Understanding PDF Compression Levels — What Gets Lost and What Stays

Compression is the single most common PDF operation, and the most misunderstood. People run it with a vague idea that it makes files smaller, download the result, and either send it with no further review or panic when the text suddenly looks blurry. Understanding what a compression level actually does — what it throws away, what it keeps, and why the same setting can feel perfect on one document and awful on another — makes the tool far easier to use.

This article takes compression apart. It explains the main things that happen inside a PDF when you compress it, what high, medium, and low quality settings really mean, and how to pick the right level for the way the file will be used.

What lives inside a PDF

To understand compression, it helps to know what a PDF contains. A typical PDF has four kinds of content:

• Text, as actual characters with fonts and positioning data
• Vector graphics, like lines, shapes, and diagrams that are described mathematically
• Raster images, like photos, screenshots, and scans, stored as pixel grids
• Metadata, including author, creation date, and structural information

Text and vector graphics are already quite efficient. Compression reduces them mostly by removing duplicate font data, simplifying file structure, and using more efficient encoding. Raster images are where the real file size usually lives, and they are where compression has the biggest impact — for better and worse.

The three things compression does

When a PDF compression tool runs, it does three things in varying combinations.

1. Image downsampling

The tool looks at every image in the PDF and asks whether it is stored at a higher resolution than the page actually needs. A photograph scanned at 600 DPI but shown in a half-page slot probably does not need to be stored at 600 DPI. Downsampling reduces the resolution to match the display size, and the file shrinks.

This is the single biggest lever. A heavy PDF often drops from fifteen megabytes to three just from downsampling.

2. Image re-encoding

The images themselves may be in an inefficient format. Compression tools often re-encode them using JPEG or JBIG2, both of which are more space-efficient than uncompressed TIFF or BMP. JPEG is lossy — it throws away detail to save space — and the quality setting determines how much detail is thrown away.

3. Object stream compression

The non-image parts of the PDF — fonts, text, structure, metadata — get compressed using general-purpose compression like Flate (the same algorithm behind ZIP). This is lossless. Nothing is lost, but the file gets smaller because the compression removes redundancy.

A "high quality" compression usually runs only object stream compression and minimal image re-encoding. A "maximum" or "low quality" compression runs all three aggressively.

What the quality levels actually mean

Most compression tools offer three or four presets. They usually correspond to something like this.

High quality / low compression

The tool does object stream compression and mild image optimization. Images keep most of their resolution and quality. Text stays crisp. The file may drop by twenty to forty percent depending on how inefficient the original was.

Use this when the PDF will be printed, when images need to stay sharp, or when the file will be used in a professional context where visible quality loss is unacceptable.

Medium quality / recommended

A middle ground. Images are downsampled to around 150 DPI (fine for most screen reading and decent for non-premium printing) and re-encoded at a moderate JPEG quality. Text is untouched. The file often drops by fifty to seventy percent.

Use this for most everyday sharing: email attachments, web uploads, client drafts, internal documents.

Low quality / maximum compression

Aggressive. Images drop to around 72 DPI (screen resolution for older displays) and re-encoded at lower quality. The file often drops by seventy to ninety percent.

Use this only when you really need the smallest possible file — an email gateway with a strict two megabyte limit, for example — and you can accept that images and signatures may look noticeably softer. Never use this for anything that will be printed or archived long-term.

Why the same level feels different on different documents

A compression setting is not a fixed promise about file size or visible quality. It interacts with the content inside the PDF.

A text-only document barely changes. Text is small to begin with, and there are no images to downsample. Compression might drop a long contract from 2 MB to 1.6 MB, and the reader will not notice any difference.

A photo-heavy document changes dramatically. A real-estate brochure with fifteen full-bleed photos can drop from 40 MB to 4 MB on a medium setting. High quality will lose less detail; low quality may leave the photos looking soft.

A scanned document is the most variable. Scans are typically large because each page is a full-resolution image. Aggressive compression can make the text unreadable; too little compression leaves the file heavy.

A mixed document — text, photos, diagrams — lands somewhere in the middle. The compression rate depends on how much of the file is images.

Because the interaction is content-dependent, it is worth running the compression once at a middle setting and checking the result. If the result is too soft, redo at a higher quality. If the file is still too big, redo at a lower quality.

The trade-offs you should care about

Readability versus size

Small files are convenient. Unreadable files are useless. Always check that body text is still crisp after compression. Zoom to 100% and scroll through a sample of pages. If text has visible halos or letters look fuzzy, the compression was too aggressive.

Signatures and stamps

Signatures and stamps often start as scanned images. They are unusually sensitive to compression artifacts. If a signature looks smeared or pixelated after compression, go back to a higher-quality setting. Bad signatures undermine trust in the whole document.

Printing quality

A PDF destined for a printer needs images at 300 DPI for color and at least 150 DPI for grayscale. A medium or aggressive compression may downsample below that, and the printed output will look pixelated. Keep print-destined PDFs at high-quality compression.

Searchability

Compression does not change whether a PDF is searchable, but it can reduce the accuracy of the text layer if OCR has already been applied. If you compressed an OCR'd file and the search results feel wrong, the OCR was likely baked against images that have since been degraded. Re-run OCR on the compressed file if you need search to work properly.

A simple rule of thumb

For a typical knowledge-worker PDF — proposals, reports, handouts, invoices — a medium compression setting produces a file under five megabytes with text that still looks clean. That is the right default.

Go higher quality when the file will be printed, archived, or reviewed by a client who cares about visual polish.

Go lower quality only when there is a specific size limit you cannot meet any other way, and check the result carefully before sending.

Compressing twice makes things worse

Every compression pass introduces a small amount of loss on any image that gets re-encoded. Compressing a file twice is not the same as compressing it once at the same level. The second pass works with the already-degraded version and adds another layer of loss.

Compress once, check the result, and keep the uncompressed source separately. If you need to go smaller, go back to the source and run a single, more aggressive pass — not another pass on the compressed output.

A checklist before you compress

Before you run the Compress PDF tool, run through this short list:

• Is the file already small enough? Skip compression for files under 2 MB.
• Is it going to be printed? Pick a high-quality setting.
• Does it contain signatures? Check the quality after compression.
• Is it destined for email? Aim to get under 10 MB so it clears most gateways.
• Do you have the uncompressed source saved somewhere? If not, make a copy before compressing.

Put it into practice

Understanding what compression is doing turns a confusing set of buttons into a straightforward decision. You are choosing how much image detail you can afford to lose in exchange for a smaller file, and that answer depends on what the file is for. Print-destined PDFs deserve high quality. Everyday sharing usually lands on medium. Emergency size-limit fixes sometimes need aggressive compression and a careful review.

When you are ready to apply compression to your own files, open the Compress PDF tool. Start at a medium setting, scroll through the result, and step up to higher quality if anything looks off. Every upload is encrypted, the compression runs in your browser, and the file is deleted automatically within two hours.