lab
Your browser cannot display the WebGPU Time Crystal canvas.

Move through the image to bend stored light

Time Crystal Studio

A moving image that remembers, folds and reshapes its own recent history.

Persistent field14 temporal samplesPointer + audio reactive
RuntimevGPU 0.3.1
Statechecking
Frame rate
Memory97.8%

Waiting for browser capability check.

This image remembers what just happened.

Most animation replaces one picture with the next. Time Crystal Studio lets each picture leave a fading trace, so movement gathers into ribbons, chambers and glowing folds.

You do not direct every shape. You choose how long the image remembers, how strongly it bends and how it responds to your pointer or music. The final form grows from those conditions.

A simple way to read the experience

  1. 01Light appears

    A moving source draws fresh energy into the frame.

  2. 02The trace stays

    Part of the previous image remains instead of disappearing.

  3. 03Memory bends

    Older traces drift, fold and separate into layers.

  4. 04You disturb it

    Pointer movement or a loaded track pushes the same stored material.

  5. 05A form emerges

    Many small changes accumulate into an image that feels sculpted from time.

A short visual memory creates the illusion of matter.

The browser keeps two copies of the image. It reads one, writes the changed result into the other, then swaps them. This fast exchange is what allows the picture to remember itself.

Images in memory
02
alternating continuously
Recent glimpses combined
14
used to shape depth
Authored looks
03
distinct visual moods
Viewing angles
03
rake, fold and tunnel

Each update keeps some of the old light, moves it slightly and adds new energy. High memory creates dense archives; lower memory clears the image more quickly. Fold, drift and fracture determine whether that history feels fluid, mineral or sharply broken.

The final drawing step looks at fourteen nearby moments in the stored image. Combining those glimpses makes a flat screen feel deep without building a conventional three-dimensional model.

For curious readers: the two alternating images are GPU textures, and the repeated read-and-write process is called a feedback loop. The three-dimensional feeling is an artistic reconstruction, not a scan of physical space or time.

How one new image is made

The process repeats many times each second. Small changes become visible only after they accumulate, which is why slow controls can have a large effect over time.

  1. 01Remember

    Start with the light left by the previous image.

  2. 02Move

    Drift and fold that stored light through the field.

  3. 03Break

    Use fracture to split smooth history into harder planes.

  4. 04Add energy

    Introduce new light from the source, pointer or audio.

  5. 05Rebuild depth

    Combine recent traces into the picture you see.

Why this matters beyond the experiment

Time Crystal suggests a motion-graphics system that can stay visually recognizable without repeating one fixed loop. A campaign could keep the same color and memory rules while every performance responds differently to a song, gesture or stream of values.

That makes it promising for music visuals, social clips, title sequences and live backgrounds. The concept is especially useful when the goal is to make time itself feel like the subject, rather than simply moving objects across a screen.

What this study does not prove

  • The field is an artistic metaphor. It does not simulate a physical time crystal.
  • Audio files stay in the browser, but live microphone input is not part of this version.
  • Smooth performance has been observed locally; it is not yet proven across a broad range of devices.
  • Social-format video export, color consistency and seamless loop quality still require dedicated testing.

What the controls change

Think of the controls as properties of a material, not as filters placed on top of a finished animation.

  1. Speed and drift

    Set how quickly the stored light travels and how restless the overall field feels.

  2. Memory

    Controls how long old movement remains visible. More memory creates denser, heavier forms.

  3. Fold and depth

    Turn flat traces into soft ribbons, stacked planes or tunnel-like space.

  4. Fracture and color delay

    Break smooth movement into seams and separate colors along those breaks.

  5. Pointer and audio

    Let a gesture or a local music file add energy to the same visual memory.

  6. Rake, Fold and Tunnel

    Change how the material is viewed without erasing what it has already remembered.

Where the research could go next

These are proposed studies, not finished features.

  1. Social motion recipes

    Create repeatable portrait, square and landscape loops with an inspected beginning and end.

  2. Live performance

    Map rhythm and musical texture to memory, fracture and camera movement in real time.

  3. Image seeding

    Let approved photographs or video enter the memory field and dissolve into the same material.

  4. Data as pressure

    Use changing values to alter retention and energy without turning the image into a conventional chart.