How to Reduce File Size: Images and PDFs, Without Wrecking Them

Almost every file that needs shrinking is large for one of two reasons, and knowing which one you are looking at decides everything that follows. Guessing at a compression preset and hoping is how people end up with a blurry document that is still too big.

For images, the image compressor works against a target size in your browser; for documents, compress PDF does the same. Neither uploads the file, which matters when the reason it is large is that it is a scan of something personal.

Find out what is making it large

For an image, the size is dominated by pixel dimensions first and compression quality second. A photo straight off a modern phone is around 4000 pixels wide. Almost nothing needs that: a full-width web image wants about 2400, an email attachment about 1600, a profile picture about 800. Halving the width cuts the file to roughly a quarter before you touch quality at all, and it is invisible if the image was never going to be displayed at 4000 pixels.

For a PDF, open it and try to select a line of text. If the cursor selects words, the document is text-based and there is very little to remove — the size is fonts and structure, both already compact. If nothing selects, the pages are images, and image resolution is where nearly all of the size lives. That single test predicts how much you can save more reliably than any tool's estimate.

Reducing an image

Work in this order, because it is the order that preserves the most quality per byte saved.

1. Resize first. Set the pixel dimensions to what the destination actually needs with the resizer. This is lossless in the sense that matters — you are removing detail nobody was going to see.

2. Then adjust quality. JPEG quality between 75 and 85 is the sweet spot for photographs: below 70, artefacts appear around edges and in flat areas like skies. On screenshots and images containing text, JPEG is the wrong format entirely and its artefacts are obvious at any setting.

3. Then reconsider the format. WebP typically produces a smaller file than JPEG at the same visual quality, often by 25 to 35%, and every current browser reads it. Converting to WebP is frequently the largest single saving available, and it is the right choice for anything going onto a web page. For a file you are sending to somebody else, JPEG remains the safer target.

One trap worth naming: converting a photograph to PNG makes it substantially larger, often several times. PNG is lossless and is right for screenshots, logos and line art. Photographs belong in JPEG or WebP.

Hitting a specific KB target

Portals and forms are fond of exact limits — 200KB, 500KB, 2MB, and the 2×2 inch photograph some visa applications require. Work toward the number rather than at it.

For a photo, set the dimensions first: a 2×2 inch photo at 300 DPI is 600×600 pixels, which is the whole calculation, and at that size hitting 200KB is comfortable. Then step the quality down and check the result each time. The point where a face starts to look mottled is visible immediately and is the boundary to stop just short of.

Do not enlarge an image to meet a minimum dimension. Enlarging adds pixels without adding detail, so the file grows and the picture gets softer — the worst of both.

Optimising a PDF

For a scanned PDF, the size is image resolution and you have real headroom. A 600 DPI scan carries roughly nine times the pixel data of the same page at 200 DPI, and 200 DPI reads cleanly on any screen. Reductions of three quarters or more are normal and often invisible.

For a text-based PDF, be prepared for very little. The genuinely lossless savings — removing duplicate objects, subsetting embedded fonts to the characters actually used, dropping unused resources — are real and modest. Any tool promising to cut a text PDF by 90% with no visible change is about to rasterise your pages, which makes the text unselectable and unsearchable. If a text PDF is too large, it is usually too long: splitting it and sending the relevant section is the better answer.

One ordering rule. If the document needs to be searchable, run OCR before compressing. Recognition depends on the resolution that compression is about to remove.

On a Mac

Preview has a built-in option — File → Export, then Quartz Filter → Reduce File Size — and it is worth knowing its reputation. It is aggressive, it has no settings, and it frequently produces a document noticeably worse than it needed to be. On a photo-heavy PDF it can work well. On a text document it does very little, and on a scan it often overshoots.

ColorSync Utility lets you build a gentler custom Quartz filter, which is the traditional Mac answer and a fiddly one. For most people, a browser tool with a visible quality control is less work and produces a better result. It also runs identically on macOS, Windows and a phone, which the Preview route does not.

The "make it high resolution" problem

Requests to increase an image's resolution come up constantly and deserve a straight answer: you cannot add detail that was never captured. Enlarging interpolates — it invents plausible pixels between the real ones — so the file gets bigger and the picture gets softer. AI upscalers do this more convincingly by guessing what should be there, which is fine for a wallpaper and not fine for anything where accuracy matters.

If a print shop asks for 300 DPI, what they need is enough pixels for the physical size: a 6×4 inch print at 300 DPI needs 1800×1200 pixels. If your image is 900×600, no amount of resampling makes it a genuine 300 DPI print. The honest options are to print it smaller or to reshoot it.

Frequently asked questions

How do I reduce an image’s file size?

Resize first, then lower the quality, then reconsider the format. Pixel dimensions dominate file size — halving the width cuts the file to roughly a quarter before quality is touched. JPEG quality between 75 and 85 is the usual sweet spot for photographs.

How do I shrink a JPEG to a specific number of KB?

Set the dimensions the destination actually needs first, then step the quality down and check the result each time. For a 2×2 inch photo at 300 DPI that means 600×600 pixels, at which 200KB is comfortable to reach.

How do I compress a PDF on a Mac?

Preview has File → Export → Quartz Filter → Reduce File Size, but it is aggressive and has no settings, so it often overshoots on scans and does very little on text documents. A browser compressor with a visible quality control gives a better result and works the same on any platform.

Why did my PDF barely get smaller?

Because it is text-based rather than image-based, so there is very little to remove — the content is fonts and structure, both already compact. Large text PDFs are usually large because they are long, which makes splitting the better answer.

Does converting an image to PNG make it smaller?

For a photograph, no — usually several times larger, because PNG is lossless. PNG is right for screenshots, logos and line art. Photographs belong in JPEG or WebP, and WebP is typically 25 to 35% smaller than JPEG at the same quality.

Can I increase an image’s resolution to make it print better?

Not meaningfully. Enlarging interpolates — it invents pixels between the real ones — so the file grows and the picture softens. If a printer wants 300 DPI for a 6×4 print, that means 1800×1200 real pixels. If you have fewer, print smaller or reshoot.

Should I compress before or after running OCR?

After. Recognition accuracy depends on image resolution and compression works by reducing exactly that. Run OCR first, then compress the result.

In short

Two rules cover most of it: find out whether you are dealing with pixels or with text before choosing a tool, and change dimensions before changing quality. For a PDF specifically, the compression guide works through the size targets in more detail.

Written by Mehran

Founder of QuickMerge. Practical guides built around the real behavior and limits of the tools. Connect on LinkedIn.