How PDFs Actually Work - A Non-Technical Guide

A format that just works (and why that is remarkable)

You have probably opened thousands of PDFs in your life without thinking about what happens behind the scenes. You double-click a file, and the document appears - perfectly formatted, with every font, image, and layout element exactly where the creator put it. It looks the same on your Mac as it does on your colleague's Windows PC, your friend's phone, or a library computer running Linux.

This consistency is so reliable that it is easy to take for granted. But if you have ever emailed a Word document and had the recipient complain that the formatting was wrong, or opened a web page that looked different in two browsers, you know that cross-platform consistency is actually very hard to achieve. PDF achieves it so well that we barely notice.

Understanding how PDFs work under the hood gives you useful context for everyday tasks. It explains why PDFs are harder to edit than Word documents, why some PDFs have selectable text while others do not, why certain PDFs are tiny and others are enormous, and why the format has survived for over 30 years while many competitors have come and gone.

This guide explains PDF internals in plain English. No programming knowledge is required. By the end, you will understand what is happening inside every PDF you open.

A brief history of the PDF

The PDF was created by a team at Adobe Systems, led by co-founder John Warnock. In 1991, Warnock wrote an internal paper called "The Camelot Project" that outlined a vision: a universal file format that could capture documents from any application, send them electronically, and display and print them on any machine. At the time, sharing documents between different computers was a nightmare of incompatible formats, missing fonts, and broken layouts.

Adobe released the first version of PDF (PDF 1.0) in 1993, alongside a viewer called Acrobat Reader. The early years were slow going. Creating PDFs required expensive Adobe software, and the format competed with alternatives like PostScript files and proprietary word processor formats. Many people saw it as overly complex for what it did.

Two decisions changed everything. First, Adobe made Acrobat Reader free in 1994, ensuring anyone could view PDFs without paying. Second, and more significantly, Adobe published the PDF specification and eventually submitted it to the International Organization for Standardization (ISO). PDF became an open ISO standard (ISO 32000) in 2008. This meant anyone could create software that reads and writes PDFs without needing a licence from Adobe.

Today, PDF is the dominant format for document exchange worldwide. Tax agencies, courts, universities, businesses, and governments all rely on it. The format has evolved through multiple versions (the current standard is PDF 2.0, published in 2017 and updated in 2020), but the core architecture has remained remarkably stable since the beginning.

The four parts of every PDF

Every PDF file, from a one-page receipt to a thousand-page technical manual, contains four structural sections. Think of them as the building blocks that every PDF viewer reads to reconstruct your document on screen.

1. The header

The header is the first line of the file. It is short - just a handful of characters - and serves two purposes. It identifies the file as a PDF (so your computer and applications know what they are dealing with), and it specifies which version of the PDF standard the file follows.

If you were to open a PDF in a plain text editor, the very first thing you would see is something like "%PDF-1.7" or "%PDF-2.0". The percent sign is part of the PDF syntax, "PDF" identifies the format, and the number indicates the version. This header is how your Mac's Finder, web browsers, and email applications recognise that a file is a PDF and decide to open it with the appropriate viewer.

2. The body

The body is where all the actual content lives, and it makes up the vast majority of the file's size. Everything you see when viewing a PDF - text, images, fonts, colours, page layouts, form fields, annotations, bookmarks - is stored in the body as a collection of objects.

Each object has a unique number (like an ID badge) and contains a specific piece of the document. One object might hold the text for page 5. Another might contain a JPEG image. Another might describe the font used for headings. Another might define the dimensions and margins of page 12. The body can contain hundreds or thousands of these objects, depending on the complexity of the document.

The key thing about these objects is that they can appear in any order within the body. The object for page 50's text might be stored before the object for page 1's text. This does not matter because the PDF viewer does not read the body sequentially - it uses the cross-reference table (described next) to find each object by its ID number.

3. The cross-reference table

The cross-reference table (often called the "xref table") is like an index at the back of a book. It maps each object's ID number to its exact byte position within the file. When the PDF viewer needs to display page 3, it looks up the IDs of the objects that make up page 3 in the cross-reference table, finds their byte positions, and jumps directly to those locations in the file to read the data.

Without the cross-reference table, the viewer would have to scan through the entire file to find each object - like searching for a chapter in a book that has no table of contents. The xref table makes PDF viewing fast and efficient, even for very large documents.

4. The trailer

The trailer sits at the very end of the file and serves as the starting point for reading the document. It tells the viewer two critical things: where to find the cross-reference table, and which object is the document's root (the top-level object from which the entire document structure branches out).

PDF viewers actually read files from the end first. They open the file, jump to the trailer, find the location of the cross-reference table, read the table to build an index of all objects, and then start rendering pages by looking up the objects they need. This end-first approach is why PDFs can be read efficiently regardless of size - the viewer always knows exactly where to start.

How text works in a PDF

Text in a PDF is more complex than it might seem. Unlike a Word document, where text is stored as flowing paragraphs that reflow to fit the window, PDF text is positioned with exact coordinates on a fixed-size page.

Positioned characters, not flowing text

When a PDF stores text, it does not simply say "here is a paragraph." Instead, it says something closer to "place the character H at position (72, 720) on the page, using font Helvetica at 12 points, then place the character e at position (79, 720), then l at (85, 720)..." and so on for every character in the document.

This exact positioning is what makes PDFs look identical everywhere. The viewer does not need to decide where to place text - it is told exactly where every character goes. There is no reflowing, no line-break calculation, no widow-and-orphan adjustment. Everything is predetermined by the PDF creator.

This is also why editing text in a PDF is fundamentally different from editing in a word processor. In Word, you type text and the application figures out where to put it. In a PDF, every character has a fixed position. If you delete a word from the middle of a paragraph, the remaining text does not automatically reflow to fill the gap. A PDF editing tool has to recalculate positions for every subsequent character, which is significantly more complex than editing flowing text.

Fonts inside the file

When you create a PDF, the fonts used in the document can be embedded directly in the file. This is one of the most important features of the format. The PDF carries its own fonts, so it does not depend on which fonts are installed on the viewer's computer.

Think about what happens with a Word document. If you use the font "Garamond" in your document and send it to someone who does not have Garamond installed, their word processor substitutes a different font. The substitution changes how the document looks - different character widths, different line heights, different overall appearance. Your carefully designed layout falls apart.

A PDF avoids this problem entirely. When the PDF is created, the font data - the actual shape definitions for every character used - is embedded inside the PDF file. When someone opens the PDF on a computer that does not have Garamond installed, it does not matter. The PDF viewer reads the embedded font data and renders the characters exactly as they were designed. The document looks identical regardless of which fonts are available on the system.

Font subsetting

A complete font file can be quite large - some fonts contain thousands of character shapes (called glyphs) covering multiple languages and special symbols. If a PDF embeds the full font file for every font used, the file size can grow significantly.

To address this, PDF supports font subsetting. Instead of embedding the entire font, the PDF creator embeds only the specific characters (glyphs) that are actually used in the document. If your document uses the letters A through Z, the numbers 0 through 9, and a handful of punctuation marks in Helvetica, the PDF embeds only those 70 or so glyphs rather than the full Helvetica font with its thousands of glyphs. This dramatically reduces file size while still ensuring perfect rendering.

Font subsetting is why some PDFs are surprisingly small despite using beautiful, complex typefaces, and it is also why you sometimes cannot add new text in a specific font when editing a PDF - the font subset embedded in the file might not include the characters you need to type.

How images work in a PDF

Images in PDFs are embedded directly in the file, just like fonts. When you see a photograph, diagram, chart, or illustration in a PDF, that image data is stored as one of the objects in the PDF's body.

Image formats inside PDFs

PDFs can contain images in several formats. The most common are:

• JPEG: Used for photographs and complex images with many colours. JPEG compression is lossy, meaning some quality is sacrificed for smaller file size. This is the most common image format found inside PDFs.
• JPEG2000: A more advanced version of JPEG that offers better quality at smaller sizes, but is less widely used. Some high-quality print PDFs use JPEG2000.
• Flate (PNG-style compression): Used for images that need lossless compression - diagrams, screenshots, images with text, or any image where quality loss is unacceptable. The image data is compressed without losing any information.
• CCITT (fax compression): Used for black-and-white images, particularly scanned documents. This is a highly efficient compression method specifically designed for monochrome images like scanned text pages.

Resolution and quality

The quality of images in a PDF depends on the resolution and compression settings used when the PDF was created. A PDF created for screen viewing might use 72 to 150 DPI (dots per inch) images, resulting in a small file size. A PDF created for professional printing might use 300 DPI or higher images, producing much larger files but sharper prints.

This is why compressing a PDF can dramatically reduce its file size. PDF compression tools like Paradise PDF, Smallpdf, and others work primarily by reducing image resolution and increasing compression. A 50 MB PDF might shrink to 5 MB by downsampling images from 300 DPI to 150 DPI. The text and fonts are barely affected - they take up very little space. It is almost always the images that account for large PDF file sizes.

Vector graphics vs raster images

PDFs can also contain vector graphics - shapes, lines, curves, and paths defined by mathematical equations rather than pixels. Logos, diagrams, charts, and illustrations are often stored as vector graphics in PDFs. The advantage of vectors is that they look sharp at any zoom level or print resolution, because they are redrawn from mathematical definitions rather than scaled from a fixed grid of pixels.

When you zoom in on a photograph in a PDF, it eventually becomes pixelated - you are seeing the individual pixels blown up. When you zoom in on a vector graphic (like a company logo or a chart), it stays perfectly sharp because the viewer recalculates the shapes at the new zoom level. This is why well-made PDFs often use vector graphics for logos, icons, and diagrams.

Why PDFs look the same everywhere

The combination of exact text positioning, embedded fonts, and embedded images is what gives PDFs their signature characteristic: they look the same on every device. Here is a summary of why each element contributes:

This self-contained nature is the fundamental difference between PDF and formats like HTML or Word documents. A web page depends on external CSS files, JavaScript files, font files hosted on servers, and the browser's rendering engine. A Word document depends on the word processor's version, the system's installed fonts, and the printer driver's configuration. A PDF depends on nothing external. Everything it needs to render correctly is packed inside the file itself.

Scanned PDFs vs text-based PDFs

Not all PDFs are created equal, and the difference between a scanned PDF and a text-based PDF is one of the most important things to understand.

Text-based PDFs

A text-based PDF is created digitally - exported from Word, Pages, a web browser, a design application, or any software that generates PDF output. The text in a text-based PDF is stored as actual characters with positioning data, as described in the text section above. You can select text, copy it, search for words, and screen readers can read the content aloud.

Scanned PDFs

A scanned PDF is created by scanning a physical page with a scanner or a phone camera. The resulting PDF contains an image of the page - a photograph of text, rather than actual text characters. To the PDF viewer, each page is just a picture. You cannot select individual words, you cannot search for text, you cannot copy and paste content, and screen readers cannot read it. As far as the PDF is concerned, the page is an image of text, not text itself.

This distinction matters practically. If you receive a scanned PDF and need to extract text from it, you need OCR (optical character recognition) - software that analyses the image, recognises the letter shapes, and converts them into actual text characters. OCR is built into many PDF tools, including Paradise PDF, and can transform a scanned PDF from an image-based document into a searchable, copyable, text-based document.

How to tell the difference

The easiest way to check: try to select text. Open the PDF and click and drag across some text. If you can highlight individual words and sentences, it is a text-based PDF. If the entire page gets selected as one block (or nothing gets selected at all), it is a scanned PDF containing images.

Another indicator: zoom in closely on the text. In a text-based PDF, the text stays sharp at any zoom level because it is rendered from font data. In a scanned PDF, the text becomes pixelated when you zoom in because you are looking at an image.

Layers of a PDF - what sits on top of what

A PDF page can have multiple layers of content stacked on top of each other, like transparent sheets laid over one another. Understanding these layers helps explain how annotations, form fields, signatures, and watermarks work.

The content layer

The base layer is the original page content - the text, images, and graphics placed by the document creator. This is what you see as the "document" itself. In most PDFs, this layer is the only one that exists.

The annotation layer

Annotations - comments, highlights, sticky notes, stamps, and drawn markups - sit in a separate layer on top of the content. This is why annotations can be added and removed without changing the underlying document. When you highlight text in a PDF viewer, the yellow highlight is an annotation object that floats above the text, not a modification to the text itself.

Some PDF operations "flatten" annotations, which permanently merges them into the content layer. After flattening, the annotations become part of the page content and can no longer be independently selected, moved, or deleted. This is commonly done when preparing a final version of a document for distribution.

The form layer

Interactive form fields - text boxes, checkboxes, radio buttons, dropdown menus - exist as interactive objects that sit on top of the content layer. When you fill in a PDF form, you are entering data into these interactive objects. The form fields know their type (text, checkbox, etc.), their position on the page, their formatting rules, and in some cases, validation rules that check your input.

Signatures

Electronic signatures and digital signatures are typically applied as annotation objects or as specialised signature form fields. An electronic signature (a picture of your handwriting) is usually an image placed in the annotation layer. A digital signature (a cryptographic certificate-based signature) is a form field that contains both a visible representation and the cryptographic data that proves authenticity.

Why editing PDFs is harder than editing Word documents

Now that you understand how PDFs store text with exact positioning and embedded fonts, it should make more sense why editing a PDF is fundamentally different from editing a Word document.

In Word, text flows like water. Insert a word in the middle of a paragraph and everything after it shifts to make room. Delete a sentence and the remaining text closes the gap. The word processor handles all the positioning automatically because it treats text as a flowing stream that adapts to the page dimensions.

In a PDF, every character is nailed to a specific position on the page. There is no concept of "flowing" text. If you delete a word from the middle of a line, the characters after it do not automatically slide left to fill the space. A PDF editing tool has to handle this explicitly - recalculating the position of every subsequent character, potentially adjusting line breaks, and dealing with the fact that the rest of the paragraph might need to be repositioned.

Simple edits - fixing a typo, changing a number, updating a date - work well in PDF editors because the replacement text is close enough in size to the original that repositioning is minimal. Major edits - rewriting paragraphs, adding new sections, restructuring the layout - are much harder because they require extensive repositioning of all subsequent content.

This is why the standard advice is to edit the source document (the Word file, the design file, the original) whenever possible, and then export a new PDF. The PDF is the final output, not the working format. If you only have the PDF and need to make significant edits, PDF editing tools can handle it, but expect the process to be slower and more manual than editing in a word processor.

The PDF specification - an open standard

One of the reasons PDF has thrived for over 30 years is that it became an open standard. The PDF specification - the technical document that describes exactly how PDF files must be structured, how objects are defined, how encryption works, how fonts are embedded, and every other detail - is published by ISO as ISO 32000.

This means anyone can create software that reads and writes PDFs without needing permission from or paying royalties to Adobe (or anyone else). This openness is why there are hundreds of PDF viewers, editors, and creation tools available across every platform. Preview on Mac, browsers like Chrome and Safari, tools like PDF Expert, Smallpdf, iLovePDF, Sejda, and Paradise PDF - they all work because the specification is public and anyone can implement it.

The current version of the standard is PDF 2.0 (ISO 32000-2:2020), which defines the format's capabilities including encryption, digital signatures, 3D content, multimedia, accessibility features, and more. Specialised subsets of the standard serve specific purposes: PDF/A for archiving, PDF/X for print production, PDF/E for engineering documents, and PDF/UA for universal accessibility.

Key takeaways

• A PDF is a self-contained package that carries everything needed to display the document - text with exact positioning, embedded fonts, embedded images, and layout information
• Every PDF has four sections: header, body, cross-reference table, and trailer. The viewer reads from the end first to find the index, then uses the index to find content
• Text is positioned with exact coordinates, which is why PDFs look identical everywhere but are harder to edit than Word documents
• Fonts are embedded inside the file, often as subsets containing only the characters actually used, so the document does not depend on which fonts are installed on the viewer's computer
• Images are embedded and can use different compression formats depending on whether quality or file size is the priority. Images typically account for most of a PDF's file size
• Scanned PDFs contain images of text, not actual text. OCR is needed to make them searchable and copyable
• PDF is an open ISO standard that anyone can implement, which is why the format is supported everywhere and has lasted over 30 years

Related reading

• what is PDF/A
• what is PDF linearization
• view PDF metadata