The short answer: find out what is making your PDF large before you compress it. If the file is mostly typed text, a lossless cleanup is usually enough and nothing visible changes at all. If it is a scanned document or a photo-heavy report, the images are the problem, and the fix is to reduce their resolution to a level that still looks sharp on screen, typically 150 DPI, rather than squeezing the whole file with a single aggressive setting. Text in a PDF is stored as outlines, not pixels, so it stays sharp no matter how hard you compress the images around it. Once you know which kind of PDF you have, the right move is obvious, and you stop wasting time on tools that promise a 90 percent reduction on a file that has almost nothing left to give.
This guide works through the diagnosis first, then the numbers behind DPI, then a step by step workflow you can use for email limits, uploads, and everyday sharing. It is written around what actually happens inside a PDF, not around a single tool's marketing claims, and it is honest about one important limit: our own PDF Compressor is a lossless optimiser. It tightens the file structure without touching image quality, which is exactly right for many files and not enough for a heavy scan. Where that is the case, this guide tells you the next step instead of pretending otherwise.
First, find out why your PDF is large
There are really only three kinds of oversized PDF, and each one needs a different fix. Open your file and try to select a sentence with your mouse. If the text highlights and you can copy it, you have a text based PDF, created from Word, Google Docs, a presentation, or a typesetting program. If nothing selects and each page behaves like a picture, you have a scanned PDF, made by a scanner or a phone camera. The third kind is a hybrid: typed text plus large embedded photos, charts, or screenshots, which is what most business reports and slide decks actually are.
| PDF type | How to spot it | Typical size | What shrinks it |
|---|---|---|---|
| Text only | Text selects and copies; file opens fast | A 50 page contract is often under 1 MB | Lossless cleanup: redundant objects, duplicate resources, font subsetting, metadata |
| Text plus photos | Text selects, but the file is heavy | A 20 page report with photos can reach 25 to 50 MB | Image recompression and downsampling; text is untouched |
| Scanned | No selectable text; every page is an image | One scanned page at 600 DPI can be 2 to 5 MB; a 20 page scan can reach 40 to 80 MB | Downsampling the page images, or re-scanning at a lower DPI |
The diagnosis matters because it sets your expectations. A text only PDF has almost no fat to trim: a lossless pass typically saves 5 to 15 percent, and a tool that promises far more is either lying or flattening your document into images, which destroys the selectable text. A scanned or photo-heavy PDF, by contrast, can often shrink by 60 to 80 percent with no visible difference on screen, because most of its bytes are image pixels nobody will ever see at full resolution. If your file barely shrank after compression, that is usually good news in disguise: it means the structure was already tight, and the weight is in the images, which need a different step.
What "without losing quality" actually means
People use this phrase to mean two different things, and confusing them causes most bad compression decisions. Strictly lossless compression changes no content at all: it removes redundant objects, deduplicates repeated images, subsets fonts so only the characters actually used are embedded, and re-encodes the internal streams more efficiently. The document looks byte for byte identical, and the saving is modest, usually 5 to 15 percent. Visually lossless compression goes one step further: it recompresses the embedded images at a lower resolution or quality level, but keeps them sharp enough that a normal reader at normal zoom cannot tell the difference. The saving is large, and the quality loss is real but invisible in normal use.
The key fact that makes both approaches safe is that typed text and vector graphics are never recompressed. A letter in a PDF is not a grid of pixels; it is a shape drawn from a font outline at whatever resolution the screen or printer asks for. Compressing the photos in a report therefore cannot make its headings blurry, and compressing a contract cannot make its clauses harder to read. Only the raster images, the photos, scans, and screenshots, are affected, and only if you choose a lossy step. If a compressed file comes back with blurry text, something went wrong: the tool flattened pages into images instead of compressing them properly, and you should go back to your original and use a different tool.
DPI and downsampling: the lever that actually moves file size
DPI, dots per inch, says how many image pixels are packed into each inch of printed page. A scanned A4 page at 600 DPI is 4,960 by 7,016 pixels, about 34.8 million pixels. The same page at 150 DPI is 1,240 by 1,754 pixels, about 2.2 million pixels. Dropping from 600 to 150 DPI therefore cuts the pixel count by a factor of 16, which is why DPI dominates every other compression setting. Quality settings and file formats matter, but they are adjustments on top of a pixel count that DPI has already decided.
That does not mean you should always pick the lowest DPI. The right target depends on how the document will be read, and picking it in advance saves you from compressing twice:
| How the PDF will be used | Sensible image DPI | Why |
|---|---|---|
| Read on screen or sent by email | 150 DPI | Sharp at normal zoom and on most phone screens, with a large size saving |
| Printed on an office printer | 150 to 300 DPI | 150 DPI prints acceptably; 300 DPI keeps fine detail in photos |
| High quality print or archive | Keep the original resolution | Do not downsample at all; use lossless cleanup only |
| Black and white scanned text | 300 to 600 DPI | Text edges need more resolution than photos, but monochrome pages compress very efficiently, so the cost is lower than it looks |
A worked example shows how the numbers play out. Take a 40 MB scanned contract, about 20 pages photographed at high resolution. A balanced recompression, downsampling the page images to around 150 DPI with a sensible JPEG quality level, typically brings that file to roughly 8 to 12 MB, a reduction of 70 to 80 percent. The text in the scan stays readable because scanned text is itself an image, and at 150 DPI a printed letter is still rendered with enough pixels to read comfortably on screen. What you lose is the ability to zoom deep into a signature or a fine print footnote and keep it perfectly crisp, and the ability to print at poster quality. For an email attachment, that is almost always the right trade. For an archival copy, it is not, which is why you should always keep the original and compress a copy.
JPEG quality is the second dial, and it has a clear sweet spot. For photographs inside a PDF, a quality level of 80 to 90 is visually near lossless for almost any photo: colours stay true and smooth areas like skies and skin stay clean. Below 70, blocky artefacts start to appear in exactly those smooth areas, and below 50 they become obvious. If a tool only offers Low, Medium, and High presets, start at Medium, open the result, zoom to 100 percent, and check a photo page and a text page. If both look fine, you are done. If the photos look soft, step up one level. If the file is still too large, step down one level, but always step from the original file, never from an already compressed copy, or you stack quality loss on quality loss.
The workflow that actually works
Work in this order, and stop as soon as the file fits its purpose. Each step is cheaper in quality terms than the next, so there is no reason to start with the most aggressive one.
Step 1: run a lossless pass first. Drop the file into our PDF Compressor, which re-serialises the document structure, drops unused objects, and optionally strips metadata. If the tool reports that the file is already optimally compressed, believe it: the structure was tight, and the remaining weight is images. This step costs nothing in quality, it runs entirely in your browser so a confidential document never leaves your device, and it sometimes solves the problem outright for text heavy files.
Step 2: if the file is still too large and image heavy, recompress the images with a desktop tool that offers DPI control, such as Adobe Acrobat's PDF Optimizer or a free desktop optimiser, setting colour images to 150 DPI and JPEG quality around 80 for screen and email use. Keep the original, compare the result at 100 percent zoom, and check that typed text is still selectable, which confirms the document was compressed rather than flattened.
Step 3: if quality is sacred and the file still does not fit, split it instead of squeezing it further. Our Split PDF tool divides a document by page ranges, for example 1 to 20, 21 to 40, and 41 to 60, with no quality loss at all, because splitting copies whole pages rather than re-encoding them. A 60 MB report split into four roughly 15 MB parts fits a Gmail message four times over, and each part keeps its full original quality. If only some pages matter to the recipient, use our Extract PDF Pages tool to send just those pages, which is often the fastest fix of all.
Step 4: rebuild only if you have to. If you still own the source document, re-export it with smaller images: set Word or PowerPoint to 150 DPI for embedded images before exporting, or resize oversized photos before they go into the document. Rebuilding from source gives cleaner results than any after the fact compression, because the images enter the PDF at the right size in the first place. If your workflow creates PDFs from images, our Merge PDF tool combines the finished pages, and stripping document properties with our PDF Metadata Remover trims a little more while also removing author names and software details you may not want to share.
Getting under an email or upload limit
Most "file too large" errors are really limit errors, and the limit is lower than the number on the error message suggests. Email attachments are encoded as text before they travel, which adds roughly a third to their size, so a 25 MB file on disk becomes about 33 MB in transit. Aim comfortably under the limit, not exactly at it.
| Service | Typical attachment limit | Safe target for your PDF |
|---|---|---|
| Gmail, personal account | 25 MB per message | 18 MB or less |
| Outlook.com | 20 MB | 15 MB or less |
| Work or school Exchange mailbox | 10 MB by default, set by the administrator | 8 MB or less, or ask your IT team |
| Web forms and job portals | Often 5 MB, sometimes 1 MB | Compress, then split if needed; a form limit is a different fight from an email limit |
Do the split maths before you compress harder. If a 60 MB report must fit a 25 MB Gmail message, you need roughly a two thirds reduction, which a 150 DPI recompression usually delivers for a scan. If it must fit a 5 MB portal upload, you need a twelvefold reduction, and no honest compression setting will get a photo heavy file there without visible damage. In that case, split the document, send only the pages the form needs, or share a cloud link instead. A link is not a defeat: for a print ready portfolio or a high resolution catalogue, it is the correct answer, because it preserves quality that no compression setting can.
What not to do
Do not recompress an already compressed file to chase a smaller number. Each lossy pass throws away a little more image detail, and the second pass starts from an already degraded image, so artefacts accumulate fast. Always return to the original. Do not use print to PDF as a compression shortcut without checking the result: printing re-renders the document, which can flatten selectable text into images and can actually grow a file that was already well optimised. Do not screenshot PDF pages and rebuild a PDF from the screenshots, which caps every page at screen resolution and destroys text selection. And do not delete pages you might need later just to hit a size target: extract the pages you need into a separate file and keep the original intact.
Keeping PDFs small from the start
The cheapest compression is the one you never need. When you scan paper documents, 150 to 200 DPI is enough for text you will read on screen, and 300 DPI is plenty for anything you will print on an office printer; 600 DPI is for archiving and for fine print you expect to zoom into later, not for a routine email. When you export from Word or PowerPoint, set embedded images to 150 DPI before exporting and discard editing data. When you build a PDF from photos, resize each photo to the size it will actually be displayed at before adding it: a 4,000 pixel wide photo printed at 6 inches wide only needs about 900 to 1,800 pixels, depending on print quality, and the extra pixels are pure file weight. Finally, avoid embedding the same large image on every page, for example a full page background logo, when a smaller version or a vector version would do. These habits keep a new PDF small without any compression step at all, which is the best outcome for both quality and file size.
The rule to take away is simple. Diagnose first, clean losslessly second, downsample images to 150 DPI for screen and email third, and split rather than crush when quality matters more than keeping everything in one file. Follow that order and a PDF that is too large to send becomes a five minute fix instead of a quality gamble. For the shorter version of this workflow, our earlier guide to compressing a PDF covers the basic tool steps, and our complete guide to working with PDFs online walks through merging, splitting, and compressing as one workflow. If the document also carries metadata you would rather not share, our PDF metadata guide explains what a PDF reveals about you and how to remove it.
Frequently Asked Questions
- Will compressing a PDF make its text blurry?
- No, not if the tool compresses properly. Typed text in a PDF is stored as font outlines, not pixels, so it is never recompressed and stays sharp at any zoom level. Only embedded photos, scans, and screenshots are affected by lossy compression. If text does come back blurry, the tool flattened pages into images instead of compressing them, and you should go back to the original and use a different tool. Text in a scanned PDF is an exception: it is part of the page image, so heavy downsampling can soften it, which is why scanned text is best kept at 300 DPI or higher when readability matters.
- Why did my text PDF barely shrink when I compressed it?
- Because there was almost nothing to remove. A text only PDF is already small: the words, fonts, and layout take up very little space, and a lossless pass can only remove redundant objects and tighten the structure, typically saving 5 to 15 percent. That is a sign the file was already well made, not a sign the tool failed. If the file still needs to be smaller, the weight is in embedded images, and you need an image recompression step or a split, not a stronger lossless pass.
- What DPI should I use for a PDF I am sending by email?
- 150 DPI is the right default for email and screen reading. It keeps text and photos sharp at normal zoom while cutting the pixel count dramatically compared with a 600 DPI scan: an A4 page drops from about 34.8 million pixels at 600 DPI to about 2.2 million pixels at 150 DPI. Use 300 DPI if the recipient will print the document on a good printer, and keep the original resolution only for archives and high quality print work.
- My scanned PDF is 80 MB. What should I do first?
- Check whether you need every page at full quality. If the recipient only needs part of the document, extract just those pages first, which can solve the problem with zero quality loss. Otherwise, recompress the page images to 150 DPI with a JPEG quality level around 80, working from a copy and keeping the original. A 40 MB scanned file typically lands around 8 to 12 MB after that step. If it still does not fit, split it into parts rather than pushing the quality lower.
- Is it safe to compress a confidential PDF in a browser tool?
- It depends on the tool. With BeYourTools, all PDF processing happens locally in your browser, so a confidential contract, bank statement, or medical form never leaves your device and is never uploaded to a server. That is the property to check for in any tool you use for sensitive documents: client side processing, not a promise to delete uploads later. If a tool uploads your file, treat it as shared, however short the retention promise.
- Why did printing to PDF make my file bigger or lose selectable text?
- Printing re-renders the document instead of editing it. The browser or printer driver draws each page afresh, which can turn selectable text into flat images and discard the efficient structure of the original file. For a text heavy document it sometimes shrinks the file by dropping metadata, but for a scanned or well optimised file it often grows it or damages it. Use a real compressor or splitter, which edits the PDF structure directly, and keep print to PDF for cases where you actually want a fresh rendering.