Which format gives more accurate results if I am going to run OCR?

PNG is generally the safer choice because it works losslessly and preserves the edge crispness of letters without any degradation. OCR engines recognise character shapes by looking at edge information; the ringing artefacts JPG compression produces around sharp edges blur that information and can increase recognition errors, especially in small point-size text. With a JPG saved at a high quality setting the difference usually stays at an acceptable level, but at low quality it degrades noticeably. On work where accuracy is critical, choosing PNG means not taking unnecessary risks.

I did a colour scan and the PNG file came out very large. What should I do?

This is to be expected; because PNG is lossless it cannot compress continuously varying content such as paper texture, shading and coloured backgrounds efficiently. If your document contains only text, scanning in greyscale or black and white shrinks the PNG file dramatically. If it has to stay in colour and size is becoming a problem, a JPG at a high quality setting is a more balanced solution. If you are using an unnecessarily high DPI value in your scanner settings, lowering it also delivers a serious gain.

Which should I use for a document with legal standing?

With documents of this kind the real rule is to keep the original scan file, that is the TIFF, unaltered; the file you convert is always a working copy. For the sharing copy, PNG is a more defensible choice because it is lossless and does not change the image at all. If you have to use JPG, keep the quality setting high, because the artefacts that form at low quality can degrade fine detail such as signatures or stamps. If your organisation or the relevant authority has specified a format requirement, you must comply with it in every case.

Do I have to split my multi-page scan into separate files per page?

If you are converting to PNG or JPG, yes, because both formats are single-image and cannot hold more than one page in a single file. When a multi-page TIFF is converted, either only the first page is taken or each page comes out as a separate file. If the pages need to stay whole in a single document, the correct destination is PDF; PDF supports multiple pages and opens everywhere. If you are going to share the document and the other party will follow the page order, PDF is almost always the better choice.

For a Scanned Document, Should You Convert TIFF to PNG or JPG?

A scanned document is a different problem from other images

Choosing a format for a photograph is relatively easy: JPG is almost always the right answer. But if you have a scanned contract, invoice, diploma or archival document in hand, the equation changes. Here the issue is not aesthetics but legibility and accuracy.

Scanned documents are usually produced as TIFF — scanners and document management systems use this format by default. But you cannot share a TIFF: it does not fit in an email, it does not display on the web, and the other party's software may not open it. It has to be converted.

The question is: PNG or JPG?

The decisive factor: what is in the scan

There is no single answer, because "scanned document" is not a single thing. There are three different types, and each gives a different answer.

Type 1: A clean black-and-white text scan

A page containing only text and lines, with a clean background. A typical contract, official letter or form.

Winner: PNG, by a clear margin.

The reason is both quality and size. Because PNG is lossless, the edges of the letters are not degraded at all. JPG, by contrast, produces grey shadowing known as "ringing" at sharp black-and-white transitions; a dirty halo forms around the letters. This is easily visible to the eye at low quality settings and makes the document look amateurish.

PNG wins on the size side too. Lossless compression is extremely efficient on flat areas containing few colours. A page consisting of black text on a white background comes out surprisingly small as a PNG — often smaller than the JPG.

So there is not even a trade-off here: PNG is both cleaner and smaller.

Type 2: A scan with paper texture, greyscale or colour

An old document, yellowed paper, a shadowed scan, a form with colour in its background.

Winner: a high-quality JPG, most of the time.

Here the equation reverses. Paper texture and shading are the content type lossless compression struggles with most — every pixel differs slightly from its neighbour and PNG cannot encode that efficiently. The result is a needlessly large file.

JPG was designed for this kind of continuous-tone content and produces far more reasonable sizes. Provided you keep the quality setting high, legibility is preserved.

Important warning: the emphasis on "high quality" here is serious. Lowering the quality on a scanned document gives a far worse result than lowering the quality on a photograph. In a photograph the artefacts blend into the texture and hide; in text they land directly on top of the letters.

Type 3: Mixed content — text plus photographs

An identity document, an application form with a photo, an illustrated catalogue page.

It depends. If the weight is on text, PNG; if the weight is on photographs, JPG. If you have no size constraint, PNG is the safe choice because it creates no degradation in any region.

If you are going to run OCR

If you plan to convert the document to text, prefer PNG.

OCR engines recognise characters by looking at edge information. JPG artefacts blur those edges and can increase recognition errors, particularly with small point-size text, fine serif typefaces and handwriting. With a high-quality JPG the difference usually stays acceptable, but at low quality it degrades noticeably.

The only cost of using PNG is file size. If OCR accuracy matters to you, that cost is worth paying.

Rules that apply in both cases

Keep the original TIFF

This is the most important rule, independent of format. The PNG or JPG you produce is a delivery copy. For documents with legal standing or destined for an archive, the original scan file must be preserved unaltered. If you need a different format later, you regenerate it from that.

Check the page count

TIFF can be multi-page; PNG and JPG cannot. In the conversion, either the first page is taken or each page becomes a separate file. If you are working with a scanned contract, make sure all the pages came out.

If the pages must stay together, think about PDF

This is the most frequently overlooked option. If you are going to share a multi-page document, PNG or JPG is not the right destination. PDF supports multiple pages, opens on every device, and preserves the page order.

Decision summary

| Situation | Choice | |---|---| | Clean black-and-white text | PNG | | OCR to be run | PNG | | Paper-textured / colour scan, size matters | High-quality JPG | | Sharing copy of a legal document | PNG | | Multi-page document | PDF | | Archival master copy | leave it as TIFF |

In summary

Choosing a format for a scanned document follows a different logic from choosing one for a photograph. With text-heavy, clean scans, converting to PNG wins on both quality and size — a rare situation. With scans containing paper texture and colour, a high-quality JPG gives a more balanced result. In both cases, do not skimp on the quality setting and do not delete the original TIFF; a document, by definition, is something you will need again.