How to Merge, Split and Compress PDFs Without Losing Quality

What is inside a PDF file

A PDF is not a picture of a page. It is a structured container made of numbered objects. Some objects describe pages, some hold the drawing instructions for each page, some hold fonts, and some hold images. A cross-reference table at the end of the file records where each object starts, and a trailer points to the root of the document, called the catalog, which in turn points to a page tree listing every page in order.

Each page has a content stream, a compact list of instructions such as move here, select this font at this size, draw this text, or paint this image in this rectangle. Text drawn this way stays as real characters, so you can search, select and copy it. A scanned document is different: each page is usually one large image, and unless optical character recognition has been applied, there is no text to search at all. Knowing which kind of PDF you have explains most of what happens when you edit or compress it.

Because objects can be shared, two pages can point to the same font or the same logo image. This is one reason PDFs can be surprisingly small, and it also matters when files are merged.

How merging works

Merging creates a new document and copies the page objects from each source file into it, along with everything those pages depend on, such as fonts, images and annotations. The new page tree lists the pages in the order you choose. Done properly, merging is lossless: no text is re-rendered and no image is recompressed, so the merged file looks exactly like the originals.

A few things are worth checking before you merge. Decide the file order first, since most tools merge in the order the files are listed. Remember that each source may carry its own copy of a font or logo, so the merged file can be larger than you expect. Interactive form fields can clash if two files use the same field names, and bookmarks or an outline from the originals may not carry across. Password-protected files must be unlocked by someone who knows the password before their pages can be copied.

Splitting, extracting and fixing page order

Splitting is merging in reverse. You choose which pages go into each output file, and the tool copies just those pages and their resources into new documents. Common patterns are splitting every page into its own file, extracting a range such as pages 5 to 12, or breaking a long file into chapters. Page ranges are normally written with commas and hyphens, so 1,3,5-7 means pages 1, 3, 5, 6 and 7.

Page order problems are common with scans, where a double-sided document may come out as all the odd pages followed by the even pages in reverse. Reordering simply rewrites the page tree, so it is also lossless. Rotation is similar: a PDF stores rotation as a page setting in multiples of 90 degrees, so turning a sideways scan upright does not degrade it. Adding page numbers, by contrast, draws new text onto each page, which is a small but real change to the content.

A good habit is to check the page count after every split or merge. If you started with three files of 4, 10 and 6 pages, the merged file should have 20 pages, and extracting pages 5 to 12 from it should produce exactly 8.

What compression can and cannot do

Compression is where expectations most often go wrong. Text and vector drawings are already stored as compact instructions, usually compressed with the lossless Flate algorithm, so a text-only PDF has little to lose. A 40 page report with no images might be 300 KB, and no honest tool will shrink it to 30 KB without destroying it. Lossless structural optimization, which removes unused objects, packs objects more tightly and recompresses streams, often saves only a few percent on files like this, and sometimes nothing at all.

Images are where the size lives. A single phone photo or a page scanned at 300 DPI in color can be several megabytes. Real reductions come from lowering image resolution and applying lossy JPEG compression. The arithmetic is straightforward: a letter-size page is 8.5 by 11 inches, so at 300 DPI it holds 2,550 by 3,300 pixels, about 8.4 million pixels. At 150 DPI it holds 1,275 by 1,650, about 2.1 million pixels, one quarter as many. Halving the resolution therefore removes roughly 75 percent of the image data before JPEG compression even starts.

The trade-off is quality. Around 96 DPI is fine for viewing on screen but small text can become fuzzy, roughly 150 DPI is a reasonable balance for most documents, and 200 DPI or more is safer for printing. Some compression methods also flatten each page into an image, which removes the text layer, so the result can no longer be searched or copied and screen readers cannot read it. Always open the compressed file, zoom in on the smallest text, and compare the new file size to the original, because recompressing an already optimized file can even make it larger.

A worked example

Imagine you need to submit a job application through a portal with a 5 MB limit. You have a two page cover letter exported from a word processor (120 KB), a three page resume (180 KB), and two scanned certificates at 300 DPI in color, each one page and 3.2 MB. Merged in that order, the file has 2 + 3 + 1 + 1 = 7 pages and is about 6.7 MB, over the limit.

Compressing the whole file with a lossless structural setting will barely touch it, because nearly all of the size is in the two scans. Instead, apply a raster setting of about 144 DPI to just the certificate pages, or recompress them before merging. Dropping from 300 to 144 DPI keeps about (144 divided by 300) squared, or 23 percent, of the pixels, and JPEG compression typically shrinks them further, so each certificate is likely to land well under 1 MB. Keep the letter and resume as they are, so their text stays sharp and searchable, then merge. The final file should come in comfortably under 5 MB with seven pages in the right order.

Privacy and local processing

Many online PDF services work by uploading your file to a remote server, processing it there, and sending back the result. For a restaurant menu that hardly matters, but PDFs often contain bank statements, medical records, identity documents, tax forms, and signed contracts. Once a file is uploaded, you are relying on the service's retention policy and security.

Processing in the browser avoids that. The document is read and rewritten on your own device, and the bytes never need to leave it. It is still sensible to work on a copy, keep the original until you have checked the output, and remember that hiding content is not the same as removing it. Drawing a black box over text in some editors leaves the text underneath, so use a proper redaction feature and check the result before sharing anything sensitive.

Doing it with ToolWorld PDF tools

ToolWorld PDF Tools runs its operations locally in your browser, so document bytes stay on your device. It can merge files in the order they are displayed, split a document into single pages or into range groups separated by semicolons, and compress using either lossless structure optimization or raster presets for screen quality at 96 DPI, balanced at 144 DPI, or print at 200 DPI. It warns that raster compression removes the text layer and that compression is not guaranteed, which matches the limits described above. The same tool also handles watermarks, password protection and unlocking, metadata, image to PDF conversion, OCR and redaction.

For page-level fixes, ToolWorld PDF Page Organizer lets you reorder pages with move up and move down controls, remove pages, rotate all pages, and add page numbers before downloading the result. If you need to edit the words themselves rather than the pages, the PDF to Word Converter turns text-based PDFs into editable DOCX files, with page selection and a preview of the extracted text. Scanned PDFs have no text layer, so run OCR first if you want editable text from a scan.