PDF won't get smaller: causes and the right next steps

Unchanged, larger or target not reached?

Before you look for causes, take a close look at the result. After processing, the compressor shows Original, Result and Saved, each as a file size in bytes, as well as the savings as a percentage. Compare these byte values rather than the rounded MB figure in your file manager, because rounding can make a small difference disappear.

Note the difference between the first two rows: successful processing does not automatically mean that your target size has been reached. If you entered a limit under Set target size, the result shows you separately whether it was met.

Both results miss the target, but for different reasons. In Example A, this attempt could not gain anything more. In Example B, 300,000 bytes are still missing. The next steps differ accordingly.

Why a result can even be larger

When a PDF is rewritten, the program rearranges objects, cross-references and data streams. If a file was already stored very compactly, this rebuild can take up more space than before. The same can happen with an image that is already heavily compressed and does not get smaller when re-encoded. Virtually printing to a new PDF can also add extra bloat or change features such as links and form fields. The compressor never gives you a larger version. If no attempt is strictly smaller and error-free, you keep the original with 0 bytes saved.

If the file can't be selected at all

The file selection accepts PDFs up to 50 MiB, which is up to 52,428,800 bytes per file. If the Select PDF button does not accept your file, read the message that appears and compare the size in bytes with this limit. The message shows whether the file is too large or whether there is another problem, such as a file extension that doesn't match. Neither is a compression result. Both are limits of the file selection. If the file can be selected but not processed, again follow the error message that appears. In that case, the PDF itself may be the cause, for example if it is encrypted or damaged. However, processing can also fail or be aborted for other reasons. In every case, your original file remains unchanged.

Which content limits the reduction?

How much can still be saved depends mainly on what your PDF consists of. The compressor uses QPDF directly in your browser and works in two ways. It stores the internal structure more compactly, for example through object streams and Flate compression. It can also re-encode suitable embedded images as JPEG. The documentation under QPDF: Optimizing File Size describes how these methods work in principle.

Why a stronger method doesn't always make the file smaller

Balanced and Smaller file also start with the lossless structural optimization. Only then do they re-encode suitable images at a lower JPEG quality. If your PDF contains no such images, or if re-encoding does not make an already efficiently encoded image smaller, every compression mode ends with the same result. In that case, Smaller file shows the same byte count as Lossless optimization.

In addition, the compressor does not reduce the resolution. A scan with a very large number of pixels keeps that pixel count, and only the encoding of the image data changes. If a scan remains too large for this reason, the scanning process itself offers more room for reduction, as described below.

Rejected doesn't mean hard to compress

A rejected file is not a result with 0 bytes saved; it was not processed at all. This applies to digitally signed or certified PDFs, encrypted or password-protected PDFs, and files whose structure is damaged or that trigger warnings during the check. In this case, it is still unclear whether the content could be reduced well. The frequently asked questions at the end of this page explain how to get a version that can be processed.

Work in a targeted way instead of compressing again and again

If the PDF stays the same size after several attempts, another run with the same file and the same method won't change anything. If, on the other hand, you compress a result that has already been reduced lossily, the file may shrink further. But if the images are re-encoded as JPEG again, more detail is lost with every pass. That is why the compressor starts each internal attempt from the original and never from the previous JPEG result. Work the same way: click Process new file, select your original file again and choose a different level under Method. That way you can always try again with the original.

When a target size is set, the compressor also tries lower JPEG quality levels in Balanced and Smaller file and stops as soon as the target is reached. Smaller file goes the furthest. If that doesn't help either, consider the following options. Always keep the original, because each of these steps creates a new file.

Remove only pages that are genuinely unnecessary

The PDF may contain pages that the recipient doesn't need, such as blank back pages or sheets that were scanned twice. You can remove unnecessary pages using the page selection. Required evidence and attachments must stay in the document, even if they take up a lot of space. Removing pages doesn't save a fixed amount. A blank page takes up little space, while a photo scan takes up a lot, and the result does not get smaller in every case. The tools don't pass files to each other automatically. So save the result and check its size in bytes. If it is no more than 52,428,800 bytes (50 MiB), you can select the saved file in the compressor again.

Split the file if the recipient accepts multiple files

Splitting only helps if the recipient explicitly accepts multiple files. In that case, each individual file must be below the per-file limit. At the same time, the total size and the number of files must not exceed the recipient's other requirements. The parts may be larger together than the original, because shared components such as fonts or images are stored again in several parts. If that fits the requirements, you can split the PDF and check each part individually.

Recreate the source

With scans and photos, the biggest savings are often possible before compression. Rescan the document with a suitable resolution and color setting, or export it again from the program in which it was created. The compressor changes neither resolution nor color mode. The guide reduce a scanned PDF explains how to balance legibility and file size. Some specialized scanning software offers its own methods that treat text and image parts separately. That is a separate approach with its own limits and is not part of this compressor.

Check the recipient's limit correctly

Only the recipient's requirements decide whether your file is small enough. Read carefully what the recipient requires and check each limit separately:

• Size per file
• Total size of all files in an upload or message
• Maximum number of files
• Page count per document
• Unit: Does MB mean 1,000,000 bytes or 1,048,576 bytes, i.e. MiB?

These limits are independent of each other. A file can meet the size limit and still have too many pages, and compression does not change the page count. With the calculator, you can check all upload limits. It reads the file size of every selected file. For readable PDFs up to 50 MiB, it also reads the page count. If the page count remains unknown, for example with password-protected or damaged files, a page limit you have set does not count as met. The calculator still checks the file size.

Why the MB display can be misleading

In the compressor, a target size of 2 MB means no more than 2,000,000 bytes. Right at limits like this, a rounded display can easily mislead you, as this worked example shows:

Programs that count 1,048,576 bytes per MB show the same file as about 1.96 MB, even though it exceeds a decimal 2 MB limit. So rely on the byte values in the result and find out which unit the recipient means. The page Reduce PDF to a target size shows how to set a limit in the tool.

Can every large file be reduced to under 2 MB?

No, there is no guarantee. The target size is an upper limit that the compressor works toward. It is not enforced, and the compressor does not pad the file to exactly this size. Work out in advance how much would need to be saved. For a file of 3,000,000 bytes to be no more than 2,000,000 bytes, you need to save at least 1,000,000 bytes, which is one third of the file (33.3 %). If the recipient explicitly requires less than 2,000,000 bytes, you need to save at least one byte more. That much is hardly realistic for a mostly text-based PDF that is already stored compactly. It may work with photos or scans, but in that case, be sure to check legibility.

When only poor quality reaches the limit

Sometimes only the strongest level reaches the limit, and then fine print, numbers or stamps are barely recognizable. In this case, legibility comes first. A document that the recipient can't read doesn't serve its purpose, even if it is below the limit. So check the saved result in your PDF viewer:

• Are all numbers, dates and small text legible at normal magnification?
• Are stamps, seals and handwritten signatures still clearly recognizable?
• Can text still be searched and highlighted if that was possible in the original?
• Can colored highlights still be told apart if the recipient needs them?

In the compressor, you can compare the same page in the original and the result and page through the comparison pages. However, a preview doesn't show whether text is searchable or links work. For that, you need the saved file. The guide Compress PDF without quality loss explains in more detail how to reliably spot quality loss.

Ask the recipient about a permitted alternative

If the legible version is still above the limit, keep it and contact the office that requires the file. Ask which alternative is permitted. For example, you might be allowed to spread the documents across several files or submit them later by another route. Many offices explicitly emphasize completeness and legibility. uni-assist, for example, requires complete, clearly legible scans in its upload procedure for application documents (checked on 13 October 2026). Only the recipient can tell you which option is available in your case.

Frequently asked questions about rejected PDFs

Why was a signed or protected PDF rejected?

A digital signature or certification is a cryptographic property of the file. Encryption protects the content with a password or permissions. The compressor rewrites the PDF when reducing it. Because it can't be reliably ruled out that such a rebuild affects the signature or the protection, the compressor does not process these files as a matter of principle. If you need a smaller version, ask the issuer or an authorized party for an unsigned or unencrypted version that you are allowed to reduce. Alternatively, find out whether the recipient will accept the file at its current size. For damaged files, exporting the document again from the program it was created in can help. A visible handwritten signature or stamp on a scan, by contrast, is part of the image and does not lead to rejection.

Can I still process a rejected PDF with Split or Delete pages?

No. These tools reassemble the pages and are even stricter than the compressor for some documents: they do not support signed, certified or encrypted PDFs, files with form fields, PDF portfolios or embedded files, or optional layers, and they can process no more than 250 pages. For a rejected file, the way forward is therefore to ask the issuing party for a version that can be processed.