Wisteria Bonsai

The Idea

I created this bonsai tree because I wanted to further and challenge my skills as a 3D artist, my ability to sculpt and model organic shapes, which I have not attempted before, as beforehand, I had only created hard surface models, which had a lot of references and simple shapes.

So how I developed this idea was with the goal of creating a complicated and organic model, and the most organic thing I could think of was a tree; at first I had a more environment focused idea but after talking to my mentor I created several drafts and was able to finalise my idea, being a wisteria bonsai tree, with a focused set of references.

Initial Block Out

Based off the references I tried to create the block out of the shape in Maya but I found that it is much harder to visualise in Maya quickly, so I moved the references into VR and used Gravity Sketch, which provided a more holistic view of the model, as I could interact with it in VR. I found that the process of drafting ideas in VR speeds up the work flow in terms of early visualisation.

As I was fleshing out the Tree block out I figured that it would be best to removed the rock that the Tree was residing on as it hindered the readability of the overall silhouette and made the scale of the bonsai more harder to figure out. Here you can see the development of the main branches, as well as more contours on the surface of the tree itself.

I then moved the block out into ZBrush to further add branches using ZSpheres and to sculpt in some detail on the surface of the tree so that the high poly will have more interesting curves and geometry. 

After getting the block out to a decent stage I started Zremeshing the 3D mesh as it would take a very long time to Retopolgise myself. However while Zremeshing I also fixed some geometry (for example inflating the tips of branches), and added in more intricate detail and ended up creating several versions of this low poly block out. However I was finally able to get to a stage that I was content with and used that as the low poly which I started UV unwrapping.

Since it was my first time I was UV unwrapping such a complex and organic shape I had a lot of difficulty in the beginning getting the cuts as seamless as possible. This process took several weeks as I was trying to get the most optimal UV islands while maintaining a high texel density. With the help of my mentor and some other people I was able to figure out the method of unwrapping the tree in terms of thinking of the branches in cylindrical shapes as well as keeping some UVs attached to converse Udim space.

Adding surface detail

After the long time that was spent specifically on modelling and sculpting the bonsai tree, a lot of the initial drive had diminished due to the large amount of effort it took to just do the tree, and wanting the project to be finished led to unmotivated attempts of the rest of the project.

I tried to spice things up with a little bit of texturing to see the overall model and I was able to boost my productivity a little, but the rest of the model really suffered in quality compared to the tree and that took a lot of the perceived skill away from the tree.

It took a while to solve how to create the moss in much higher quality as well as adding surface detail to it. Which led to a lot of research into foliage cards, MASH and other forms of showing details of leaves and thin organic material. Starting with a paint over to visualise the new moss I used lots of reference and created a nice silhouette and overall flow of the mesh.

After getting an idea of the overall shape I wanted the new moss to be, I started with varying the soil and adding some surface detail so that the moss that would lay on top of it would be more interactive with the soil. 

Using Zbrush I blocked out all of the moss balls and zremeshed it several times to get a lower poly mesh to retopologise, this process was very tedious and frankly tiring. The reason why I couldn’t zremesh it was because it couldn’t maintain the shapes of the sculpt and reduce the polycount.

Hand retopologising the moss took around 2 weeks and it really challenged my knowledge of topology and helped me develop a stronger understanding of it. 

After the Retopo I UV unwrapped it, however with varying results, due to some unfolding issues as I used the legacy version rather than the new one.

The results of the moss sculpt I did not blend well into the soil and tree enough so I went back into ZBrush and had to sculpt more surface variation as well as deformed geometry to make it more organic looking.

I also developed the post and added some simple smart materials to visualise the piece.

Here is where I started finialising the UVs and topology

Here was a quick texture pass

Lots of trial and error on different ways of getting moss

I’ve tried different methods in getting the moss fibers using:

Fibermesh, Nhair, MASH, Xgen

Through a lot of trial and error I experimented with all these different workflows and gained a decent understanding of them, however I still couldn’t get the results I wanted for the moss. MASH was able to give me the results I wanted but my inexperience in it was the main issue, as well as the fact that it may be unsuited for this particular project. In the end I settled with Xgen as that had the most resources and gave the best results for my skill level.

Below are some results from FiberMesh

For the foliage I used Wisteria Leaves Atlas’ provided by Quxiel Megascans Library. I Multi-Cut the leaves I wanted from the texture maps (creating cards), then deformed them using bend deformers in Maya and soft selection on specific vertices.

Here is the stage I also mocked up the positions of the Wisteria Flowers I would be making for the tree, as shown with the ovals.

Here is where I Started sculpting the petals for the Wisteria Flowers. I started off with a simple poly model in Maya and brought it into Zbrush to work on it more.

This process took a week or so as I had to learn how to PolyPaint in Zbrush as well as how to manipulate masks in to provide details in the Albedo.

I had difficulty with the placement of the petals, due to lack of detailed reference and my inexperience. I think I could’ve spent more time on this to get it closer to the references I had but I just felt like getting it done to a decent standard. 

After consultation with my mentor I felt that I should work more on the details within the texture itself as there was issues with resolution and very harsh colours from the pot itself. Though Wisteria trees are greyer and lighter in general, I chose to make it more browner in its colour, to bring more contrast to the petals (and to make it more easier on the eyes). 

I created a simple studio lighting setup here, with 3 Point Lighting. From my mentor, I was able to understand more in depth the use cases for the individual lights within the 3 point lighting system.

I learnt more about Maya’s Light Editor as well and will be using it more for future projects.

Concluding Thoughts

I had a lot of difficulty in managing my time, as it was hard to estimate the time it would take on things that I have never done before. It took a very long time to get some parts of the project to a certain standard before moving on to the next step. Planning ahead, this is something I really need to work on, as I wouldn’t have gone through burnout if I paced myself, and set milestones where I was able to rest during the project.

I have a habit of avoiding challenging parts of a workflow and trying to find another way to solve the issue faster; this can be seen with the moss section of my project, there was a lot of trial and error, and fumbling around trying to avoid certain pipelines.

I thought this project would’ve been completed way faster, but the lack of knowledge I had made the project a lot longer than it would’ve. The main issues I had were cohesive modelling, sculpting duration, texturing detail, foliage.

Overall, this was a challenging project and I came to realise my limits, and what I enjoyed as well as disliked throughout the process. 

Thank you for youre time and for reading this project :3