It took me a long time to realise that the joy of creation is in the process – the dreaming, the experimentation and ultimately finding a way to bring your vision into the real world in some form.
When I make a physical object like a painting or build an instrument, I always feel a little empty once the creative process is done. I never have a sense of attachment to the creation itself. I feel like this is the case for a lot of Artists.
This could be why I’m fascinated by generative systems. I get to be both the creator and experiencer to some degree. The process never feels complete, the results are always unpredictable, unrepeatable.
Perhaps it’s about creating something that will still exist when I am gone, a fragment, a digital memory in a global network that continues to evolve without interference.
I had an idea for a new project. It is a living system - a space where sound and visuals are generated, layered, and allowed to evolve in slow, deliberate ways. Each listening session is unique, shaped by time, chance, and the subtle constraints of the system itself.
First we need to make some decisions about the platform this will exist on. I really enjoyed building RNDM and the fact it is so accessible, it runs using simple code in your browser. It also reminds me of the early 2000s when web designers were making super creative arty websites just for fun, no monetisation tactics or paywalls. It was an amazing community to be a part of as a young designer, sites like K10K shaped who I would become as a creative.
I’ve wanted to create a generative audio system for some time, something that could be considered a concept album that never repeats, but each “piece” still has it’s own theme.
If it’s browser based, we’ll need a visual component. Something that also has an evolving/morphing style - perhaps there are some shared values that effect how the sound and visuals evolve over time.
I remember experimenting with processing some years ago and creating an image that was effected by “pixel swapping” - whereby individual pixels within an image swap locations to deconstruct the original, reforming into something abstract. This could be an interesting option.
For this project I will be leaning into some “vibe coding” using ChatGPT 5.2 as this will dive into some areas my basic understanding of code won’t cover. I’m interested to see how this process goes, to be honest I am not a big fan of AI but I’m trying to detach from my cynical views and approach this tool with an open mind. This may very well turn into a 10 part series where I rant about how ChatGPT is dogs shit. We’ll see. I want to try and structure these sessions into more of an educational focus rather than “here is a finished product”. There will be no AI involved in the creation of the audio or visuals.
So there are essentially 3 aspects to this idea:
Audio - sound that constantly evolves over time
Visual - Photographs that evolve/change alongside the audio
Environment - Both the audio and visuals exist in the same ecosystem where shared values direct the outcome
Defining how the environment functions and the rules that influence the system feels like the first step.
It would be interesting to connect it to the real world in some way. Maybe the work responds to the listener’s local moment, altering its musical scales, textures, tonal character and visual states depending on the time of day or weather patterns. In this way, the piece is never detached from the world it is experienced in; it exists alongside the listener’s own passing day.
The media (I’ll use this term to reference both the sound and visual components together) is chosen, layered, and withdrawn according to simple rules: only so many elements may exist at once; recently used media recedes; long-unheard/unseen media slowly returns. Nothing repeats in quite the same way, yet nothing feels arbitrary.
The visuals could respond to the overall energy of the sound. As the music breathes, so does the image. When a new image appears, it begins whole again, only to slowly dissolve into abstraction.
Perhaps there is some kind of session parameter, each session gets assigned a certain “vibe” to make it unique. When you start a new session, you experience something quite different. Can we make these sessions recallable somehow? not an exact copy, but more like you plant the same seed, only to experience it grow into something different.
Some lofty ideas! How do we achieve this in a practical sense…
Defines the personality of each session. Makes the experience reproducible and shareable.
Audio layer selection probabilities
Master effect selection and initial parameter values
Image selection order
Visual effect randomness
Same seed - same starting point, but different evolution each time
Evolution occurs slowly over time
Seed displays on screen and can be copied/recalled (UI)
Dawn, day, dusk, night can influence:
Musical scale
Audio layer pools
Master effect parameter ranges
Image pools
Effect availability (certain effects for certain timeframes)
Density of activity (e.g. more minimal at night)
Perhaps there is a different tempo or musical scale for each timeframe. The album you hear at 2am is not the same album you hear at 2pm. Maybe you prefer dusk to dawn, in the same way that people might like track 1 more than track 2 on a traditional album.
Limited concurrent layers
Media fades in/out over time
Recent media is less likely
Long unheard/unseen media slowly re-emerges
Some layers never coexist
Some media is rare
The same parameters effect the audio and visual states in a different way
This feels like a good starting point! Next I will check the feasibility of these ideas and write a brief for the project. I find this is a good idea when there are a lot of technical requirements, I need to make sure I define exactly how it should work to avoid any scope creep down the track that could derail the whole thing.
I will also look at the broader creative concept and then hopefully start to bring it all together. What do you think? Come along for the ride and let’s see if we can build this thing… this part of the process is very tech heavy but I promise it will get creative soon!
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.