1. Home
  2. Blog
  3. Resize or compress first? Measuring what actually shrinks an image
Guide

Resize or compress first? Measuring what actually shrinks an image

When an image is too big, there are two levers: change the pixel dimensions, or lower the quality. People reach for the quality slider first. The numbers say that is usually the weaker lever.

How we tested

Five photographs were saved as high-quality JPEGs (quality 95) to stand in for camera or download originals, with sizes from 0.14 to 0.87 megapixels. Each went through ConvertPicto's resize tool at 75%, 50% and 25%, and through the JPG compressor at 85%, 70% and 60%. We also ran each resized file through the compressor at 70% to see what doing both gives. Everything ran in headless Chromium on the real pages.

The results

Every figure is the average size as a percentage of the original JPEG.

What we didAverage size vs original
Compress to 85%, full size53%
Compress to 70%, full size35%
Compress to 60%, full size28%
Resize to 75% of width30%
Resize to 50% of width16%
Resize to 25% of width6%
Resize to 75%, then compress to 70%23%
Resize to 50%, then compress to 70%12%
Resize to 25%, then compress to 70%4%

The resize tool saved its output at its default 85% quality, so the fair comparison for the resize rows is the first row. At the same quality setting, shrinking the width to 75% gave 30% of the original size, against 53% from compression alone.

Why resizing wins

Image data grows with the number of pixels, and pixels grow with the square of the width. Halving the width leaves a quarter of the pixels. That is why 50% width landed near 16% of the file size, while dropping the quality slider from 85% to 60% only reached 28%.

One photo, in kilobytes, to make it concrete. This is the largest of our five, 1000 x 872 pixels:

VersionSize
Original JPEG (quality 95)395 KB
Compressed to 85%187 KB
Compressed to 60%89 KB
Resized to 50% (500 x 436)52 KB
Resized to 50%, then compressed to 70%38 KB

The last row is about 10 times smaller than the original. The 60%-quality version at full size, which also looks softer, is still 2.3 times larger than that.

What this looks like in practice

The pattern repeated across all five photos, not just the largest one. Whatever the starting size, resizing to 75% of the width always beat compressing to 85%, and resizing to half the width always beat compressing all the way down to 60%. The bigger the starting image, the bigger the absolute saving, which is why this matters most for phone and camera photos, often 3000 pixels wide or more, that end up displayed in a column a few hundred pixels across. Our test images were small (0.14 to 0.87 megapixels), so a modern 12-megapixel phone photo will show an even larger gap.

There is also a quality argument for resizing. Lowering the quality slider adds compression artefacts such as blocky skies and ringing around edges. Resizing keeps every remaining pixel at the same quality, so the picture is smaller but not dirtier. Combining the two, a resize to half and a compression to 70%, is a good default for blog and web images.

A short checklist before you export

  1. Decide where the image will appear and how wide it will be shown.
  2. Resize to that width, or twice that width if the audience is on high-density screens.
  3. Compress the resized copy to about 70% to 85% quality and look at it at 100% zoom.
  4. Keep the untouched original in a separate folder.

The catch

Resizing throws away detail permanently. It is only the right lever when you know how large the image will actually be shown. A few rules that work:

  1. Find the display size. A blog column might show an image 800 pixels wide. A phone screen is rarely wider than 1200 physical pixels. Resize to about that, or at most twice the CSS width for sharp high-DPI screens.
  2. Resize first, compress second. Compressing before resizing wastes effort, because the extra detail is discarded anyway.
  3. Do not resize images you will print or crop later. Keep the full-size original and make smaller copies.
  4. Do not enlarge. Upscaling cannot invent detail and increases the file.

If you only need a smaller file and the dimensions are already right, compression alone is the correct tool. Our guide on reducing image file size without losing quality covers that order of operations. For the quality side of the story, see our JPG vs WebP quality measurements.

Resize an imageScale to exact pixels or percent — free, no upload
Compress a JPGReduce JPG file size — free, no upload
Caveat: five small photographs is a small test. The direction of the result, that resizing beats quality reduction, follows from how pixels scale, but the exact percentages will differ on your images.

Tools mentioned in this article

More from the blog

We checked all 28 image conversions pixel by pixel
28 conversions, each run in headless Chromium and compared against an independent decoder. What is identical, what is lossy, and why.
4 min read
How big will your image PDF be? We measured it
Page size does not change the file size, but quality does, a lot. Measured PDF sizes from JPG and PNG sources.
4 min read
JPG vs WebP quality settings: what each level really costs in file size
Nine quality settings, two formats, five photos, measured. Where the file-size curve is gentle, and where it falls off a cliff.
4 min read
PNG vs JPG vs WebP: what we measured on photos, screenshots, text and charts
There is no best format, only a best format for each kind of image. Measured sizes and quality scores for photos, screenshots, text and charts.
4 min read
Writing a real PNG compressor in vanilla JavaScript (no WASM, no libraries)
Median-cut quantization, Floyd–Steinberg dithering, hand-built PNG chunks, and the four browser traps that cost me an evening each.
6 min read
Are online image converters safe? What actually happens to your files
Most converters upload your file to a server you know nothing about. Here's how to tell which ones do, and when it actually matters.
6 min read
JPG vs PNG vs WebP vs AVIF: which image format should you use?
Five formats, one decision. Here's what each one is actually good at, with a rule of thumb you can apply in ten seconds.
6 min read
How to combine multiple images into one PDF
Every method, on every device — plus the page-size and quality settings that decide whether your PDF is 200 KB or 20 MB.
5 min read
How to make a favicon: every size and format you actually need
Which sizes you actually need, the exact HTML to paste, and the three reasons a favicon usually fails to appear.
5 min read
How to reduce image file size without losing quality
Do these three things in the right order and a 4 MB photo becomes 80 KB with no visible difference. The order matters more than the tools.
6 min read
What is a WebP file, and how do you open one?
Why the internet keeps giving you files you can't open, how to open them anyway, and when WebP is actually the right choice.
5 min read