How Much Can You Compress a PNG?

You've got a PNG that's too large. Maybe it's slowing your page load, blowing your storage budget, or getting rejected by a platform's file size limit. You open a compression tool, hit compress — and walk away with a file that's 8% smaller. That's... not what you were hoping for.

The frustrating truth is that "how much can a PNG be compressed" doesn't have one answer. It has five — depending on the image content, the strategy you apply, and whether you're willing to let any visual information change. Done wrong, compression yields almost nothing. Done right, the same PNG file can be 80% smaller, visually identical, and served in a format browsers prefer. This guide breaks down every scenario with real numbers.

1. What PNG compression actually means

PNG files store pixel data using a two-stage compression pipeline: a reversible filtering pass that converts pixel values into small differences, followed by DEFLATE — the same algorithm inside ZIP files. This pipeline is always mathematically lossless. The pixel values decoded from a PNG are bit-for-bit identical to the originals.

That lossless guarantee is what most people mean when they say "compress a PNG." But it's only half the picture. There are actually three distinct compression strategies available for PNG files, each with a very different outcome:

• Lossless recompression — strips invisible metadata, reduces unnecessary color depth, and recompresses the DEFLATE stream more aggressively. Zero pixel change. Typical savings: 10–30%.
• Lossy quantization — reduces the image's color palette from millions of colors to 256 or fewer using a process called palette quantization. Pixel values change, but the difference is usually imperceptible. Typical savings: 50–80%.
• Format conversion — exports the image as WebP, JPEG, or SVG instead of PNG. Not compression in the traditional sense, but produces the smallest files for most use cases. Typical savings over PNG: 25–90%.

Understanding which strategy you're applying — and which your tool of choice actually uses — is the prerequisite to understanding why your numbers look the way they do.

2. Why compression ratio matters for your site

Images are the single largest contributor to page weight on most websites. According to HTTP Archive data, images account for around 50% of the average page's total transfer size. PNG files, specifically, tend to be among the worst offenders because they are generated by design tools at maximum quality with no output optimization in mind.

The compounding effects reach further than raw load time:

• Core Web Vitals. Google's LCP (Largest Contentful Paint) metric is directly tied to how fast your hero and above-the-fold images load. A hero image that loads in 0.8 seconds instead of 3.2 seconds can move your LCP score from "Poor" to "Good" — a measurable ranking signal.
• Mobile experience. On a 4G connection averaging 20 Mbps, a 500 KB PNG takes about 200 ms to download — before the browser has parsed it, decoded it, and composited it. A 100 KB equivalent takes 40 ms. That 160 ms difference is perceptible and accumulates across every image on the page.
• CDN and bandwidth costs. If you're paying per gigabyte for CDN egress, images that are 70% smaller cost 70% less to serve at scale. For high-traffic sites, the savings are non-trivial.
• Storage and backup size. Unoptimized asset libraries compound over time. An image pipeline that processes on upload keeps your storage costs flat as your content library grows.

3. Step-by-step: how to get maximum savings

Maximum compression isn't a single button press — it's a sequence of decisions made in the right order. Here's the exact process to follow for any PNG that needs to be smaller.

4. Common mistakes that kill your compression ratio

These aren't edge cases — they're patterns seen across thousands of production sites. Each one represents recoverable savings sitting idle.

Using PNG for photographs

Photographs have millions of distinct colors and complex noise — exactly what PNG handles worst. A 500 KB JPEG and a 2.5 MB PNG can look identical on screen. PNG's lossless guarantee is meaningless for a photo whose "original" data is already a lossy sensor reconstruction. Use JPEG at quality 82–88 or WebP lossy. The PNG version buys you nothing.

Exporting RGBA when there's no transparency

Design tools default to RGBA because it's safe. But if an image has no transparent pixels, that alpha channel stores 255 (fully opaque) for every single pixel — 25% extra data carrying zero information. Always check whether your export actually uses transparency before choosing RGBA.

Skipping metadata stripping

Metadata stripping is the fastest, cheapest, always-safe optimization available. An iPhone photo embedded in a PNG can carry 10–30 KB of EXIF data — GPS coordinates, camera model, shutter speed. Design tool exports carry Adobe XMP blocks and color profiles. All of it is invisible to the browser renderer. Strip it unconditionally.

Serving large images in small slots

Compressing a 2000×2000 PNG and then displaying it in a 200×200 slot means the browser downloads 100× more pixel data than it renders. No compressor recovers those bytes after the fact. Resize to display dimensions — accounting for 2× retina — before any compression pass. Dimension-correct images at proper compression will always beat oversized images at aggressive compression.

Treating all images the same

A flat logo with 8 colors and a gradient-heavy illustration respond completely differently to the same compression settings. Flat graphics with few colors should use indexed mode with low or no dithering. Complex images need higher color counts and full dithering. Running a single preset across all images guarantees suboptimal results across the board.

Not checking output visually

Lossy quantization at aggressive settings can produce visible color banding in gradients, shifts in precise brand colors, or edge artifacts on anti-aliased text. A file size number is not a quality metric. Review every lossy-compressed image at 100% zoom before shipping, especially images with subtle gradients or exact color requirements.

5. Real examples: what to expect by image type

Compression ratio is almost entirely determined by image content. Here's what to expect for the most common PNG types found in production.

6. Compression method comparison table

Use this table to quickly identify the right strategy for your specific image and performance goal.

7. Compress your PNG now

The fastest way to see what compression ratio your specific PNG file can actually achieve is to run it. The Rebrixe PNG Compressor applies the full optimization pipeline — metadata stripping, color depth analysis, and both lossless and lossy modes — in your browser. No upload, no account, no waiting.

If you've determined that PNG isn't the right format — which, for most web images, it isn't — here are the format-specific tools for each alternative:

8. Frequently asked questions