Put a Logo or a Photograph in a PDF

A letterhead needs a logo. A report needs a photograph of the thing it is about. The document builder has an Image block for both, and it deliberately does less than an image in a word processor does. Knowing what it does, and what it refuses to do, saves you the half hour of fighting it.

Each image sits on its own line, at the width you give it, in the position you put it in the block list. Text does not wrap around it. Two images do not sit side by side. An image cannot go inside a table cell, and a logo cannot sit to the left of an address block on the same line. None of those are settings switched off somewhere. They are not built, and saying so is more useful than letting you hunt for them.

What you get in exchange is predictability. The image takes the room it needs, everything below moves down by that much, and the layout never has to make a clever decision about how text should run around a shape.

The width field takes a measurement in points, where 72 points is one inch. Leave it blank and the image fills the text column. Whatever you type is capped at the column width, so an image can never spill into the margins. The height is not yours to set: it comes from the source image's own pixel dimensions, so the proportions cannot be squashed by accident, and an image taller than the content box is shrunk to fit rather than cut in half.

• A wordmark at the top of a letter: 108 pt, an inch and a half wide.
• A badge, stamp or small mark: 72 pt, one inch.
• A figure in a report: leave the width blank and let it fill the text column.

Alignment is left, centre or right, and it applies to the image as a block rather than to anything around it. A centred mark above a centred heading is the one arrangement that looks intentional with no other effort.

Points are not pixels, and this is where a built PDF gets unexpectedly heavy. The width you choose decides how large the image is drawn on paper. It does not decide how many pixels go into the file: the picture is embedded as you supplied it, so nothing is resampled down to the size you are drawing it at.

A 600 pixel wide logo drawn at 108 points works out at roughly 400 pixels per inch, which is more than any printer needs and costs almost nothing. A 4,000 pixel photograph drawn at that same 108 points is still a 4,000 pixel photograph inside the PDF, and every one of those bytes travels with the document for the rest of its life.

A PDF page can carry exactly two raster formats: PNG and JPEG. A PNG or a JPEG goes in as it is. Anything else, meaning WebP, AVIF, GIF or SVG, is converted the moment you choose the file rather than at export. That timing is deliberate: finding out that a picture cannot be used as it stands is much cheaper before you have built a document around it.

• The conversion targets PNG, which is what keeps transparency intact through it.
• An animated GIF contributes its first frame. A PDF page has no animation to carry the rest into.
• An SVG is rasterised at its natural size. It stops being a vector, so it will soften if it is scaled up later. Export the artwork at the size you plan to draw it, or larger.
• A photograph converted to PNG usually grows. PNG is lossless, and a WebP or AVIF photo re-encoded as PNG can be several times the size it arrived at. Hand over a JPEG instead.

A logo with a transparent background has to arrive as a PNG. JPEG carries no transparency at all, so the same logo saved as a JPEG brings a rectangle of whatever colour it was flattened onto, usually white. On a white page that rectangle is invisible, right up until somebody prints the document on coloured stock or reads it against a dark background.

There is no background removal here and no way to knock a colour out. If the transparency is not already in the file you hand over, it is not going to appear in the PDF.

The image block has an alt text field, and it is worth being precise about what that currently achieves. The value is kept with the block and carried through the layout, so a later tagged export has something to write. Today the exporter emits no tagged structure tree, so the words you type there do not reach the finished PDF as something assistive software can announce.

Fill it in anyway, because it costs one sentence. What you must not do is let it stand in for the accessible version of the information. If an image carries meaning the document would lose without it, put that meaning into a paragraph underneath. A caption is real embedded text: everything can read it, and it also helps the sighted reader who could not tell what the chart was meant to show either.

1. 1

   Put the image block first

   Add it before anything else so the mark is the first thing on the page, then choose the file. A block with no file chosen yet is skipped on export rather than reserving a blank gap.

2. 2

   Set a width and an alignment

   Between 72 and 144 points suits most wordmarks. Left aligned reads as correspondence, centred reads as a certificate.

3. 3

   Add a Spacing block underneath

   An image leaves the same 10 points behind it that a paragraph does, which is far too tight under a logo. A Spacing block of 18 to 36 points gives it room.

4. 4

   Then the address, as one paragraph

   Line breaks inside a paragraph block are kept, so the whole sender block can be one right-aligned paragraph with a return between each line.