y2k-zones

An interactive macOS terminal command that applies Y2K glam effects to a video through a zone-based, click-and-paint interface.

Features

• Blur zone - freehand-painted blur with an optional pink/lilac/mint/gold color tint (not a plain gaussian blur), plus an edge softness slider that controls how feathered/soft the transition at the mask border is (0 = hard edge, 100 = very soft/gradual).
• Glitter zone - animated, twinkling sparkles (white / pink / gold / rainbow) with adjustable density and glow.
• Glass zone - frosted, matte Y2K-style glass (soft blur + desaturation + a light rim highlight along the zone edge), with an opacity slider so the original image shows through.
• Global dust - film grain + slight desaturation + vignette for a dusty, vintage feel.
• Global color filter - a full-video color wash (pink, lilac, emerald, baby blue, warm gold, or a custom color) with adjustable intensity.
• Original audio is preserved in the output.

Download

Option A - with git:

cd ~/Desktop
git clone https://github.com/alishapckg/Glitter-WIP.git
cd Glitter-WIP

Option B - without git:

1. On the repository's GitHub page, click the green Code button → Download ZIP.
2. Unzip it (double-click the downloaded file).
3. In Terminal, cd into the unzipped folder:

   cd ~/Downloads/Glitter-WIP-main

Then run the installer:

chmod +x install.sh
./install.sh

This installs the Python dependencies (opencv-contrib-python, numpy), ffmpeg, and registers the y2k-zones command on your system. Verify it worked with:

which y2k-zones

To get updates later, pull the latest changes (or download the ZIP again) and re-run ./install.sh.

Requirements

• macOS (Apple Silicon or Intel)
• Python 3.9+
• opencv-contrib-python, numpy
• ffmpeg on your PATH (used only to copy the audio track)

Install the Python dependencies:

pip3 install opencv-contrib-python numpy

Install ffmpeg (if you don't already have it):

brew install ffmpeg

Usage

y2k-zones input.mov
# or with a custom output name
y2k-zones input.mov output.mp4

An OpenCV window opens showing the first frame of your video plus a control panel on the right. Walk through 5 steps in order:

1. Blur zone - paint the area(s) to blur, pick strength, edge softness, a color tint, and optionally turn on "Track motion".
2. Glitter zone - paint where sparkles should appear, pick density, glow, color, and optionally "Track motion".
3. Glass zone - paint the frosted-glass area, pick strength, opacity, and optionally "Track motion".
4. Global dust - no painting needed, applies to the whole frame.
5. Global color filter - pick a color wash and intensity for the whole video.

Controls:

• Left mouse drag - paint the current zone (freehand brush, not a rectangle).
• Mouse wheel - change brush size (while hovering over the video).
• Sliders / toggles / color swatches - click and drag directly with the mouse, right in the window.
• Enter - confirm the current step and move to the next one.
• Esc - cancel and quit without processing.

Any step can be left at its defaults or turned off with its toggle - you are not required to paint every zone.

After step 5, processing starts automatically. Progress is printed to the terminal, and the final video (with audio) is written next to your input file as <input>_y2k.mp4 unless you specified a different output path.

Notes

• Masks are static by default: whatever you paint on the first frame is reused for every frame in the video. Turn on "Track motion" on a given zone if you want it to follow a moving subject instead (see the feature list above for how it works and its limitations).
• If ffmpeg isn't found, the tool still produces a video, but without audio, and prints a warning.
• Pure OpenCV/NumPy pipeline - no PyTorch, no ML models, nothing GPU-bound. "Track motion" uses OpenCV's classical CSRT tracker (appearance-based, not object recognition).