Against the Current: A Struggle-Fueled Character Odyssey
This project involved sculpting a character in ZBrush, tackling challenges in grooming, and navigating rendering issues. Rigging and animation were facilitated using Character Creator and iClone, and simulations were crafted in Houdini.
Presenting my midterm project, a product of six days' dedicated work at Pixl Visn Media Arts School. While the process presented numerous challenges, including three days spent troubleshooting and finding a way to render the project, certain planned aspects couldn't be completed due to time constraints. Additionally, the project exhibits some render glitches that, although bothersome, were unavoidable within the time frame, and I acknowledge that the overall render quality could have been improved with more time.
My midterm project encountered significant challenges primarily related to the groom, particularly in the process of exporting and rendering it. Despite the enjoyable experience of creating the groom in Houdini, a software I deeply appreciate, complications arose during the subsequent stages.
Previously, I had successfully exported grooms from Houdini as Alembic files for rendering in Maya using Redshift, a choice motivated by its efficiency for mid and end-term projects where time is of the essence. However, a few months ago, I discovered that Redshift couldn't read GPU caches in Maya, making the typical Alembic import excessively resource-intensive. To address this, I had developed a workaround involving Bifrost and resampling the strands/curves with two plugins, albeit not a perfect solution.
Confident in my ability to navigate this workaround, I began my midterm project, but upon attempting to import the groom through Bifrost into Maya, consistent crashes ensued. It became evident that the complexity of the groom exceeded the system's handling capacity. Although I tried pruning hair in Houdini to reduce the load, compromising the groom's quality, the troubles persisted.
Arnold, an efficient renderer capable of handling GPU caches seamlessly, proved a viable alternative. Despite successful imports and commendable performance, the decision to render in Arnold was not ideal due to its comparatively slower speed for mid and end-term deadlines. Thus began days of attempting to render the groom in Redshift, experimenting with V-Ray in Maya (although encountering issues with character displacement), and even exploring direct rendering in Houdini with V-Ray and Karma, all without success.
In hindsight, I realize that my insistence on finding a quicker rendering solution led to days of futile attempts, significantly delaying the project. Ultimately, on the last day of submission, I resorted to rendering in Arnold, simultaneously employing six school computers to expedite the process. The experience, marred by renders crashing every 30 minutes and incorrect layer outputs, proved to be a true challenge. Despite the unexpected obstacles, I am genuinely surprised and relieved to have successfully completed the project in the end.
Just before the midterm, I successfully completed a freelance project in collaboration with two friends and fellow students, resulting in some earned income. With this newly acquired funding, I made the decision to invest in Character Creator 4 and iClone 8, two software tools I had long desired, particularly for their appeal to character artists in terms of presentation.
This acquisition significantly enriched my workflow. I seamlessly transferred my ZBrush sculpt of the character onto a Character Creator basemesh, and subsequently imported the character into Character Creator 4. The software not only facilitated the integration of my sculpt but also provided a comprehensive full-body and facial rig for the character. Taking the process a step further, I exported this rigged character into iClone for the animation phase.
For the walk cycle, I utilized mocap data, while manually animating the chair animation and facial expressions—a task that admittedly showcased my novice animation skills. Nevertheless, given the limited time I had, I'm surprisingly satisfied with the results. This workflow allowed me to bring my custom ZBrush sculpt to life through rigging and animation, tasks that I'm not proficient in. I continue to handle my sculpting, grooming, texturing, and other aspects independently, but the integration of Character Creator and iClone enabled me to overcome the challenges of rigging and animating, making this experience both enjoyable and productive. It marked my first time using such a workflow, and it certainly added a new dimension of creativity to my projects.
Sculpting has always been a cherished phase for me, and naturally, I turned to ZBrush for this project. Devoting the first day to sculpting allowed me to swiftly transition to other crucial phases like texturing. Recognizing that intricate close-ups weren't in the project's scope, I ket it simple in the sculpt, focusing on capturing the essence and form efficiently.
For texturing, I opted for Substance Painter, a choice I make when dealing with a manageable number of UDIMs. However, for projects with a higher UDIM count, I turn to Mari. Despite the time constraints, I found the texturing process enjoyable, dedicating approximately 2 to 3 hours to this phase.
Below, I've included a selection of iRay renders for your review. Please note that the mesh lacks displacement or subsurface scattering (SSS) in these renders; their purpose is solely to showcase the textures.
All simulations for the project were crafted in Houdini, a process I find immensely enjoyable despite its time-consuming nature. One notable challenge lies in the size of the Alembic files generated during the simulation process, which are notably large. This bulkiness adds a layer of complexity to the workflow, making it somewhat challenging to manage. However, it seems to be an inherent characteristic of the simulation process, and I navigated through it to achieve the desired results.
The creation of the throne was a swift process, as it wasn't the primary focus and was predominantly obscured by the character. Approximately 2 hours were dedicated to the modeling phase, with an additional 30 minutes allocated for texturing. This efficient workflow allowed me to allocate more time to the character and animation aspects, ensuring a balanced and timely completion of the overall project.
In concluding this project, I reflect upon the challenges and triumphs that have shaped its development. From overcoming hurdles in grooming, rendering, and animation to exploring new workflows with Character Creator, iClone, and Houdini, this project has been a journey of exploration and problem-solving. Sadly it was mainly problem-solving.
The use of tools like Substance Painter for texturing added depth and detail to the character, despite the limited time dedicated to this phase. Simultaneously, the intricacies of working with Houdini for simulations brought both joy and challenges, highlighting the balance between enjoyment and the inherent time-consuming nature of such processes.
I appreciate you taking the time to go through the comprehensive overview of this complex project, filled with numerous obstacles.