# Why Is My PDF So Large? Find What Takes the Space

- Author: FastBankConvert (https://fastbankconvert.com)
- Published: 2026-10-17
- Category: Speed, size and limits
- Canonical: https://fastbankconvert.com/blog/why-is-my-pdf-so-large

> Why is my PDF so large when it is only a few pages? Embedded images, fonts, scans in color and repeated content, and how to see which one it is.
Your PDF is so large because of what is embedded in it, not its page count. Page images from a scanner or phone are the usual cause, especially in colour. Fully embedded fonts, a logo stored again on every page, attachments and saved edit history add the rest. Listing the file's images and fonts shows which one it is.

## Why is a PDF so large when it has only a few pages?

A PDF's size comes from the objects stored inside it. Text is cheap: a byte or two per character plus a font reference. Images are expensive: every pixel is stored. So a five-page file of plain text may be under 100 KB, while five scanned pages of the same document can be several megabytes or more.

Think of a page as a list of drawing instructions. "Put these characters here in this font" is short. "Paint this photograph, 2,500 pixels wide, here" carries the whole photograph. Once you see a PDF that way, a large file with few pages stops being a mystery. Somewhere in it is a picture, a font or a stored file that is bigger than all the text put together.

The question to ask, then, is not how to shrink the file in general but which object is heavy. Each cause has a different fix, and the wrong fix can damage the file. Recompressing images does nothing for a file that is heavy because of an attachment, and removing fonts from a text PDF can break how it displays.

## Colour scans and photos as the weight in a large PDF

Page images are the most common reason a PDF is large. A scanned page is one picture covering the whole sheet, and its size grows with resolution and colour depth. A colour scan stores three values per pixel where greyscale stores one, so scanning a black-and-white statement in colour roughly triples the raw data for no gain in readability.

Resolution adds to this quickly. Doubling the dots per inch quadruples the pixels, because it doubles both width and height. A scanner set to a high resolution "to be safe" produces pages that are far heavier than text recognition needs. Phone photos are similar: a modern camera captures a large image, and saving it into a PDF without resizing carries all those pixels along.

Compression type matters as well. JPEG keeps photos small but handles sharp text edges poorly. Formats designed for black-and-white pages, such as the fax-style CCITT and JBIG2 encodings, store text scans very compactly. A file that holds colour JPEGs of what are really black-and-white pages is the classic oversized scan. For ways to fix one without hurting text recognition, see [what to do when a scanned PDF is too large](/blog/scanned-pdf-too-large).

## Fully embedded fonts in a PDF file

Fonts can make a PDF large when they are embedded whole rather than as subsets. An embedded font is a copy of the font file stored inside the PDF so text displays the same everywhere. A subset keeps only the characters used. A full embed of a large font can outweigh the text itself.

Most PDFs from banks and accounting software embed subsets, which keeps fonts small. Problems tend to come from documents assembled by hand. A file merged from many sources can carry the same font several times, once per source document, because each brought its own copy. A document exported with an "embed all fonts fully" setting can also grow quickly when it uses several typefaces.

Fonts rarely explain a file of tens of megabytes on their own. They are more often the reason a text-only PDF is a few megabytes when you expected a few hundred kilobytes. If the images check out as small, fonts are the next place to look.

## Repeated logos, attachments and edit history inside a PDF

Some of a PDF's weight never appears on the page. A logo can be stored once and reused, or stored again on every page. Files can carry attachments. Some editors save changes by appending them to the end of the file, so earlier versions stay inside. Each of these adds size you cannot see by scrolling through the document.

Look for these hidden sources:

- **Repeated images.** A statement generator that stores the bank logo separately on each page multiplies its size by the page count. A sharp, high-resolution logo repeated on 50 pages adds up.
- **Attachments.** PDFs can contain embedded files, such as a spreadsheet or another PDF. Most viewers show them in an attachments or paperclip panel.
- **Edit history.** Some tools write changes as additions at the end of the file. A document edited and saved many times can carry its earlier states. Saving it as a new file, rather than saving over it, usually drops the old versions.
- **Hidden layers and form data.** Optional content layers and filled form fields can add objects that are not visible in the default view.
- **Merged duplicates.** Combining PDFs can leave the same page or image in the file twice.

Of these, repeated images and merged duplicates are the most common in statements. Attachments are rare in bank PDFs but common in documents passed around by email.

## How can you see what takes the space in a PDF?

Free command-line tools from the Poppler project list a PDF's images and fonts with their sizes. `pdfimages -list` prints every image, its page, resolution and stored size. `pdffonts` prints every font and whether it is embedded or a subset. Between them they explain most oversized files in a minute or two.

The [pdfimages manual page](https://manpages.debian.org/testing/poppler-utils/pdfimages.1.en.html) describes the listing columns. The useful ones for this job are page, width and height in pixels, color, enc (the encoding, such as jpeg or ccitt), x-ppi and y-ppi (the resolution as placed on the page) and size (the stored size). Here is an example listing for a made-up four-page file, shortened to the useful columns:

| page | width | height | color | enc | x-ppi | size |
| --- | --- | --- | --- | --- | --- | --- |
| 1 | 5100 | 6600 | rgb | jpeg | 600 | 4.1M |
| 2 | 5100 | 6600 | rgb | jpeg | 600 | 3.9M |
| 3 | 5100 | 6600 | rgb | jpeg | 600 | 4.0M |
| 4 | 5100 | 6600 | rgb | jpeg | 600 | 3.8M |

Reading it line by line: every page is one full-page image, so this is a scan. Each is in rgb colour at 600 ppi, stored as jpeg, at about 4 MB per page. That one look tells you the file is large because it was scanned in colour at a high resolution, and that rescanning in greyscale at a moderate resolution would shrink it most.

The [pdffonts manual page](https://manpages.debian.org/testing/poppler-utils/pdffonts.1.en.html) explains its columns: emb says whether a font is embedded, sub whether it is a subset, and uni whether it has a map back to Unicode text. A long list of fonts marked embedded but not subset points to the font cause. The same font name repeated many times points to a merged file.

If you prefer to stay in a viewer, a quick test still helps. Search a page for a word you can see. If it finds nothing, the page is an image. Then compare the file size to the page count. Several megabytes per page almost always means scans. Developers who want to inspect a file in code can start from [extracting a table from a PDF in JavaScript](/blog/extract-table-from-pdf-in-javascript), which walks the same page objects.

The table below gathers the signs. The examples are typical, not exact thresholds.

| What you notice (example) | Likely cause | How to confirm |
| --- | --- | --- |
| Several MB per page, text not searchable | Scanned or photographed pages | `pdfimages -list` shows one large image per page |
| Search works, but file is still large | Big logo or background images | `pdfimages -list` shows the same image size on every page |
| Text-only file of a few MB | Fully embedded or duplicated fonts | `pdffonts` shows fonts not subset, or repeated names |
| Size far above what the pages suggest | Attachments or saved edit history | Attachments panel, or the size drops after saving as a new file |

## When a large PDF file size does not slow conversion

A large file size slows the upload, but not always the conversion. A digital PDF with a heavy logo uploads slowly, then converts as fast as any text file, because the converter reads the text and ignores the picture. A large scan is different: its size reflects pages that need text recognition, and that is slow.

The opposite case catches people out. A small PDF can be the slow one. A 2 MB digital file with 200 pages of dense transactions can take longer to convert than a 40 MB file of five photographed pages, because reading work grows with pages and rows, not megabytes. It also produces a much bigger spreadsheet. If a long statement leaves Excel sluggish after conversion, [speeding up Excel with many rows](/blog/excel-slow-with-many-rows) deals with that side.

So before you shrink a file, ask what you are trying to fix. If the upload is slow or the file is over a size limit, shrinking the images helps. If the conversion is slow and the file is digital, shrinking does nothing and splitting by page count is the answer. FastBankConvert is built to read a typical digital statement in seconds, expects scans to take longer, and shows on screen how long each conversion took, so you can see whether a large file was really a slow conversion. [The homepage walks through a sample statement](/) from drop to download.