Reduce image file size

Start with the biggest lever

If an image is too big, check its pixel dimensions before anything else. File size grows with the number of pixels, and that number grows with the square of the width. A 4000×3000 photo has 12 million pixels. Resized to 2000×1500, it has 3 million. That’s a quarter of the data, before you touch quality or format.

Most phone photos are 12–48 megapixels. A website, an email, or a chat app rarely shows them wider than 1200–2000 px. So resizing first, then compressing, is almost always the right order.

The five levers

1. Dimensions (width × height)

Resize to the largest size the image will actually be displayed at. For a high-density (retina) screen, use up to twice the displayed CSS width. Going bigger than that adds bytes nobody sees.

Use Resize Images and set a width, a height, or a percentage. “Lock aspect ratio” and “Don’t enlarge smaller images” are on by default, so you won’t distort photos or blow up small ones.

2. Format

The same picture can differ several times in size depending on the format:

• JPG suits photos. It has no transparency.
• PNG is lossless. It’s ideal for screenshots, logos, and graphics with sharp edges, but it’s heavy for photos. A PNG photo is often 5–10 times larger than a good JPG.
• WebP handles both photos and graphics, supports transparency, and is usually 25–35% smaller than JPG at similar visual quality.
• AVIF is often smaller still, but encoding is slower.

If an image is going on the web, Convert to WebP is a quick win for almost any input. The most common mistake is a photo saved as PNG. Converting it to JPG or WebP can cut it by 80% or more.

3. Quality

For lossy formats (JPG, WebP, AVIF), the quality setting trades detail for size. Gains are steep at the top of the scale. Dropping from 95 to 80 often halves the file, while dropping from 80 to 65 saves much less and starts to show artifacts. A setting of 70–80 is a good default for photos. Compress Images uses 75 for JPG and WebP.

4. Metadata

Photos carry EXIF data such as camera model, settings, timestamps, sometimes GPS, and often an embedded preview thumbnail. On a large photo this is a small share of the file. On a 30 KB web thumbnail it can be a big one. Converting, resizing, or optimizing in Pixellin drops EXIF automatically. To remove it without touching the pixels, use Remove EXIF Data.

5. Color count

This lever is for graphics, not photos. A screenshot, chart, or logo often uses far fewer than 256 distinct colors. Quantizing a PNG to a limited palette and then optimizing it losslessly typically saves 50–80% with little or no visible change. Pixellin’s PNG compressor lets you choose anywhere from 2 to 256 colors and keeps transparency. For GIFs, reducing the palette to 128 or 64 colors helps in the same way.

Decision table: which lever first?

Your situation	Pull this lever first	Then
Phone or camera photo for a website, email, or form	Resize to display size	Compress JPG at 70–80, or convert to WebP
Photo saved as PNG	Change format to JPG or WebP	Resize if it’s still large
Screenshot or UI capture	Reduce colors (PNG compression)	Resize if it’s wider than needed
Logo or icon with transparency	Reduce colors, keep PNG or WebP	Consider SVG if you have the vector source
Images for your own site	Convert to WebP or AVIF	Resize to 1–2× display width
Form that requires “under 100 KB”	Use target-size compression	It lowers quality first, then dimensions
Animated GIF	Reduce palette, optionally drop frames	Consider video or animated WebP for big animations
Already small JPG that won’t shrink	Check dimensions and metadata	Avoid re-compressing repeatedly

Step by step: shrink a batch of photos

This sequence handles the most common case, a set of photos that are too large to send or upload.

1. Open Resize Images and drop in your photos, up to 200 at a time.
2. Choose Width and enter the largest width you need, such as 1600 px for email or 1920 px for a full-width web image.
3. Set the output format to JPG or WebP, then select Process.
4. Download with Download all (ZIP).
5. If the files are still too large, open Compress Images, add the resized files, keep quality around 75, and process again.

If you have a hard limit, such as a job portal or government form that accepts only files under 50 KB, skip the guesswork and use Compress to Exact File Size. It searches for the highest quality under your target. If that isn’t enough, it reduces the dimensions step by step and tells you the final size and dimensions.

How much can you expect to save?

Rough figures for typical inputs:

Starting point	Action	Typical result
12 MP phone JPG, ~3–4 MB	Resize to 1920 px wide, quality 75	~250–450 KB
Same photo	Also convert to WebP	Usually another 25–35% smaller
1–2 MB PNG screenshot	Quantize to 256 colors	Often 50–80% smaller
PNG photo, ~8 MB	Convert to JPG at quality 80	Often under 1 MB

Actual results depend on the content. Noisy, detailed images, such as low-light photos or foliage, compress less than clean, smooth ones.

Mistakes that keep files large

• Compressing without resizing. Quality changes alone can’t match the saving from fewer pixels.
• Re-saving the same JPG many times. Each save adds artifacts. Work from the original and compress once at the end.
• Using PNG for photos. Lossless is the wrong tool for camera images.
• Upscaling. Making a small image larger adds bytes without adding detail.
• Chasing tiny targets with heavy compression. If quality must drop below about 60 to hit a limit, reducing the dimensions usually looks better.

Pixellin never hands back a larger file from its compressors. If a result would be bigger than the original, the original is kept and labeled “Already optimized.” All processing happens on your device, so even large batches of private photos never leave your browser.