Reduce PDF File Size — Keep Quality (2026)

You compressed a PDF to fit an email limit and now the report's photos look like a 90s GIF. The print quality is gone, the screenshots are unreadable, and you're sending it anyway because the deadline's in 10 minutes.

This guide covers what actually causes PDFs to be large, which compression methods preserve quality and which sacrifice it, and how to get the biggest size reduction with the smallest visible change.

What "without losing quality" really means

Strict definition: the output must be byte-for-byte identical in visual appearance to the input. By that definition, only lossless compression counts.

Lossless techniques shrink the file by removing redundancy in how the data is stored — not in what it represents. Examples:

• Re-compressing already-stored streams with better algorithms (Flate, ZIP)
• Removing unreferenced objects (orphaned images, deleted-but-not-purged content)
• Subsetting fonts (storing only the glyphs the document uses, not the entire typeface)
• De-duplicating identical resources (same image embedded twice → stored once, referenced twice)

Lossless compression typically reduces a PDF by 5–30%. For a 30 MB report, that's 21–28.5 MB. Useful, but not always enough.

Pragmatic definition: visibly indistinguishable in normal viewing — same text crispness, same readable photos at default zoom. Under this definition, much more aggressive compression is possible without quality complaints from real-world users.

This guide covers both.

What's making your PDF large

Most PDF bloat comes from a small number of common causes:

If your PDF is mostly typed text and charts, lossless compression already keeps it small — there's not much to remove. If it has lots of photos or scanned pages, lossy compression is where the real shrinkage happens.

The two-pass approach (recommended)

Best results come from running both passes:

1. Lossless pass first. Removes structural bloat without changing how anything looks.
2. Lossy pass second (only if needed). Reduces image sizes with controlled quality loss.

Lossless alone is enough for typed reports and structured documents. Photo-heavy PDFs need both.

Pass 1: Lossless techniques

Subset fonts

If your PDF embeds a font but only uses 50 characters of it, the embedded font weighs 200 KB even though those 50 characters need maybe 5 KB. Subsetting strips the unused glyphs.

Most modern PDF generators subset fonts by default. Some (older Word versions, some web printers) embed full fonts. Run a re-save through any modern PDF generator to enforce subsetting.

Save: typically 50–500 KB per font.

Remove embedded files and attachments

PDFs can contain other files as attachments — the original Word document, supporting spreadsheets, supplementary images. These add up quickly.

To check:

1. Open the PDF in any reader that shows attachments (Adobe Reader → Attachments panel; Foxit → Attachment view).
2. Note any attached files you don't actually need to ship.
3. In the same reader, delete the attachments and re-save.

Save: anything from 0 to several MB.

Flatten form fields and annotations

Interactive form fields, comments, and annotations add overhead. If the form is filled and won't be edited again, flattening converts them into permanent page graphics — same look, less metadata.

Use a Flatten PDF tool or your PDF editor's "Flatten" option.

Save: 5–10% on form-heavy documents.

De-duplicate identical resources

Some PDFs embed the same logo on every page as a separate copy. A 50 KB logo on a 100-page report is 5 MB of duplication.

Most modern PDF compressors do this automatically. If your tool doesn't, re-saving through one that does (e.g., any standalone PDF compressor) handles it.

Save: highly variable — often 0, occasionally 50%+.

Run lossless re-compression

PDFs use stream compression (typically Flate). Older or poorly-tuned PDF generators don't compress as well as modern algorithms. Re-saving through a modern engine often shaves another 5–15%.

This is what "lossless PDF optimisation" tools do. The output looks identical but uses better internal compression.

Pass 2: Lossy image compression

Once lossless techniques are exhausted, the only remaining lever is image quality. Lossy compression trades visible image detail for size.

The best PDF compressors offer multiple quality levels:

PDF text and vector graphics stay sharp at every level — only embedded raster images are downsampled.

What survives compression

These elements stay crisp regardless of compression level:

• Native PDF text (selectable text, not text-in-images)
• Vector graphics (charts generated by Excel/PowerPoint, vector illustrations)
• Tables, lines, borders drawn natively
• Form fields (until flattened)
• Hyperlinks and bookmarks

These elements degrade with lossy compression:

• Photos (JPEG, PNG embedded as images)
• Scanned pages (entire pages stored as images)
• Screenshots (often saved as JPEG/PNG inside the PDF)
• Diagrams stored as raster images (e.g., a chart someone screenshotted instead of pasting as vector)

A PDF that's mostly typed report + a few stock photos: compress aggressively, lose photo detail, text reads identically.

A PDF that's mostly screenshots and diagrams: compress more carefully — readable text inside the screenshots starts going fuzzy at High and breaks at Very High.

Edge cases worth knowing

Scanned PDFs (entire pages are images)

A scanned PDF is essentially a stack of JPEGs in a PDF wrapper. The text on each page is part of the image — there's no native PDF text to preserve.

Compression strategy:

1. Convert to grayscale if the original is color and color isn't meaningful (typed pages, official letters). Save: 50–70%.
2. Reduce image quality aggressively. Scanned typed text is forgiving — even Very High compression keeps letters legible.
3. Run OCR if you need searchable text. OCR adds a tiny invisible text layer; image stays the same size.

A 50 MB color scan often drops to 2–5 MB at greyscale + Medium compression with the typed text still perfectly readable.

PDFs with mixed content

If your PDF has some pages of text and some scanned pages mixed, split into two files, compress each appropriately, then merge:

1. Use Split PDF to separate by page ranges.
2. Compress the text-only pages at Low (no visible loss).
3. Compress the scanned pages at Medium or High.
4. Merge PDF the results.

This avoids forcing aggressive settings on text pages just to shrink the scanned ones.

PDFs that won't shrink at all

If a PDF refuses to compress meaningfully, it's usually because:

• Already compressed. Adobe Acrobat's "Optimised PDF" or any reputable compressor leaves no slack. Re-compressing gains 0–5%.
• Already image-only at low quality. Can't compress what's not there.
• Almost entirely text. A 1 MB text-only PDF won't shrink to 100 KB; the text is already minimal storage.

In these cases, accept the size or use Split PDF to send in chunks.

PDFs that get LARGER after compression

Counterintuitive but real. Some scenarios:

• Compressing an image-light PDF with image-focused settings. The "compression" overhead can outweigh savings on a text-heavy document.
• Re-encoding an already-optimised JPEG as JPEG. Each re-encode adds artefacts AND can grow the file.
• Adding metadata or layers in the compression process.

If your tool's output is larger than the input, don't use that tool for this file. The original was already optimal.

How to choose the right level

Match the compression level to what the PDF will be used for:

When in doubt: try Low first. If it doesn't shrink enough, try Medium. Avoid jumping straight to High unless the file size requirement is non-negotiable.

Verifying the result

After compression, before sending or distributing:

1. Open the compressed PDF at 100% zoom.
2. Compare side-by-side with the original (open both, alt-tab between them).
3. Zoom to 200% on the most detail-critical page (usually a photo or chart).
4. Search for text with Ctrl+F to confirm text is still selectable.
5. Print preview the first page if it'll be printed.
6. Check file size matches your expectation.

If any of these fail, undo and try a less aggressive setting.

Step-by-step: compressing without quality loss for an email

Most common scenario: you have a 35 MB report, need it under 25 MB for Gmail.

1. Open a PDF compressor (e.g., PDFGrover's Compress PDF).
2. Upload the file.
3. Pick Low quality first.
4. If the output is under 25 MB, done. Verify it looks right and send.
5. If still over, try Medium. This is where most email-too-big cases finish.
6. If Medium is still too large, you have a photo-heavy file. Move to High and verify image readability.
7. If High still doesn't fit, the file genuinely needs to be split or sent via cloud link. See our compress for email guide for that workflow.

Common mistakes

Compressing then editing then re-compressing. Each save cycle compounds losses. Make all edits first, compress last.

Trusting "100% no quality loss" claims. Some tools market lossless settings that actually downsample images at 300+ DPI to lower-resolution lossless formats. The output is "technically lossless" at the new resolution but visibly degraded compared to the original. Verify by opening side-by-side.

Compressing for the wrong recipient. A PDF compressed for emailing to a colleague is over-compressed for printing. Keep the original file as your master copy.

Forgetting to flatten before compressing. A form-field-heavy PDF compresses better when flattened first. The form fields, JavaScript, and annotations are removed; only the visible content remains.

Compressing PDFs with embedded fonts you've already subsetted. Diminishing returns. Most modern PDFs have already subsetted; a re-pass through an aggressive font compressor sometimes corrupts encoding without size gain.

Summary checklist

To shrink a PDF without visible quality loss:

1. Identify what's making it large (images? fonts? attachments?).
2. Pass 1 — lossless: remove attachments, flatten forms, re-save through a modern compressor.
3. If still too large, Pass 2 — lossy: compress at Low first, step up to Medium only if Low doesn't fit, and only go to High when nothing else works.
4. Always verify the output before sending — open at 100%, check key images at 200%, confirm text is selectable.
5. Keep the original as your archived master.

For routine compression, PDFGrover's Compress PDF handles the lossless and lossy passes in one tool, supports files up to 500 MB, and processes everything without watermark or signup.