movie-barcodes

Python program to generate a movie "barcode", or a visualization of the colour palette used in a film. It is generated by compressing each frame into a single column of pixels that are stacked sequentially to form a barcode-like image. Makes for a neat wallpaper.

There are many implementations available, and I don't claim this is the best, but it's clean, simple, efficient, and easily-customizable. Automatically adjusts sample rate according to the length of the video. Runs in under 2 minutes for a typical 2 hour movie.

Supported on Python 3 and OpenCV 3+. Tested on macOS.

Requirements

• OpenCV
• numpy

Example Results

Dunkirk

Blade Runner 2049

Ratatouille

Tangled

Frozen

The Lord of the Rings: The Fellowship of the Ring

Uniform column mode

WALL-E

The Handmaiden

Usage

python barcode.py -video <PATH_TO_VIDEO> [-u]

• -video: The path to the movie to be processed. Only tested on .mp4 files, but should work with any format supported by OpenCV.
• -u: Optional flag to compress all frames into a single color, thus producing uniform columns.

Additional Notes

There are some constants defined in the code that can be changed if desired.

• Set OUT_WIDTH and OUT_HEIGHT to the desired output dimensions.
• Set SAMPLE_HEIGHT to change the granularity of the columns. SAMPLE_HEIGHT should be at most the input video height and at least 1 (which is equivalent to using the -u flag). This is a matter of personal preference; larger values preserve more detail while smaller values yield smoother results.
• Tip: use low-resolution videos. It won't effect the result at all and it will run much faster.