, but I will still play around with different combinations when I begin the concepts.
I also took the rig I will be using for Rahela and took screenshots of a neutral pose at various angles that I can draw over for the concepts.
This week we were experimenting with Virtual Production Cameras within UE5. A few of us had some technical issues setting up the app in class, so I tried it on my home PC where having access to my home wifi seemed to help. We used the ‘Unreal VCam‘ app on our phones and connected this to a virtual production scene in UE5. While I found the process very cool and useful for VFX style production, I am unsure how much I will use the process in my own work. I would like to research the use of it more to see how I might be able to make use of it, but for my current project ideas the cameras are either hand animated or physically held and then tracked for compositing purposes.
Project Progress
I started doing some visual research for ideas, and based on project ideas Serra mentioned last week I also considered window sets seen in store fronts.
My idea for my FMP is based in a futuristic cyberpunk version of London, so incorporating punk themes to this project feels suitable. Above are some cool window sets I found and liked the vibe of, including some punky ones and one of a runway show with punk/alternative themes.
I also did some visual research for cool unique digital media styles that I could potentially experiment with, both for animation and otherwise.
When I showed some of these things to Serra today, she suggested I could potentially have different window sets in different visual styles, which would allow for me to experiment with multiple styles in one project. I do really like this idea, but I later realised this might take a lot of time depending on how long the Marvelous Designer side of things takes me.
I think the next steps for me will be figuring out the scope of the project overall, and designing some clothes to make in Marvelous Designer.
Today we were introduced to the unit work we will be doing with Serra. For her side of the unit, we will be doing experimental work where it will be important to try new things and take some risks with the techniques we experiment with. The work we do for this project can contribute in some way to our FMP if we like (such as experimenting with visual styles we might want to explore and expand upon further for our FMP).
Project Progress
Over the break between Term 2 and 3, I was looking into Marvelous Designer. While I want to focus on 3D Animation primarily for my FMP, I do want to still include my own aspects to the visuals outside of animation. My current plan is to use the M-Rigs by Ramon Arango to save time on making my own rigs, but to modify them to personalise the characters to my own designs. This could include altering the textures, the hair models, and the clothing. Hence my interest in Marvelous Designer, as I am hoping to make my own clothes for the M-Rigs using this program. From some research into the process over the break, I found that I could animate the rigs and then import the animation to Marvelous where I can simulate the clothing over the top.
Therefore, for the project I would like to learn Marvelous Designer and make some clothing for at least one of the rigs. I have never used the program before, so I am hopeful this will be a great opportunity to do so. I do also want the project to be more than me learning to program, so I had the idea to create a catwalk scene with interesting visuals to match the design of the clothing.
However I do still want to focus more on the animation side of things, and Serra mentioned that (like in Term 1) we can choose whether to spend more time on George’s project or this one, so I will most likely be spending more of my time on the animation work. I do still want to use this project to develop aspects of my FMP as much as possible though, and I am excited to explore Marvelous Designer and some cool visual techniques with this work.
