How to compress a PDF for email without making the text blurry

A 40 MB PDF that will not attach to an email is almost never a text problem. A page of plain text weighs a few kilobytes; a single uncompressed scanned page or an embedded print-resolution photo can weigh several megabytes. Blurry text after compression is also rarely the compressor's fault — it happens when a tool "flattens" the document by turning every page into a picture and then squeezing that picture. This guide shows how to find where the bytes actually are, shrink those parts, and keep the text layer intact so the document stays crisp and searchable.

Step 1: Know the limit you are aiming at

Attachment limits differ, and the limit that matters is the smaller of your limit and your recipient's. Widely published limits are 25 MB for Gmail (sending and receiving), 20 MB for Outlook.com and 25 MB for Yahoo Mail. Many company and university mail servers are stricter — 10 MB is common, and some government and healthcare portals accept only 5 MB or less per file. Two practical rules follow. First, aim for the portal's number, not your mail app's: if a form says 10 MB, deliver 9 MB, not 9.9 MB, because some systems count the encoded size, and email encoding inflates an attachment by roughly a third. Second, if nobody has stated a limit, 10 MB is a safe default target that passes almost everywhere.

Step 2: Work out what is making the PDF big

Open the PDF and look at its ingredients before you compress anything:

• Scanned pages. If you cannot select the text with your mouse, the page is a photograph of a page. A colour scan at 300 dpi is typically 2–8 MB per page as an uncompressed image, and 200–600 KB per page as a sensible JPEG. Twenty scanned pages are a 40 MB file waiting to happen.
• Embedded photos. A document exported from a word processor keeps photos at their original resolution. A 6000 × 4000 phone photo placed in a report at a quarter of a page still travels at full size unless it is downsampled.
• Fonts. Fully embedded fonts add 100 KB to over 1 MB each. Subsetting — embedding only the characters actually used — usually reduces this to a few tens of kilobytes.
• Leftovers. Undo histories, embedded file attachments, document thumbnails and uncompressed object streams can add surprising bulk to files that have been edited many times.

Step 3: Compress in the right order

1. Downsample images first. This is where 80–95% of the savings live. For a document that will be read on screen, 150 dpi for embedded colour and greyscale images is the standard sweet spot: sharp at normal reading zoom, a fraction of the bytes of 300 dpi. Keep 200–300 dpi only if the recipient will print the document. Never go below about 120 dpi for images that contain small text, charts or signatures — those are the first things to turn mushy.
2. Re-encode images as JPEG (or JBIG2/JPEG 2000 where offered). Photographs stored as PNG or uncompressed TIFF inside a PDF compress dramatically as JPEG at quality 80–90 with no visible change on screen.
3. Keep the text as text. In your export or compression settings, reject any option described as "convert text to outlines", "rasterise pages" or "print to image". The moment text becomes pixels, it needs far more bytes to look sharp, it can no longer be searched or copied, and screen readers can no longer read it.
4. Subset fonts and remove leftovers. Enable font subsetting, and discard embedded thumbnails, attachments you did not intend to include, and editing metadata.
5. Check the size, then check the quality. If the file is still too big, lower image resolution one step (150 → 120 dpi) rather than dropping JPEG quality sharply; resolution cuts bytes quadratically, quality cuts them roughly linearly and introduces visible artefacts sooner.

You can do the whole job in your browser with the PDF compressor: pick a quality level, download the result, and compare its size against your target before you attach it.

A worked example

Suppose a 24-page report is 38 MB: 20 scanned pages at roughly 1.8 MB each, plus four pages of text and charts. Downsampling the scans to 150 dpi and re-encoding them as JPEG typically brings each page to about 150–250 KB, so the scans drop from ~36 MB to roughly 4 MB. Font subsetting and removing the embedded thumbnail trims a few hundred kilobytes more. The result lands near 4.5 MB — under every common limit — and because the four original text pages were never rasterised, their text is exactly as sharp as in the original file. The only pages that changed are the scans, and at reading distance a 150 dpi scan is indistinguishable from the 300 dpi original.

Common mistakes that make text blurry

• "Print to PDF" through a photo or fax preset. Some virtual printers render every page to an image at 150 dpi or lower. The output is small and uniformly fuzzy.
• Compressing twice. Running an already-compressed PDF through a second aggressive pass re-encodes JPEG images again; artefacts compound. Always start again from the original.
• Screenshotting pages into a new PDF. A screenshot is capped by your screen resolution and bakes in your zoom level. It is the fastest route to unreadable small print.
• Judging quality at 100% zoom on a small laptop screen. Check at 200% instead — that is where soft text shows — and check the worst page (smallest fonts, densest table), not the title page.
• Shrinking below the real requirement. If the limit is 25 MB, a 20 MB file with 300 dpi images is a better deliverable than a 2 MB file with mushy scans. Compress to the requirement, not to the minimum possible.

When the file still will not fit

If a document cannot reach the limit without unacceptable quality, split it rather than destroying it: send volume 1 and volume 2 as separate emails, or use the PDF splitter to extract the sections the recipient actually needs. For a portal with a hard per-file cap, splitting at a chapter boundary beats one unreadable file every time. And if the recipient only needs three pages of a 90-page document, send three pages — the most reliable compression is not sending bytes nobody asked for.

Frequently asked questions

Why is my one-page PDF 5 MB?

Almost certainly one embedded image at full camera resolution, or the page is a scan. Downsample the image to 150 dpi for screen delivery and the same page usually falls below 300 KB.

Does compressing a PDF reduce its quality?

It reduces image quality by an amount you choose; the text and vector graphics are untouched by a proper compressor. Quality loss becomes visible when images are downsampled too far or re-encoded repeatedly, which is why you should always compress from the original file.

What is the best target size for email?

Under 10 MB works with nearly every mail system. Under 25 MB works for Gmail and Yahoo recipients. Remember that email encoding adds about a third, so a 19 MB attachment can hit a 25 MB message limit.

Can I compress a password-protected or signed PDF?

Encrypted files must be unlocked first, and recompressing a digitally signed PDF invalidates the signature — the signature covers the exact original bytes. Send signed documents unchanged, or re-sign after compression.

My PDF is mostly text but still large. What else can it be?

Fully embedded fonts (common with several typefaces or CJK fonts), embedded file attachments, and accumulated editing data. Font subsetting plus a "remove attachments/metadata" pass usually fixes it without touching quality at all.