How to Reduce PDF File Size Without Ruining Quality
In most oversized PDFs, images take up most of the space. You can usually make the file much smaller by downsampling those images to the resolution you actually need: about 150 ppi for reading on screen and 300 ppi for print. Start with lossless cleanup, then downsample images and recompress photos as JPEG. Done this way, these steps rarely cause visible quality loss.
Problems start when a tool compresses everything the same way. Text in scanned pages turns blurry, screenshots pick up smudgy artifacts, and some shortcuts quietly remove links, bookmarks, and form fields. This guide covers what makes a PDF big, what each reduction method costs you, and how to check the result before you send it.
One term comes up throughout. ppi (pixels per inch) is how many image pixels land on each inch of the page at the size the image is placed. Many tools call this "dpi," and in this context the two mean the same thing.
What Makes a PDF File Large
A PDF is a container. Text drawn with fonts takes up very little space, but some other kinds of content can make the file balloon.
Embedded images at high resolution
A PDF stores each image at the pixel size it was embedded with, which is often the full original, not the size it appears on the page. A 12-megapixel phone photo shrunk to a 3-inch thumbnail can still carry all 12 million pixels. Some authoring apps downsample pictures when they export, but many don't, and this is the most common reason a 10-page report ends up at 40 MB.
Scanned pages
A scanned page is one large image, with no live text in it. A letter-size page (8.5 × 11 in) scanned in color at 300 ppi is 2,550 × 3,300 = 8,415,000 pixels. At 3 bytes per pixel, that comes to about 25 MB of raw data per page before compression. Grayscale needs a third of that. Pure black-and-white (1 bit per pixel) needs 8,415,000 ÷ 8 ≈ 1.05 MB, which is 1/24 of the color version, and it compresses far better on top of that.
Embedded fonts
PDFs embed fonts so the document looks the same on every device. Most tools embed a subset, meaning only the characters you actually used, and that costs little. A fully embedded font can cost a lot more, especially a large Chinese, Japanese, or Korean font, which can run to several megabytes.
Duplicated resources
If you merge several PDFs, each one may bring its own copy of the same logo, background, or font. The merged file then stores those copies again and again, even though one would do. This is a common side effect of merging PDF files.
Incremental saves
The PDF format lets an editor add changes to the end of the file without rewriting it. That makes saving fast, but the old versions of edited objects stay inside the file. A document that has been annotated, signed, and edited many times can carry a lot of this dead weight.
Other extras
Attached files, embedded thumbnails, hidden layers, large metadata blocks, and leftover editing data from design apps all add size. They're often invisible when you read the document.
How to Find Out What Is Taking Up Space
Figure out what's large before you change anything, because the right fix depends on it.
- Try to select the text. If you can't, the pages are scanned images, and image settings will matter most.
- Divide file size by page count. Pages of pure text are cheap. If each page costs megabytes, the cause is almost certainly images or scans.
- Look for a space audit. Some full-featured PDF editors include a space-usage report that breaks the file down into images, fonts, and document overhead. If yours has one, use it.
- Think about where the file came from. Merged files suggest duplicate resources. Files that were edited many times suggest incremental-save bloat. Exports from design software may have full-resolution images and editing data.
Image Downsampling: Choosing the Right Resolution
Downsampling lowers an image's pixel count to a target ppi. The image keeps its size on the page, and it has fewer pixels to store. Because pixel count goes with the square of the resolution, halving the ppi cuts the pixel count to a quarter.
Worked example: phone photos in a report
Say a report contains a 4032 × 3024 phone photo placed 6 inches wide.
- Current resolution: 4032 pixels ÷ 6 inches = 672 ppi. The height on the page is 3024 ÷ 672 = 4.5 inches.
- Target for on-screen reading: 150 ppi.
- New pixel size: 6 in × 150 = 900 pixels wide, and 4.5 in × 150 = 675 pixels tall.
- Pixel count before: 4032 × 3024 = 12,192,768. After: 900 × 675 = 607,500.
- Ratio: 607,500 ÷ 12,192,768 ≈ 0.05, so about 5% of the original pixels.
Compressed file size doesn't fall in exact proportion to pixel count. Still, the image data shrinks dramatically, and at the size it's shown on screen, the photo looks the same.
Resolution guidelines by use
| Intended use | Photos and color images | Scanned text and line art |
|---|---|---|
| Email, web, on-screen reading | 100-150 ppi | 150-200 ppi grayscale, or 300 ppi black-and-white |
| Office or home printing | 200-300 ppi | 300 ppi; 600 ppi for black-and-white with fine print |
| Commercial or professional print | 300 ppi at final size (confirm with the printer) | Follow the printer's specifications |
| Archiving | Keep the original resolution | Keep the original resolution |
Most tools let you set a threshold, so only images above a certain resolution get downsampled. A sensible setting is to downsample to 150 ppi any image above about 225 ppi. That leaves images that are close to the target alone, so they aren't needlessly re-encoded.
JPEG Recompression and When to Avoid It
Downsampling reduces the number of pixels. Recompression changes how those pixels are stored. JPEG is a lossy format: it throws away detail the eye is unlikely to notice in photos. Flate (also labeled ZIP) is lossless, so it keeps every pixel exactly.
- Photos: Use JPEG at a medium-high quality setting. The quality scales aren't the same from tool to tool, so compare results at 100% zoom rather than trusting a number.
- Screenshots, charts, diagrams, and text-heavy images: Use lossless compression. JPEG creates halos and blotches around sharp edges, which makes small text hard to read. The same reasoning applies to standalone image files, as explained in JPG vs PNG vs WebP.
- Black-and-white scans: Use a dedicated 1-bit compression such as CCITT Group 4 or lossless JBIG2.
Two warnings. First, recompressing a JPEG that has already been compressed adds new artifacts on top of the old ones, and the damage builds up each time you run a file through a compressor. Second, lossy JBIG2 replaces similar-looking shapes with one shared shape, and it's known to swap similar characters like 6 and 8. That's unacceptable for invoices, contracts, or anything with numbers. Turn lossy JBIG2 off for documents like these.
Removing Unused and Duplicate Data
Cleanup doesn't touch image quality, so it's the safest step and the one to try first.
- Save a fresh copy. Rewriting the file with a full save, usually by choosing "Save As" or exporting rather than plain saving, drops the old object versions left by incremental saves.
- Merge duplicates. Optimizers can detect identical images and fonts and keep one shared copy.
- Use compact structure. PDF 1.5 and later can pack the file's internal objects into compressed "object streams." Optimizers often turn this on. The result needs a reader that supports PDF 1.5, which almost every reader in use today does.
- Remove extras you don't need: attachments, embedded thumbnails, hidden layers, private editing data, and unused form or script data.
Don't strip tags, the structure data screen readers rely on, unless you're sure accessibility doesn't matter. They usually take up little space, and removing them makes the document harder to use for people with assistive technology.
Built-In Optimize and Reduce Size Features
Most PDF editors offer two kinds of size reduction.
- One-click "reduce size" or "compress": This applies preset image settings. It's quick, but you can't see exactly what changed. Some built-in filters are aggressive and can make photos noticeably soft.
- Advanced optimize panels: These let you set the target ppi, the threshold, the compression type for color, grayscale, and black-and-white images, and which objects to discard. Use this kind when quality matters.
On the command line, the free tool Ghostscript can rewrite a PDF using presets. On Windows, the program is usually called gswin64c instead of gs:
gs -sDEVICE=pdfwrite -dPDFSETTINGS=/ebook -dNOPAUSE -dBATCH -dQUIET -sOutputFile=smaller.pdf input.pdf
The /screen preset targets about 72 ppi, /ebook about 150 ppi, and /printer and /prepress about 300 ppi. For lossless cleanup only, qpdf --object-streams=generate input.pdf output.pdf restructures the file without touching the images. Whatever the tool, check the output afterward, because different tools keep different interactive features.
Print to PDF: Fast but Lossy in Other Ways
Printing a PDF to a virtual PDF printer, like the ones built into Windows and macOS, makes a brand-new file, and that sometimes shrinks it. The cost is that the printer passes on only what it would put on paper. Here's what you usually lose:
- Clickable links and the table of contents bookmarks
- Fillable form fields (they get flattened into static text or blanks)
- Comments, attachments, and layers
- Accessibility tags, and sometimes clean text selection
Print to PDF also doesn't reliably make files smaller. Depending on the driver, it may keep the full-resolution images or turn complex pages into images, and the file can end up bigger. It's a reasonable last resort for a simple document someone will only read or print. Avoid it for forms, contracts, or anything with links.
Methods Compared: Size Savings vs. Quality Impact
| Method | Typical size reduction | Visual quality impact | What can break |
|---|---|---|---|
| Fresh full save | Small to large (depends on edit history) | None | Existing digital signatures |
| Remove duplicates and unused objects | Small to moderate | None | Attachments or layers, if removed by mistake |
| Font subsetting | Small to moderate | None | Later editing of text in the PDF |
| Downsample images to 150 ppi | Large for photo-heavy files | Low on screen; visible in print | Zoomed-in detail |
| JPEG recompression | Moderate to large | Low on photos; high on text and screenshots | Readability of small text in images |
| Scans to black-and-white | Very large | Loses color and shading | Photos, stamps, colored signatures |
| Lossy JBIG2 | Very large | Can swap characters | Accuracy of numbers and letters |
| Print to PDF | Unpredictable | Varies by driver | Links, forms, bookmarks, tags |
Quick Checklist Before You Send the Smaller File
- Keep the original. Compression is one-way, so always work on a copy.
- Pick the target first: about 150 ppi for screen or email, 300 ppi for print.
- Do the lossless steps first (fresh save, duplicate removal, cleanup), then check the size. That may be enough.
- Downsample images with a threshold, and use JPEG only for photos.
- Zoom to 100% and 200% and look at the smallest text, especially on scanned pages and in screenshots.
- Click a few links and bookmarks, and try typing in a form field.
- Check any numbers in scanned tables against the original.
- If the document will be signed, compress it before signing, because re-saving, optimizing, or compressing a signed PDF breaks the signature.
Frequently Asked Questions
Why did my PDF get bigger after I compressed it?
This usually means the tool re-encoded images that were already well compressed, sometimes at a higher quality setting than the original. It can also happen when a tool turns vector graphics or transparent areas into images. Try a lossless cleanup instead, or set a downsampling threshold so images already near the target resolution are left alone.
How do I get a PDF under a specific size limit, like 2 MB for an upload form?
Divide the limit by the page count to get a per-page budget. For example, 2 MB across 20 pages is about 100 KB per page. Text pages need only a fraction of that, so most of the budget can go to images. If photo-heavy or scanned pages still go over, lower the image resolution or convert text-only scans to grayscale or black-and-white.
Does zipping a PDF make it much smaller?
Usually not by much. The text, fonts, and images inside a PDF are typically already compressed, so a ZIP archive has little left to squeeze. Zipping mainly helps when you bundle several files together, not when you need to shrink one PDF.
Is it safe to use an online PDF compressor for confidential documents?
Online compressors upload your file to someone else's server, and how long they keep it depends on the service and its policies. For contracts, medical records, financial statements, or anything confidential, a desktop editor or a local command-line tool is the safer choice because the file never leaves your computer.
Will compressing a PDF make it unreadable for screen readers?
Image compression doesn't affect accessibility, but some optimizers remove tags or document structure to save a little more space, and that hurts screen-reader users. Leave the option to discard tags or structure turned off, and avoid print-to-PDF for documents that need to stay accessible.