JPEG vs PNG: Which Is Smaller, and Which Should You Use?

Few questions come up as often as whether to save an image as JPEG or PNG. The honest answer is that neither format is smaller in general. Each was designed for a different kind of picture, and choosing the wrong one can make a file several times larger than it needs to be, or make an image look worse than it should. This guide explains how the two formats work, shows which kinds of images each one wins on, and gives you a decision table, quality settings and a local workflow so you can stop guessing.

The short answer

• Photographs and images with smooth gradients: JPEG is typically far smaller than PNG at similar visual quality.
• Screenshots, UI, text, logos, line art and diagrams: PNG looks sharper and is often comparable or smaller, because JPEG struggles with hard edges.
• Transparency: PNG supports it. Standard JPEG does not.
• Web delivery in 2026: WebP and AVIF are usually smaller than either, with broad browser support.

How JPEG works

JPEG is a lossy format built for photographs. It converts the image to a luminance-and-color representation, divides it into small blocks of pixels, applies a mathematical transform to each block and then throws away detail that the human eye is least sensitive to. Color information is commonly stored at lower resolution than brightness, a step called chroma subsampling, because we notice brightness detail much more than color detail.

The quality slider you see in an editor controls how aggressively that detail is discarded. A lower quality number means a smaller file and more visible artifacts. Because the discarding happens in blocks, the typical failure mode is blockiness and a fuzzy halo around sharp edges, which is especially noticeable on text.

How PNG works

PNG is a lossless format. It stores the exact value of every pixel, using filtering and DEFLATE compression to remove redundancy without discarding anything. It supports an alpha channel for full transparency, and it handles flat colors and sharp edges extremely well, because repeated runs of identical pixels compress very efficiently. It struggles on photographs, where almost every pixel differs slightly from its neighbors and there is little repetition to exploit.

Why photos favor JPEG and flat graphics favor PNG

The two formats exploit different properties of an image. JPEG exploits the eye's tolerance for lost fine detail, which is large in natural scenes with noise and texture. PNG exploits repetition, which is large in interface graphics with big areas of uniform color. A photograph saved as PNG is large because there is little repetition. A screenshot saved as JPEG either looks bad, if the quality is low, or gets surprisingly large, if the quality is high enough to keep text clean.

Decision table

Real example: how to test on your own files

Do not rely on a rule of thumb for borderline content. Take a representative image and save it three ways: a PNG, a JPEG at quality 85 and a JPEG at quality 70. Compare the file sizes and zoom to 100 percent. For a photograph, the JPEG will be a fraction of the PNG size and look the same. For a screenshot of a text-heavy interface, you will often find the PNG is similar in size or smaller and the JPEG shows fringing around letters. That test takes two minutes and beats any generalization.

Transparency and JPEG

Standard JPEG has no alpha channel. If you convert a PNG with a transparent background to JPEG, the transparent area has to become a solid color, usually white or black. That is fine for a product photo that will sit on a white page, but it will ruin a logo meant to sit on different backgrounds. If you need both small size and transparency on the web, WebP and AVIF both support alpha and are usually smaller than PNG.

Quality settings that work for JPEG

These ranges are widely used starting points, not rules. The right setting depends on the image and the encoder. A modern encoder such as MozJPEG produces smaller files than a basic one at the same visual quality, and progressive JPEGs load in stages, which improves perceived speed on slow connections.

Avoid generation loss

Every time you open a JPEG, edit it and save it again as JPEG, you re-apply lossy compression to an already-compressed image and quality degrades a little more. Keep your master file in a lossless format such as PNG, TIFF or the editor's native format, and export JPEG only as the final step. The same applies to compressors: always compress from the original, never from a previously compressed copy.

Converting between JPEG and PNG

JPEG to PNG

Converting a JPEG to PNG does not restore lost quality. You get a lossless copy of an already lossy image, and the file will usually be much larger. Do it only when you need transparency editing or a format a tool demands, not to improve quality.

PNG to JPEG

This is the common direction, and it is a good idea for photographs. Choose a quality of 80 to 85, fill any transparency with a suitable background color and compare the result at full size. Do not do this for screenshots with text unless you check the edges.

Beyond JPEG and PNG

If you publish to the web, you have better options. WebP offers both lossy and lossless modes plus transparency, and AVIF can reach even smaller sizes at comparable quality, at the cost of slower encoding. Browser support for both is broad in 2026, but keep a JPEG or PNG fallback if you must support very old browsers or email clients. Our comparison articles on PNG versus WebP and AVIF versus WebP go into the details, and the best image format for websites guide gives a recommendation by use case.

Compress both formats locally

Choosing the right format is half the job. The other half is encoding it well. A JPEG exported from a basic editor can usually be re-encoded with an optimized encoder to a noticeably smaller file at the same visual quality, and a PNG can be optimized losslessly or quantized.

TinyPixels handles both on your own machine. It uses MozJPEG for JPEG, oxipng for lossless PNG and imagequant for lossy PNG, processes whole folders in parallel and can convert to WebP or AVIF when you are ready. Because everything runs locally on macOS or Windows, nothing is uploaded, which matters for client photos and unreleased screenshots. Pro is a one-time purchase with no subscription.

A simple workflow

1. Decide the content type: photograph, flat graphic, text-heavy screenshot or image needing transparency.
2. Pick the format from the decision table. If unsure, save one test copy in each.
3. Resize to the largest size you will display before encoding.
4. Encode with an optimized encoder and a sensible quality setting.
5. Compare at 100 percent zoom and keep the smallest version that looks right.
6. Keep the original master file untouched.

Common mistakes

• Saving screenshots as JPEG and wondering why text looks smudged.
• Saving photographs as PNG and shipping files five to ten times larger than needed.
• Re-saving a JPEG repeatedly.
• Converting JPEG to PNG expecting better quality.
• Forgetting to resize. Pixel count affects size more than format does.

Chroma subsampling, progressive encoding and other settings

Beyond the quality number, three encoder options influence JPEG size and appearance. Chroma subsampling stores color at lower resolution than brightness. The common 4:2:0 setting saves space and is invisible on most photographs, but it can smear the edges of saturated colored text or thin red lines on a flat background. If a JPEG must contain such content, a setting with less subsampling keeps edges cleaner at the cost of size.

Progressive encoding writes the image in several passes, so a rough version appears quickly and sharpens as data arrives. Progressive files are often slightly smaller than baseline files and feel faster on slow connections, so they are a sensible default for web photos. Finally, metadata such as camera EXIF data, GPS location and embedded thumbnails can add tens of kilobytes. Strip it when publishing, both to save space and to avoid leaking the location where a photo was taken.

Choosing by workflow

• Photographers delivering to clients: keep the master in a lossless or RAW format, export JPEG at 85 to 92 for delivery.
• Developers shipping to a website: export WebP or AVIF with a JPEG fallback, resized to the layout width.
• Designers handing off interface: keep PNG for crisp graphics, compress it, and export photographs as JPEG.
• Support and documentation teams: PNG for screenshots with text, because blurry text wastes a reader's time.

Email, social and messaging considerations

Platforms often recompress what you upload. Social networks and messaging apps re-encode images to their own settings, so uploading an already heavily compressed JPEG compounds the loss. For these destinations, upload a high-quality JPEG or a clean PNG at the platform's recommended dimensions and let the service do its single pass. For email, keep attachments small by resizing first, since many mail providers reject large messages, and use JPEG for photos and PNG for screenshots as usual.

Print is another case with its own rules. Printers prefer high-quality, high-resolution files, often in formats and color spaces specified by the print shop. Do not apply web-style compression to files destined for print, and keep embedded color profiles when the printer asks for them.

Conclusion

JPEG is smaller for photographs, PNG is better for crisp graphics and transparency, and neither is the best choice for modern websites. Match the format to the content, resize first, use a good encoder and always keep an original. A short test on your own files settles most arguments, and a local batch tool makes the winning choice easy to apply at scale.

Frequently asked questions