prismatical

A psychedelic[ish] seed based (for now) visual synthesizer with aNiMaTioOOn.

What it does

• Takes/Generates a string as a seed and generates an unique canva from it
• The seed controls the shape the frequenceis and the starting/transition color palette
• You can animate the visual in real time using keyboard toggles

Why?

It has been a long time since I don't touch C++ so this is a funny and tryppy way to get used to it again. Besides that, I have no clue. Maybe I do. Most likely it is my artistic side giving glimpses of development or it is just a nice way to waste my time. Only time will answer.

Build

Requires: cmake, make, SDL2

mkdir build && cd build
cmake ..
make

Run

./prismatical                # default seed
./prismatical Ineedholidays  # your own seed

Controls

Key	Action
space	Generate a random seed
1	Toggle phase drift — the shape breathes
2	Toggle color flow — hues rotate along the curve
3	Toggle freq morph — the shape warps
escape	Quit

You can also type any word directly in the terminal and press Enter to jump to that seed.

Available configs

All visual parameters live in src/config.h so you don't need to investigate this code. You can trust this repo I guarantee there is no crypto mining stuff or deep scan to steal your data.

Constant	What it does
COLOR_HUE_SPREAD	How wide the color swings are (degrees)
COLOR_HUE_CYCLES	How many color patches appear along the curve
COLOR_SATURATION	Color intensity (0 = grey, 1 = vivid)
ANIM_PHASE_SPEED	How fast the shape breathes
ANIM_COLOR_SPEED	How fast the colors rotate
ANIM_FREQ_SPEED	How fast the shape warps
CURVE_STEPS	Curve density — more steps = smoother but slower

Roadmap

Fractals

Fractal generators alongside the existing wave patterns. Same seed-based approach producing an unique fractal with its own identity.

Audio input & FFT

Feed a song or microphone input. The idea is that the audio will be broken into its frequency components in real time using FFT, and those frequencies directly drive the visuals in some kind of pattern. It could be e.g. bass shakes the shape, mids shift the colors, highs add detail and motion.

Real time audio reactive animation

Every pattern change via fft becomes a visual event. Shape, color, speed and pattern all respond live to what the audio is doing reacting to the full frequency spectrum continuously.

Style fingerprint

The idea here is to try to identify patterns and classify the audio based on its frequency distribution. The synthesizer could identifies recurring patterns energy, rhythm, harmonic structure and maps them to a similar visual style based on the group classification. Similar sounds get similar looking visuals. Different genres look clearly different. A funny name rotule would be cool as well.

ESP32 ?

Maybe some crazy future stuff will happen here.

Stack

• C++23 — language
• SDL2 — window and pixel rendering
• CMake — build system

Platform support

Platform	Status
Linux (native)	Working
Windows	Hell no
MacOs	It should work
WASM / Browser	Planned — the goal is to compile for WASM and embed directly in a webpage, but the web shell and browser rendering layer are not built yet

Development happens on Linux first. Once features are stable empirically, based on my judgment, the WASM target will be added so the synthesizer can run in a browser with no install required (I hope).