Two formats with opposite goals
The Library of Congress describes PDF as designed to represent the layout of page-oriented documents. Its imaging model comes from the PostScript page description language. A PDF says, in effect, draw these glyphs in this font at these positions on a page of this size. PDF 1.7 became the international standard ISO 32000-1:2008, and PDF 2.0 followed as ISO 32000-2.
DOCX, the default format of Microsoft Word since Word 2007, was standardized as ECMA-376 in 2006 and approved as ISO/IEC 29500 in 2008. A .docx file is a ZIP package of XML parts. Rename one to .zip and you can open it. The main text sits in word/document.xml as paragraphs, runs and text, with styles, themes, headers, footers and media in separate parts. Nothing in it fixes where a line breaks. The word processor works that out each time it opens the file.
| DOCX | ||
|---|---|---|
| Built to | Show fixed pages the same everywhere | Be edited; layout is worked out by the app |
| Standard | ISO 32000-1:2008, ISO 32000-2 | ECMA-376 (2006), ISO/IEC 29500 (2008) |
| Inside the file | Page descriptions: text, fonts, images, graphics | A ZIP package of XML parts |
| Fonts | Embedded for faithful display | Referenced; display changes if missing |
| Structure | Only if the file is tagged | Paragraphs, runs, styles |
| Reflow | Only tagged PDF | Always; text flows to the page |
Two details from the format descriptions explain most conversion surprises. A PDF needs its fonts embedded to render faithfully, and has a logical structure, with headings and reading order, only if its creator added tags. Many files made by printing to PDF have no tags at all. A DOCX, on the other hand, displays differently when its original fonts are missing, and the Library of Congress notes that a conversion round trip back to an identical document should never be expected.
Word to PDF: what changes
Going from Word to PDF means freezing a layout. The question is whose layout. Word's own Save as PDF uses Word's layout engine, so the PDF keeps Word's line and page breaks. Any other converter reads the document's content and lays it out again. Our Word to PDF tool does the second, in your browser, with this result:
| In the .docx | In the PDF |
|---|---|
| Text, alignment, bold, italic, sizes | Kept, as selectable text |
| Lists with levels | Kept |
| Tables, merged cells | Kept, widths worked out again |
| Pictures | Kept, in the text flow |
| Links, page breaks | Kept |
| Fonts | Replaced by the font you pick |
| Footnotes, endnotes | Listed at the end |
| Headers, footers, columns | Not rebuilt |
The font row matters most. When fonts change, line lengths change, so page breaks move, and a document that ended at the bottom of page 3 can spill onto page 4. That is normal for any converter that is not Word. The text in the resulting PDF is real text with an embedded font, so you can still select, search and copy it, and a screen reader can read it. The tool's fonts cover Latin, Greek and Cyrillic letters; other scripts are replaced with a question mark, and it tells you how many.
PDF to Word: what changes
The other direction is harder. A PDF records where letters sit, not which letters form a paragraph. Our PDF to Word converter reads every piece of text with its position, size and font and rebuilds the structure: lines with normal spacing become paragraphs, larger text becomes a heading, lines that start with a bullet or a number become a list, and words hyphenated at line ends are joined again.
| In the PDF | In the .docx |
|---|---|
| Paragraphs and text | Comes over |
| Headings (three levels) | Rebuilt from text size |
| Bold and italic | Comes over |
| Bullet and numbered lists | Comes over |
| Page breaks | Optional |
| Tables | Lines of text |
| Images | Left out |
| Columns, text boxes, exact positions | Not rebuilt |
| Scanned pages | Need OCR first |
That makes it the right tool for letters, reports, contracts and CVs you want to rework, and the wrong one for a brochure whose layout must stay exactly as it is. The headings use Word's built-in Heading 1 to 3 styles, so the navigation pane and a table of contents work straight away. Tables arrive as lines of text you can paste into a Word table.
A scanned PDF is a special case: each page is an image, so there is no text to convert, and the result is empty. It needs OCR first. The guide how to fix a scanned PDF walks through that.
File size after converting
A PDF made from a text document is mostly fonts and text, which are already compact. Photos and scanned pages are what make PDFs large. That is also why the Compress PDF tool works on images only: it scales photos and scans down to a chosen resolution and saves them again as JPEG, while text, fonts, vector drawings, links and form fields are copied unchanged. A text-only PDF exported from Word hardly shrinks, and the tool says so instead of making the file bigger.
Which one to send
- Send a PDF when the reader must see your pages exactly: applications, forms, invoices, anything to be printed, and documents that should not be casually edited. For long-term archiving, see PDF vs PDF/A.
- Send a .docx when the reader is meant to edit, comment, track changes or reuse the text.
- Keep the source. Edit the Word file and export a fresh PDF, rather than converting a PDF back and forth. Each trip through a converter loses something, and the format descriptions say an identical round trip is not to be expected.
- Check accessibility. A PDF only has headings and reading order if it is tagged. The guide how to make a PDF accessible covers what that takes.
Before you convert
- Decide whether the result is for reading (PDF) or for editing (DOCX).
- From Word to PDF, export from Word itself if the layout must match line for line.
- From PDF to Word, check whether the PDF has text: try selecting a word. If you cannot, it is a scan.
- Expect to fix columns, tables and text boxes by hand after PDF to Word.
- Compress only PDFs with photos or scans; text-only files are already small.
Frequently asked questions
Why does my Word file look different after converting from PDF?
A PDF stores where each letter sits on the page, not which words form a paragraph or a heading. A converter has to guess the structure back from positions and sizes. Simple documents come back well; columns, tables and text boxes are where the guesses fail.
Will a PDF made from Word look exactly like the Word file?
Only if the program that lays out the pages is Word itself. Any other converter, including ours, reads the document's content and lays out the pages again, and fonts that are not available get replaced, so lines and page breaks land in other places. When the PDF must match Word line for line, save it as PDF from Word.
Can I edit a PDF directly instead of converting it?
You can change some things, like adding text, filling form fields or reordering pages, but the PDF is built to describe finished pages. Reflowing paragraphs after a big edit is what a word processor does well and a PDF does not. For real rewriting, convert to Word, edit, and export a new PDF.
Which format should I send?
Send a PDF when the reader should see exactly your pages, for example for forms, applications, invoices and anything that will be printed. Send a .docx when the reader is meant to edit, comment with tracked changes or reuse the text. If you are unsure, ask which one the recipient needs.
Why is my PDF to Word file empty?
The PDF is probably a scan, where every page is a picture of text. There is no text for the converter to read. It needs optical character recognition first, which turns the pictures back into text.
Sources
- Library of Congress, Sustainability of Digital Formats: PDF (Portable Document Format) Family (Adobe origin; PostScript imaging model; designed to represent the layout of page-oriented documents; ISO 32000-1:2008 and ISO 32000-2; embedded fonts for faithful rendering; logical structure only with tagging; tagged PDF and reflow; many print-to-PDF files untagged; printing and text extraction can be disallowed)
- Library of Congress, Sustainability of Digital Formats: DOCX Transitional (Office Open XML), ISO/IEC 29500 and ECMA-376 (default in Word since 2007; ECMA-376 in 2006, ISO/IEC 29500 in 2008; ZIP package with Open Packaging Conventions; document.xml with paragraphs, runs and text; styles, headers and footers as parts; display differs without the original fonts; no identical round trip in conversion)