What's the actual difference between the two algorithms?

Uniform-cube maps every pixel onto this category's fixed 216-color grid (the same palette the shared encoder always quantizes to anyway), so it's really a preview of what re-encoding will do. Median-cut instead builds a custom palette by recursively splitting the image's actual colors at the median of their widest channel, which usually preserves the source's real dominant colors more faithfully.

How is this different from GIF Color Reducer?

GIF Color Reducer always uses one fixed (popularity-based) algorithm and only exposes a target color count. This tool instead makes the algorithm choice itself the point, letting you directly compare uniform-cube quantization against median-cut on the same source.

Does the color count setting apply to both algorithms?

No, only median-cut mode uses a target color count (2-64). Uniform-cube mode always uses the shared encoder's fixed 216-color palette.

Which algorithm should I pick if I'm not sure?

Median-cut generally looks closer to the original image for a given color count, since it adapts to actual content; uniform-cube is faster and shows exactly what a plain re-encode will produce.

GIF Quantizer - Uniform Cube vs Median-Cut Algorithm Online

Overview

Introduction

This tool lets you directly compare two different color-quantization algorithms on an animated GIF, entirely in your browser.

It's built for anyone who wants to understand or control the quantization algorithm itself, rather than just a target color count.

What Is GIF Quantizer?

A client-side quantization tool offering two distinct algorithms: 'uniform-cube', which reuses this category's own shared 216-color web-safe palette (the exact same quantization the encoder performs on every GIF this site produces), and 'median-cut', a custom recursive algorithm that adaptively builds a palette from the image's actual colors.

Median-cut repeatedly finds the color bucket with the widest range along any single channel and splits it at its median, until the target number of buckets exists; each bucket's population-weighted average becomes one palette color.

How GIF Quantizer Works

In uniform-cube mode, every pixel across every frame is mapped directly onto the shared codec's fixed palette using the same function `encodeGif` itself uses, so this mode is effectively a live preview of default re-encoding.

In median-cut mode, every frame's pixels are first bucketed into a coarse histogram, then that histogram is recursively split at medians until the target palette size is reached, and every pixel is remapped to its nearest resulting color; both modes then re-encode a new animated GIF.

When To Use GIF Quantizer

Use uniform-cube mode to preview exactly what a plain re-encode through this site's shared GIF codec will look like.

Use median-cut mode when you want a custom target color count that preserves the source's actual dominant colors more faithfully than a fixed grid.

Often used alongside GIF Color Reducer, GIF Ditherer and GIF Quality Changer.

Features

Advantages

Exposes the underlying quantization algorithm directly, rather than hiding it behind a single fixed choice.
Median-cut mode adapts to the image's real content instead of assuming a fixed color grid.
Uniform-cube mode doubles as an honest preview of this site's default GIF re-encoding behavior.

Limitations

Median-cut is more computationally involved than the fixed-table lookup uniform-cube mode uses, so it takes a little longer on large or long GIFs.
Neither mode performs dithering; combine with GIF Ditherer afterward if the reduced palette introduces visible banding.

Examples

Comparing algorithms on a colorful animation

Input

A colorful 15-frame GIF, median-cut with 24 colors

Output

A 24-color re-encode whose palette closely tracks the animation's actual dominant colors

Median-cut adaptively splits the color space based on the image's real content rather than a fixed grid, so it captures the animation's true dominant hues.

Best Practices & Notes

Best Practices

Try uniform-cube first to see the shared encoder's default behavior, then switch to median-cut if you want a custom, more faithful color count.
For very colorful or gradient-heavy content, median-cut with a moderate color count (24-48) usually looks noticeably better than a fixed grid.
If you just want a simple target count without comparing algorithms, GIF Color Reducer is the more direct tool.

Developer Notes

Median-cut is implemented as a from-scratch recursive bucket-splitting algorithm operating on a coarse pre-bucketed histogram (not raw per-pixel data) for performance, built once from every frame's pixels combined so the whole animation shares one consistent palette; uniform-cube mode instead calls the shared codec's own `buildGifGlobalColorTable`/`colorToGifIndex` directly, guaranteeing it matches `encodeGif`'s default behavior exactly.

GIF Quantizer Use Cases

Comparing a fixed uniform-color-grid quantization against an adaptive median-cut palette on the same source GIF
Previewing exactly what this site's shared GIF encoder will do to a file's colors before downloading
Building a custom, content-aware limited palette for a stylized animation

Common Mistakes

Setting a color count expecting it to affect uniform-cube mode; that mode always uses the fixed 216-color palette regardless.
Assuming median-cut and popularity-based quantization (used by GIF Color Reducer) always produce the same palette for the same target count; they use different selection strategies and can differ visibly.

Tips

Use GIF Color Index Extractor afterward to see exactly how many distinct slots your quantized result ends up using.
For a genuinely limited retro palette (8-16 colors), median-cut usually preserves recognizability better than the uniform cube.

GIF Quantizer