Converting a PDF to greyscale without growing the file

Converting a document to greyscale is usually done for one of two reasons: to cut printing costs, or to shrink the file. It reliably achieves the first. It achieves the second most of the time, and understanding the exception saves you from shipping a larger file than you started with.

What colour actually is inside a PDF

A PDF does not store “blue”. It stores a number, and a colour space that says how to interpret it.

• DeviceRGB — three values, for screens
• DeviceCMYK — four values, for printing presses
• DeviceGray — one value, from black to white

Converting to greyscale means rewriting every colour value into DeviceGray. The standard conversion is a luminance calculation: green contributes most to perceived brightness, red less, blue least. That weighting is why a well-converted greyscale image looks natural rather than flat.

Text and vectors convert for free

Here is the good news, and it covers most documents.

Text in a PDF is a set of vector glyph outlines with a colour value attached. Convert that value and the outlines are untouched. The text stays exactly as sharp as it was, at any zoom, and stays selectable and searchable.

Vector graphics behave identically. A chart drawn as lines and filled shapes converts perfectly — the geometry is unchanged, only the fill colours are rewritten.

So a text-heavy document with vector charts converts losslessly and gets slightly smaller. Grayscale PDF removes colour while preserving text and vector content where possible, and reports what it had to re-encode.

Images are where it gets expensive

An image is different in kind. It has no geometry to preserve — it is a compressed block of pixel data.

To convert it, the tool must:

1. Decode the compressed image back to pixels
2. Convert each pixel to a single grey value
3. Re-encode the result

Step three is the problem. The new compression is a fresh lossy pass over an image that has already paid that cost once, and there is no guarantee it lands smaller than the original.

A tightly compressed colour scan is the classic case that grows. If the original was encoded at aggressive settings, the re-encode often cannot match that efficiency, and a file that was 4 MB comes back at 5 MB — now also slightly degraded.

This is why the tool warns when a tightly compressed scan can grow rather than silently producing a worse file. Compare the byte counts and keep the smaller one.

When it saves real money

The print-cost case is the strongest reason to do this, and it works because of how printers count.

Most office devices and print shops bill colour pages at several times the mono rate. Crucially, many count a page as colour if it contains any non-grey value — a pale tint behind a table header, a coloured hyperlink, a logo in the footer.

So a fifty-page report that is visually monochrome can be billed as fifty colour pages because of a blue heading style nobody thought about. Converting the whole document removes most of that ambiguity, but not always all of it: spot inks such as Pantone colours, some gradients, JPEG 2000 and CMYK JPEG images, very large JPEGs and small inline images are left in colour and listed in the result. Converted JPEGs also keep an RGB label even though their pixels are grey.

For a print run of any size, that is the difference that pays for the effort.

Greyscale is not black and white

These get conflated and they are not close.

• Greyscale keeps 256 levels. Photographs keep their tonal range, shading survives, gradients still look like gradients.
• Bilevel (1-bit) allows only pure black and pure white. Text becomes astonishingly small — a scanned page can drop to a few kilobytes — but every photograph becomes a harsh dithered pattern.

Bilevel is right for a scan of typed text and nothing else. For any document with an image in it, greyscale is the choice that still looks like a document afterwards.

What you lose that you might need

Be deliberate about this before converting, because it cannot be undone from the converted file:

• Colour-coded information. A chart whose series are distinguished only by colour becomes unreadable — several grey lines with no way to tell them apart. If the chart has no pattern or label distinction, converting destroys its meaning.
• Red-marked edits. Tracked changes, annotations and corrections marked in red become ordinary grey text, indistinguishable from the body.
• Status indicators. Red/amber/green cells in a converted table all read as similar greys.
• Brand colour, if the document is going to a client.

Keep the original. The conversion is one-directional; the colour values are gone from the new file and no tool recovers them.

The right order when you also want it smaller

If the goal is a smaller file, sequence matters:

1. Convert to greyscale first, and check whether the file grew.
2. Then run Compress PDF over the result.

Compressing first and converting second means the images are re-encoded twice, paying the lossy cost on both passes for no benefit. One re-encode is unavoidable; two is waste, and it shows.

If greyscale conversion made the file larger and print cost is not your concern, discard it and compress the colour original instead.

Before you send it

• Compare byte counts against the original and keep the smaller file.
• Look at every chart. This is where conversion silently destroys meaning.
• Read a page of body text to confirm nothing was flattened — text should still select.
• Check the images at normal reading zoom, not at 400%.
• Check the metadata if the file is going outside your organisation; the title, author and producer carry over unchanged, so the title may still name the colour original.