Convert BMP to PDF

That gap is the whole reason this conversion exists. Scanner drivers, fax gateways, older document systems and a good deal of Windows-era imaging software write BMP because it is the format they have always written, and the place the document has to go accepts PDF and often nothing else.

The conversion is therefore administrative rather than creative. Nothing about the picture needs to improve; it needs to arrive in a container the recipient will take, at a page size that does not make somebody on the other end wonder what happened. The rest of this page is about that second half, because it is the half that goes wrong. A PDF that contains the right document at the wrong size is accepted by the upload, opens on somebody’s screen looking absurd, and comes back a week later as a request to send it again.

It is the default because most images converted to a PDF are going into an email or an upload rather than onto paper, and a page that is exactly the picture has no letterbox and no guessed-at paper size. For a photograph, a diagram or a screenshot that is the right answer.

For a scan it is a trap, and the arithmetic is the reason. A PDF point is 1/72 of an inch and the page is sized one point per pixel, so an A4 sheet scanned at 300 dpi — 2,480 by 3,508 pixels, about 26 MB as a bitmap — becomes a PDF page roughly 34 by 49 inches. It contains the right document and it opens as something the size of a poster.

Both are offered and both do the obvious thing: the page becomes 595 by 842 points for A4 or 612 by 792 for US Letter, and the picture is placed centred inside it. That is what somebody printing or filing the document expects to receive.

Orientation follows the same logic. It is chosen automatically unless you set it, and it is ignored when the page size is "fit to image", because a page that is already the shape of the picture has no orientation to change. A landscape scan on an A4 page wants the landscape setting rather than a rotated bitmap: rotating the image and leaving the page portrait produces a document somebody has to turn their head to read, and it is the more common of the two mistakes.

Three things are worth thirty seconds. The page dimensions, which any viewer will show you and which are the failure this page keeps warning about. The page count, which should be one per bitmap and surprises people who expected a document. And the legibility at 100 per cent, since a scan that was hard to read as a bitmap is exactly as hard to read as a PDF.

The reason to check rather than to assume is that a submission is usually a one-shot interaction with a system that will not tell you what was wrong with the file. An upload form accepts anything with the right extension under the right size, and the person who eventually looks at the document is not the person who can explain the problem to you. Every one of the three checks above catches a failure that this conversion can produce without erroring.

A PDF can embed a JPEG or a PNG directly. A bitmap is neither, so it is decoded to pixels and re-encoded as PNG on the way in — which is lossless, so every value the BMP held is a value in the document, but it is not free in time the way passing a JPEG straight through is.

What that means in practice is that this conversion is honest about quality and slightly slower than converting a photograph, where the JPEG is passed straight through and costs nothing at all. There is no quality setting because there is no loss to control, and no amount of waiting produces a better picture than the one the bitmap contained. If a scan is unreadable in the BMP it will be unreadable in the PDF, and the fix is at the scanner rather than here — 300 dpi in greyscale is the setting that makes a page of type reliably legible.

There is no assembly step here. Each file in becomes a single-page document out, so converting a folder of eight scanned pages gives you eight PDFs rather than one eight-page document.

That is a real limitation for the most common reason people convert a scan, and it is better said than discovered. Merging is a separate job with its own tool; if the destination will accept several files, sending them as they are is usually faster than looking for one.

Whatever the scan says, the document holds a grid of pixels. Nothing in it can be selected, searched, copied or read aloud, and a recipient who expected a text document will notice within seconds.

If the words themselves are what is needed, that is optical character recognition and it is the BMP to TXT conversion on this site. The two are often wanted together — a PDF to submit and the text to paste into a form — and they are genuinely different operations, one of which can be confidently wrong while the other cannot. This one either produces the document or fails with a message; recognition produces fluent text whether or not it matches the page, which is why the two are kept on separate pages rather than offered as one button.

Much smaller, because the uncompressed bulk of a BMP does not survive being re-encoded. A 26 MB A4 scan at 300 dpi produces a document of a few megabytes; a 6.2 MB screen capture at 1920 by 1080 produces a smaller one still. The size is roughly what a PNG of the same picture would be, plus a small amount of document structure.

If the result is still too large for an upload limit, the lever is the resolution rather than the format. A page scanned at 600 dpi carries four times the pixels of the same page at 300, and almost never four times the useful information — 300 dpi is enough for text and 150 is enough for something that will only ever be read on a screen. This page offers page size, orientation and margin rather than a resize control, so the reduction belongs upstream: re-scan at a sensible resolution, or convert the bitmap to a smaller JPG or PNG first and make the PDF from that.

When a fixed page size is chosen, the image is reduced to fit inside the page and its margin, keeping its proportions. If it is already smaller than the space, it is left alone and centred.

That asymmetry is deliberate. Enlarging a small bitmap onto an A4 page turns a crisp logo or a low-resolution scan into a blurred one, and a converter that did it silently would be making a picture worse in order to fill a page nobody asked it to fill. A small image on a large page looks like a small image; that is honest and it is recoverable.

Drop the folder. Each bitmap is decoded, re-encoded and wrapped in its own PDF on your own machine, keeps its filename with the extension swapped, and the set comes back as one ZIP. The ceiling is 100 MB per file, which is about 33 megapixels of uncompressed bitmap — comfortably more than any document scan.

Doing it locally is the part that matters for this pair specifically. The things people scan are contracts, payslips, medical letters, insurance claims and identity documents, and every one of them is a file that should not be handed to a third party in exchange for a container change.