The visualizer UI gets an aesthetic (and functional) overhaul, and the physical interface finally connects to the digital realm!
A code tutoring session saw us through the successful implementation of the buttons and encoders for the music box interface.
Knocked out the serial DAT parser in TouchDesigner on the quick when it became clear that there was no simple JSON implementation for CircuitPython (yet).
We built the controller logic from the ground up, which ended up being quite simple and painless in the end (due to Jesse actually knowing what he’s doing) and BOOM!!! There we were with out first true working prototype : an actual tangible object that controlled sound and graphics in TouchDesigner. It was simple and janky, but it was actually there. It felt like literally months of clawing around in the dark, grasping at nothingness, finally coalesced into something. I relished the brief moment.
And then back to the work! Due to some formatting changes to what the music box component expects, it means my ‘digital UI’ is totally scrap moving forward. If we want to tweak the instrument without the physical interface, its going to mean building a new GUI...
After a next steps conversation, I went home and wired up and coded the third input on the device, and put some attention and effort into getting the visualizer up and running with our new controller. This process led to a significant refactor of my data and rendering networks that I feel pretty good about, especially knowing that future inevitable changes will be easier to implement.
Dug in and stayed up late significantly reworking the visualization UI for the musicbox project, both for functional readability, and to bring it in alignment with some other aesthetic endeavors from a while back that predate this feed.
I imagine I’ll post some of that stuff eventually.
Definitely seriously frustrating not being able to get the SOP’s to do what I want them to do. Hours and hours of just grinding my brain into dead end after dead end. But ultimately I think this has netted out into something OK. At the very least, it solves our problem, which is most important at this early stage.
Got some new parts in and wired up to the physical interface. Trying out a Metro M0 Express by Adafruit for this stage of the prototyping. The knowledge is finally sinking in, as I was able to get this all connected without referring to an intro to electronics tutorial; a first!
Threw some rough and dirty code on the board just to make sure the inputs were wired correctly. Things seem to be in order. There’s not much here yet, and I dunno if it would at all be of use or interest to folks, but the repository is live on GitHub (which in itself is yet another thing I’m learning!).
A series of crops from the previous generative art tutorial. Working with replicators in TouchDesigner to layout crops of all the images created over a given period of time...
A few test renders from a procedural design tutorial from the excellent Matthew Ragan (as a step towards investigating SOPS). I take no credit for the aesthetics of the visuals posted, as it hardly deviates from the lesson...
While its just a tutorial, I am stoked to be posting because this is actually my first foray into generative art that has actually yielded an interesting result.
I did end taking things a bit further than the video, in that I made a sort of fade loop to keep the forms relatively self contained instead of just continuing to paint the screen forever. I thought it might be nice to save an image of each of these forms...
And so, I scripted the images to be automatically exported every time the sequence has run through its loop. Until tonight, I haven’t exported anything from TouchDesigner, as its primarily a live realtime environment. Pretty basic stuff, but I saw an opportunity to tackle something I’ve wanted to try for a while. Success!
One of the things I really like about TouchDesigner and this way of working, is that you get these great feedback loops where as you solve a problem, a whole new series of options present themselves to you.
Moving forward, I’ve already made a few tweaks so that some properties are randomized with each new iteration, including speed of the color change over time, and length of time that the form is moving around painting the screen. Next time I fire things up, I’ll let it run in the background for a bit and see what these tweaks get me.
A few screenshots to give an overview of some of the (maddening) work from today. For the ‘music box’ project, we’ll ultimately be connecting the physical controller through TouchDesigner to control various parameters of the generative music, and potentially some on screen graphics.
I started the day by trying to build out a simple visual feedback graphics using Surface Operators (SOP’s), which is TouchDesigner’s 3d object paradigm. Normally I would have just mocked something up in Photoshop and brought in the individual layers to be controlled with data, but I’m trying to be less avoidant of 3d in general these days. Unfortunately what I’d thought would be a simple thing was quickly reality checked, and I was effectively unable to do even the most rudimentary modification. And so it is when you’re just starting out.
After several frustrating (and very time consuming) dead ends, I moved on to building out the control data network, including a digital UI that approximates the controls available on the physical interface (3 knobs and a button). While it wasn’t as sexy, I had more success here, and well, this part is actually what will be driving the thing. To get the problem solved, I leaned into working with Data Operators (DAT’s) and Python scripting; things I usually avoid because it moves away from the thing I love about TouchDesigner : the node based flow. But to manage state and update information about multiple objects with similar parameters, its either impossible or exorbitantly complex to lay out. (That or I just dont know how to approach it effectively...)
Once I had it functioning, I realized I had more or less had ‘hard coded’ the digital controls to the processing, whereas ultimately I’d need a way to pass in data from the physical box itself. I was determined to ‘refactor’ the code. With just a little bit of cleaning up, I have everything separated into nice little discreet modules. It was a small (and much needed) victory after hours and hours of feeling like I was getting nowhere.
---
Moving forward, its clear that I’ll need to go into deep tutorial mode to approach the graphics using SOP’s, which could significantly delay getting this prototype up and running. Given what we’re trying to accomplish, I think for now its best to go back to the drawing board to come up with a simpler approach to visualizing the control feedback.
I’m working with a friend to design and build a physical interface for a generative music project he’s developing. After our previous design session, we determined an ultra bare bones “minimum viable product” that would let us start twiddling knobs and access most of the functionality of a more robust controller.
MakerCase was used to generate the box (thank goodness for people building and sharing tools like this!) Once I had the layers of MDF bolted together though, I thought it was actually more interesting than being hidden in an enclosure. Mental note for future aesthetic...
I’ll most likely need to drill out some hole for the USB/battery connection for the microcontroller, but since I dont know which I’ll be using yet, I’m waiting. Plus the last time I got fussy about making that port nice, it ended up needing hella revision.
Very RAW at the moment, but will certainly get more interesting. Next step, get the microcontroller wired up...
Today I revisited some old work, and put some effort into updating and organizing a digital portfolio...
I’ve been away so long that I was surprised and somewhat excited to find that Cargo Collective has a completely new revamped hosting and design service.
Web design is one of my least favorite things to do; but as a non coder, this new iteration on their platform made it significantly less painful than times past. A little bit more work to do before the site is ready to go live...
Also on my mind is branding and identity. As a professional artist, I’ve had the great privilege of being part of some incredible collectives, and have not had much need of putting myself as an individual out there...
I have had several iterations of personal blogs/sites in the past, but none of the working names has ever really felt right, and I’ve always been uneasy or uninterested in sharing my work under my given name.
Enclave Atelier is something that I’ve had in my back pocket for a while now, developed for another project, but it keeps popping back up. I’m in need of an identity refresh, so lets see how it feels.
Minor code touch ups for the ‘mystery box’ prototype and navigating Git workflow...
This is a project I’ve been working on for a while now, learning the basics in several new domains of knowledge and activity : circuit python (code), micro controllers (hardware), and laser cutting (fabrication).
romaincrisis
prguitarman
kianamaiart
germansingergirl
baratheonpilled
izzy806