Process: The Workflow

When delving into why a work is made, sometimes it's easy to lose track of how it was made. In the whole process of making this work, I wouldn't have been able to create all the elements needed without good systems in place for execution and organization, from keeping track of work and deadlines, to an idea of what needed done moving forward. This section is dedicated to the ‘how’ structures. Think of this as a kind of systems summary. This is a place for the scaffolding of the work; if you don't want to see how the sausage is made, or if that's boring, it's also convenient to skip, but for those die-hard process folks, have at it!

The Master Document

Midway through my second semester of work, I was at an impasse. Getting into exactly what it was I was making when I first started was a grind, compounded by all the new challenges of working remotely, so it made sense to have a slow start. Half way through however, it felt like I had the roadmap ahead. I knew generally what I needed to do to create what I wanted to make, but it wasn't concrete anywhere. I finally decided to make a master spreadsheet that outlined all of the steps ahead, marked with definitive milestones in an easy-to-understand interface. What I ended up with was a beautiful three sheet document, with conditional formatting to see ‘complete,’ ‘in process,’ and ‘not yet started’ work at a quick glance; a clear picture of the next steps, broken down to make them more manageable; and a good sense of what I'd already done and where to go next. Having this document changed everything. It was so much easier to break off the small chunks, know what I needed to do, and knock it off the list. To a point, it was so easy to do work on the list, I had to add writing about process work to a new page just so I would stop working for a second to document the process!



The master document also helped me conceptualize how to systematize the work of making animations for Public from Nothing, the first stage of my thesis. Having made a style of shifting images I liked for one week of songs, I reverse engineered their directionality into a scheme for every week so I could keep the same animation timing, keeping it bounded within ten seconds, but have an interesting composition that didn't seem like a copy/paste.

Folders, Systems, Organization

Another important part of the process was making sure I had a clear system for file organization and systems applied to workflows. I know it sounds boring, but having a clean and easy-to-understand file tree is so important when working: it means knowing where to put things next, never losing track of where your most recent file is, etc. So much of the work in this undertaking is ultimately about collection and organization. A photo of hair every time I shower, 30 album covers every week, a series of 3D terrains, a set of animations.

 

Public from Nothing

Public from Nothing is almost entirely an organization-as-aesthetic output project, so much so that I tried to explore how I could automate it with scripts in After Effects. I was able to develop, easily enough, a system through which I could make each new video composition in about 30 minutes. Each week, I cloned my Discover Weekly playlist (since the playlist is otherwise overwritten each week), downloaded a .csv file of all the song info from a third party site, and used a tool created by app developer Ben Dodson to search for each album cover in the iTunes database and find high resolution photos.

In After Effects, I also created a system for animating and easily replacing. After finding a shift animation I liked with an ideal velocity, I translated it into four different directions moving both outward and inward. I devised that if each cell contained an album that moved out, a static album beneath it, and another album that moved in, I could show all 30 albums in 10 squares that could then be UV mapped onto columns in Blender. Mapping the movement to a null object and assigning an album image to it, I could repeat the animation endlessly without recreating it. After an invaluable meeting with my Motion Design professor Michelle Astrug, I learned to replace elements easily with Alt+drag, and I was set. Using the process to structure a change in animations over time from my spreadsheet, I created folders for each week, with animation bones in individual folders. When I got new album covers I just dragged them in and the whole thing was done. From After Effects, I brought them into Blender, UV mapped the first week's video file onto a column, and copy/pasted for each new week to maintain the UV map--and quickly the whole lot came together.

A Vision

When it came to photographically documenting my hair for A Vision, I knew I could easily create a system to process photos. Because the lighting in my bathroom was the same all the time, I knew that as long as my camera settings, especially ISO and aperture, were the same week to week, processing the photos became easy. I created a preset in Adobe Camera Raw with the lighting and context fix to make the hair more clean-edged. This I saved to favorites, so that each week I just had to find the best picture, right click in Bridge, apply the correction, and export full size and web versions. I've done enough quick-turnaround photography in the past (having done vehicle photography for hundreds of cars a month) that I knew it was a process I could automate.

New Terrains

New Terrain was probably the hardest work to systematize since each activity that happens within the virtual space is unique. I was able to develop a process for creation of each space and eventually a formalization for size, angle of the initial plane, and positions of cameras to capture tile views and revolving non-VR version.

The biggest system breakthrough was when I discovered a Python extension to render 360 images in Eevee render engine instead of Cycles. I knew I wanted to do VR environments before I knew exactly how Blender did its render work, since this is really only my first undertaking in Blender. I didn't realize that out of the box you need to use Cycles, an engine that takes much more time to render than Eevee, the default engine. Getting that bit of code, I was able to go from rendering one frame in 17 minutes to rendering two frames per minute, which was a game changer in how much I could produce. VR environments for Public from Nothing could be done in two and a half hours instead of close to two days, and VR spaces for New Terrains could be done almost instantaneously.

In the end, the ability to set good systems for a series of works primarily concerned with collection meant that I could take a little bit of each part and grow it to whatever amount I needed to feel like the work was as full as it should be. I was able to move much faster towards the end with my master spreadsheet, and creating guidelines for development of each piece of a project meant I could power through all the elements much more easily.