I don’t know where the time has gone. Already it’s September and the morning chill has me in a wool-lined hoodie. The sky is beginning to show hints of that deep Payne’s Grey that comes exclusively with early fall and I’m admittedly a bit giddy about it. I don’t know what my personal Pandemic winter is going to look like yet, but I’m really grateful for how I was able to spend the summer. There are worse ways to be spending your time than tinkering around in a farmhouse while trying to figure out how to create artwork in virtual reality.
Still, I’m not sure how this leap of time happened. I must have started this piece around a month ago, and granted there’s been plenty to distract between then and now, it’s a wakeup call that it simply has taken too long (and still has a ways to go). If I continue at this rate, I’d be lucky to have a handful of pieces by April, so I think it’s probably time to carve out a production plan, at least loosely speaking. Okay enough berating myself, it was August, and it did include writing my thesis proposal, presenting to the CFC committee, getting all my VR tech set up, and not to mention painting my apartment, turning a year older, and…. August. Fall is great because the FOMO of trying to focus on indoor work on a gorgeous day is at least a bit less of an issue.
Following the excellent advice of my Advisor (Emma Westecott), I also began the process of documenting the work in video format. This entailed the time-eating act of editing but I had forgotten how much fun it is. I had been recording small snippets of my production as I painted in Tilt Brush, but after compiling and splicing the footage with a voice-over guiding the process, it became clear that I need to be doing a lot more of that. Basically I need to be recording full painting sessions then making a practice of immediately exporting, reviewing editing and documenting the footage. Adding a voice-over to explain my process is essential too, even just as a note-taking tool. There’s a lot more I need to familiarize myself with about Premiere Pro, for instance, how to add tags as bookmarks in different parts of a clip for future reference. Where video editing is concerned, I’ve historically had a fairly counter-productive practice where instead of taking the time to learn, I consume endless bits of time trying to get things done the hard way. Okay here’s my list, while it’s still fresh in my mind:
– How to move entire blocks of Video and Audio ahead so I can open a gap for a new piece of footage without having to select everything manually. I know there’s gotta be a way to do this.
– How to rename clips for easy referencing
– How to add bookmarks to clips and/or to the playhead if possible
– How to make easy audio edits (I did try to research this but what was suggested wasn’t available to me)
– How to build sub-projects / sequences then nest them into a main sequence

Overall the process has been great. First and foremost, it’s been encouraging to finally have a substantial prototype. I can see the promise and value of it as an educational tool. But being able to talk through my process was a major throw-back to my days of teaching digital illustration, where I would talk the class through a live-demo for anywhere from an hour to the full three-hour time block. I feel like it comes naturally to me at this point; those little tangents that feed into the lesson like subplots, building a cultural context, observing and admitting the faults in my technique and suggestions for how to improve for next time, and the observances made that tend to surprise me later for their expertise, at least in terms of owning my craft. The latter best exemplifies the concepts of Flow: the heightened attention that comes with thinking on your feet (or in this case, speaking into a microphone without a script). It was a nice surprise to discover that I could address this audience of no one (as of yet) the same way I could to a group of students in a classroom. Even better to have the added benefit of editing out the ‘uuhmms’ and pregnant pauses as well as doing away with the rambling moments which I’ve always used as sort of verbal placeholders while my mind is occupied thinking of what to talk about next.
This first attempt is just shy of 15 minutes in length and far from perfect. Still it served as an excellent template to experiment with. I’ve got a few ideas already of how to approach the second attempt:
– As I mentioned above, incorporate video imports and voice recordings into the regular painting sessions.
– Create two versions of the video: an extended version where I’m able to walkthrough the entire process, then another (or several) that chop up the material covered into smaller chunks that are easier for the attention span to digest on the fly. I could see this serving really well at a later point in attracting viewers to my learning channel, ie., by tagging the shorter videos with labels that are technique specific so they come up in searches, then using them to navigate viewers to the longer videos where they can become more engaged with my own brand of teaching.
– Get efficient with title graphics, adding cut-to graphics for informative or entertainment value, and generally using the screen as a canvas for visual footnotes to compliment the voice over. It seems that Oculus uses a square aspect ratio for filming whereas the YouTube format is 16:9, meaning there’s ample blank space on either side of the footage to make use of (assuming this difference can’t be adjusted in the Oculus settings).
– I used my phone’s microphone for the voice recording. It was an easy process and I was pretty satisfied with the results, although next time I’ll try placing the phone in a mug and not move it throughout, while keeping myself at a consistent distance to avoid fluctuations in the recording volume.

Now, on to the actual project…

I’ll summarize what I already discuss in the video but with the added notes that came after the many edits to my voice-over. I didn’t really know where or how to get started with a project, so I just picked something and launched into it without much planning. That something was a Japanese print that happened to be in the room where I’m working, one among a series that my grandmother collected at some point, possibly while on a cruise with my grandpa in their still-active twilight era. I was drawn to the fact that there was so much atmosphere loaded into the single frame, along with enough constructible elements to make it a proper challenge, and that it was (and still is – this was an over-ambitious starting point to say the least).

I had sat through at least half of the tutorial videos about working in Tilt Brush from the series by YouTube instructor Shameless Mayhem, and figured it was sufficient to get me started (plus really just needed to get going at that stage of the summer). The main issue with that though, was that I hadn’t actually given myself the time to practice each of the techniques provided, and given the amount of tasks required to create this scene, that made this project overwhelming at every stage.
I spent about fifty years (three weeks) rendering out the grass alone. I discuss the two excuses I’d been telling myself in the video : 1) I wanted this grass to be physically accessible anywhere the viewer happened to find themselves in the scene, and 2) I made it an opportunity to train my motor skills to adapt to the 3D spatial environment. Both excuses hold up in court, but I’ll admit a third and fundamental reason was that I took refuge in the simple repetition of the endless brush strokes to shield myself from the daunting task of plotting out a more efficient way of rendering. That being said, I will argue that I was also trying to think like a painter more than a 3D environment designer, with all the time-saving tricks up their sleeve, like replicating assets. Being familiar with both ways of working (painting repetitively on canvas vs. planning visual patterns in a digital piece) I was weary of foregoing the creative flow of painting strokes for the sake of technically efficient application devoid of physical expression. In hindsight I figure there has to be a middle ground.

I equate it to being similar to creating custom brushes in Photoshop. As an example, one way of adding a sense of texture to skin when digitally painting a face is to create a brush comprised of a dozen odd dots, ranging in opacity from nearly translucent (near the edges of the brush) to opaque (around the core of the brush). A brush like this will allow you to not only cover more surface area than you would with a single dot in far less time, but it in effect looks much better.

Another example of this practicality is in painting hair, be it with traditional mediums or digitally. When presented with the challenge of rendering hair, people are always inclined to render hundreds of strands, convinced that the time invested will pay off in realistic looking volume. This tactic backfires 100% of the time. The reason being, a sense of volume can’t be conveyed with endless strokes of a brush any more than it can by rolling strands of clay out into a thousand tiny spaghetti strings and having them float together in midair. It needs to be built on a foundation that conveys a sense of form. This goes for any hair type, whether it’s a thick, full head of long brown hair or thin, balding curly grey hair. In the former you’d start with a dark brown, opaque base that takes the shape of the hairstyle in question, whereas in the latter case you’d start with a slightly translucent cloud of mid to dark grey. In both cases, from there you would then ebb out the volumetric shapes of the body of hair by alternating between highlights (high values) and shadows (low values). Finally you would apply a custom brush comprised of a few dots, similar to what I described above but maybe only four or five dots, and paint in the detail of the strands.

Even then you would only apply the strands in key areas – just enough to add a sense of texture. What you’re doing is creating the impression of hair, rather than attempting to literally reproduce the hair as you perceive it in physical reality. And that’s the key word: Impression.

The challenge with painting virtually though, is that you are in fact dealing with a form of physical reality. If you imagine an architectural model of a natural setting, ie. fields surrounding the track of a train set, depending on the quality of the model you might likely be able to see each individual blade of wheat if you observed closely. But that definitely doesn’t imply that each strand was created from scratch. More likely there were blocks, or patches of grass created first, then used repetitively to cover the surface area, perhaps with just enough alteration (achieved by using two or more blocks) to prevent the eye from detecting a pattern.

A similar idea could be applied for the grass. Taking the time to create a few assets that would connect together like the tongue-in-groove style sockets found on the pieces of toy train sets could potentially allow for higher quality material generation in a much shorter period of time. Eventually we should see features like converting assets into symbols that can be edited dynamically at any time (like Illustrator or since the onset of Adobe CC Creative Cloud, any library assets). But for the primitive state we find ourselves in with VR apps, we’ll have to settle with a UX that opts for simplicity over efficiency. It just means that the endless copy/paste repetition of assets (blocks of dirt and grass) will each be considered separate instances thus the design is baked in. Any forthcoming edits will done on an individual basis, with the only feature permitting quick selection being the ability to group each asset to itself, as a sub-collection or entire collection. On that note, another missing feature here is that of nesting groups within groups… ungrouping a collection in Tilt Brush means disband all the sub-groups within it as well.