# PDF vs JPG vs PNG: Which Format for Scans, Photos, Forms and Printing

Use PDF for documents and anything with several pages, JPG for photographs, and PNG for screenshots, diagrams and text. Why, with file sizes we measured.

---

- **Canonical URL:** https://dothecalculation.com/blog/files/pdf-vs-jpg-vs-png
- **Category:** Files & PDFs
- **Author:** Do The Calculation Team
- **Published:** 2026-09-29
- **Reading time:** 17 min read
- **Publisher:** Do The Calculation (https://dothecalculation.com)
- **Methodology:** https://dothecalculation.com/methodology

---

## PDF, JPG or PNG: the short answer

Use PDF for documents: anything with text, anything with more than one page, anything that will be printed, signed, filed or uploaded to a portal. Use JPG for photographs, where its compression makes files small without visible loss. Use PNG for screenshots, diagrams, logos and images of text, where every sharp edge needs to survive. A PDF can hold JPG and PNG pictures inside it, so the three are not rivals so much as a container and two kinds of image.

The rest of this guide explains what each format actually is, what happens to quality and file size when you choose one over another, with numbers we measured ourselves, and which format suits common jobs, from scanning a contract to posting a screenshot. It also covers what government portals ask for, what resolution to scan at, and how archives decide which formats will still open in fifty years.

Tool: [Turn photos or scans into one PDF](https://dothecalculation.com/tools/jpg-to-pdf) — The free JPG to PDF converter runs in your browser, keeps JPG photos at their original quality and never uploads your files.

## What each format actually is

### PDF: a container of pages

A PDF (Portable Document Format) describes pages of a fixed size, and what is drawn on each: real text in a given font, vector shapes such as lines and table borders, and images. Adobe created it so that a document looks the same on any screen and printer (https://www.adobe.com/acrobat/about-adobe-pdf.html), and it is now an international standard, ISO 32000; the current version, PDF 2.0, is described by the Library of Congress as a format for exchanging and viewing documents independently of the software that made them (https://www.loc.gov/preservation/digital/formats/fdd/fdd000474.shtml).

Because the text in a PDF made from a word processor is real text, you can search it, copy it, and zoom in forever without it going blurry. A scanned PDF is different: each page is one big picture, and its text is only searchable if someone ran text recognition (OCR) over it and added an invisible text layer.

### JPG: a lossy photograph format

JPG (or JPEG, the same thing) stores a grid of pixels using the compression method standardised as ISO/IEC 10918-1, which jpeg.org itself describes as a lossy format (https://jpeg.org/jpeg/). Lossy means the encoder throws away detail it predicts you will not notice: very fine texture and sharp changes in colour, which it approximates in blocks of 8 × 8 pixels. On a photograph of a landscape or a face, that is hard to see and saves a great deal of space. On black text on white paper, it leaves faint grey smudges around each letter.

### PNG: a lossless image format

PNG (Portable Network Graphics) also stores a grid of pixels, but losslessly: the W3C specification describes it as a format for the lossless, portable, well-compressed storage of raster images (https://www.w3.org/TR/png-3/). Every pixel comes back exactly as it went in, however many times the file is saved. It compresses flat areas and repeated patterns very well, so screenshots and diagrams make small files, but it cannot shrink the random detail of a photograph much, so photo PNGs are large. PNG can also store transparency, which JPG cannot.

## Lossy vs lossless, in practice

Zoom a JPG of a text page to 400 percent and look at the letter e. At quality 95 the edge looks clean. At 85 a faint halo of slightly darker grey pixels appears around the curves, visible only if you look for it. At 70 the halo is obvious, the inside of the letter is no longer pure white, and there are small square patches where the encoder’s 8 × 8 blocks meet. The same zoom on a PNG shows exactly the pixels the scanner produced: the soft grey edge the scanner itself created, and nothing else.

On a photograph of trees at the same quality settings, you would struggle to point to a difference between 95 and 85 at normal viewing size, and 70 shows only a slight loss of texture in the leaves. That is the whole case for JPG: the loss lands where photographs hide it and text does not.

**What each format keeps and loses**
|  | PDF | JPG | PNG |
| --- | --- | --- | --- |
| Compression | Depends on what is inside: text and vectors are exact; images keep their own compression | Lossy | Lossless |
| Best for | Documents, forms, multi-page files, print | Photographs | Screenshots, diagrams, logos, text images |
| Sharp text edges | Perfect if the text is real text | Soft, with halos at lower quality | Exact |
| Transparency | Yes, inside the page | No | Yes |
| Several pages in one file | Yes | No | No |
| Searchable text | Yes, if the text is real or OCR was run | No | No |
| Re-saving loses quality | No, unless images inside are recompressed | Yes, a little each time | No |

That last row matters more than people expect. Every time a JPG is opened, edited and saved as a JPG again, it goes through the lossy step again and loses a little more. A photo that has been cropped, rotated and re-saved five times in different apps is visibly worse than the original. Keep an untouched original, and when you put photos into a PDF, use a tool that embeds the JPG as it is rather than re-encoding it.

## File sizes we measured

To put numbers on this, we took two test images: a 300 DPI colour scan of an A4 page of printed text (2481 × 3508 pixels, including the faint grain every scanner adds), and a 12-megapixel photograph (4032 × 3024 pixels, the size most phones produce). We saved each as JPG at three quality settings with the Pillow imaging library, and as PNG, then put each file into a PDF on its own page with our [JPG to PDF converter](/tools/jpg-to-pdf) in its Original quality setting. These are the byte counts of the files.

**A4 text scan at 300 DPI (2481 × 3508 px)**
| Saved as | Image file | PDF containing it |
| --- | --- | --- |
| JPG, quality 95 | 1,492,944 bytes (1.4 MB) | 1,494,150 bytes |
| JPG, quality 85 | 682,273 bytes (666 KB) | 683,476 bytes |
| JPG, quality 70 | 523,828 bytes (512 KB) | 525,031 bytes |
| PNG, colour | 7,194,187 bytes (6.9 MB) | 5,260,924 bytes (stored losslessly) |
| PNG, greyscale | 3,083,333 bytes (2.9 MB) | 3,415,879 bytes (stored as colour) |

**12-megapixel photograph (4032 × 3024 px)**
| Saved as | Image file | PDF containing it |
| --- | --- | --- |
| JPG, quality 95 | 2,980,339 bytes (2.8 MB) | 2,981,551 bytes |
| JPG, quality 85 | 1,985,623 bytes (1.9 MB) | 1,986,834 bytes |
| JPG, quality 70 | 705,808 bytes (689 KB) | 707,017 bytes |
| PNG | 10,149,472 bytes (9.7 MB) | 11,768,866 bytes (stored losslessly) |

Three things stand out. First, a PDF that holds a JPG is only about 1,200 bytes larger than the JPG itself: the PDF adds a page description and a few lines of structure, and the picture goes in unchanged. Converting a JPG to PDF does not have to cost size or quality. Second, PNG is a poor choice for a photograph or for a scan with scanner grain: the photo PNG is more than five times the size of a quality-95 JPG that looks the same. Third, going from quality 95 to 85 roughly halved the scan and cut the photo by a third, while 85 to 70 saved much less on the scan and a lot on the photo, because the grain in the photo was what the encoder was paying for.

> **Where PNG wins** — The scan above is a photograph of paper, grain and all, which PNG compresses badly. A page of text drawn digitally, such as a PDF page converted to an image or a screenshot of a document, has flat white areas and clean edges. We converted pages of a text PDF to PNG at 300 DPI with our PDF to JPG converter and got files of about 600 KB each, smaller than the same pages as quality-85 JPG, and perfectly sharp. For screenshots and exported pages, PNG is usually both better and smaller.

## Text edges, colour and transparency

Text is where the formats differ most. A PDF made from a word processor or a web page keeps the text as characters and draws them at whatever size you zoom to, so it is always sharp. The moment the page becomes a picture, as a JPG or PNG, the text is frozen at one resolution: zoom in and you see pixels. As a PNG, those pixels are exactly what was rendered; as a JPG, they also carry compression halos.

Colour works the same way in all three for everyday use: files are in the sRGB colour space that screens use, unless a colour profile says otherwise. PDFs prepared for commercial printing may use CMYK, the four inks of a printing press, and a print shop may ask for that. JPG can store CMYK too, but most web tools expect RGB.

Transparency is simple: PNG has it, JPG does not, and a PDF can place a transparent image on a page, but the page itself is paper, so whatever shows through is the page colour. A logo with a transparent background put into a PDF appears on white unless the page is coloured.

## Metadata: what travels with the file

Photos from phones and cameras carry EXIF metadata: the date and time the photo was taken, the phone’s make and model, the exposure settings and, if location was switched on, GPS coordinates accurate to a few metres. JPG, PNG and HEIC can all carry it. Send a photo of a receipt taken at home and you may be sending your home address as coordinates.

A PDF has its own document information: a title, author, subject, keywords, the application that created it and creation and modification dates. Word processors fill in the author from your account name. And a PDF that contains a JPG may also contain that JPG’s EXIF block, because the image goes in byte for byte. Our JPG to PDF converter strips EXIF, XMP and IPTC from photos before embedding them, without recompressing the picture, and writes only a title and its own name into the PDF. To check any PDF, open File, Properties in Adobe Acrobat Reader, or Tools, Show Inspector in Preview on a Mac.

## Several pages, and keeping them in order

A JPG or PNG is one picture. A five-page contract photographed as five JPGs arrives as five files that the recipient has to open in the right order, and may not: email clients and portals often sort attachments by name, and IMG_0010 can land before IMG_0009 in a plain alphabetical sort. A PDF keeps the pages in one file in the order you set, with page numbers, which is why nearly every office, court and portal asks for documents as PDF.

If you have the pages as photos, combine them into one PDF before sending, checking the order and that every page is upright. If you need a single page of a PDF as an image, for a slide or a portal that only accepts JPG, convert just that page rather than the whole document.

## Searchability and OCR

A JPG or PNG of a page is a picture of text, not text. You cannot search it, copy a sentence from it, or have a screen reader read it aloud until optical character recognition (OCR) turns the shapes of letters back into characters. Adobe Acrobat, Microsoft OneNote, Google Drive and Apple’s Live Text all do this, with results that depend heavily on the image: 300 DPI, straight, evenly lit, with no JPG smudging, gives the best accuracy.

A PDF made from a document that was typed is already searchable. A scanned PDF becomes searchable once OCR adds an invisible text layer behind the image, which many scanners and scanning apps do automatically. Converting a searchable PDF to JPG throws that text away, so keep the PDF as well.

## Editing, forms, signatures and security

PDFs can do things images cannot. They can contain fillable form fields, digital signatures that prove who signed and whether anything changed afterwards, links, bookmarks, attachments and even JavaScript for interactive forms. They can be encrypted with a password that must be entered to open the file, or restricted with a permissions password that asks readers not to allow printing or copying.

Those features cut both ways. A PDF form is useful; a PDF with scripts or attachments from an unknown sender deserves the same caution as any other attachment. Converting a signed PDF to images keeps the look of the signature but removes the digital signature itself, so the recipient can no longer verify it. An image has none of these features, which makes it simpler and, for that reason, what some upload systems prefer for photos of ID documents.

## Archiving: PDF/A and what archives recommend

PDF/A is a version of PDF designed for long-term archiving, standardised as ISO 19005 (https://pdfa.org/resource/pdfa-in-a-nutshell/). It forbids features that could stop a file opening in the future, such as fonts that are not embedded, encryption and external dependencies. The Library of Congress describes PDF/A-1 as a constrained form of PDF 1.4 intended for long-term preservation of page-oriented documents (https://www.loc.gov/preservation/digital/formats/fdd/fdd000125.shtml).

The Library of Congress publishes Recommended Formats Statements saying which formats it prefers to receive. For textual works it lists XML-based formats and EPUB first, then PDF/UA and PDF/A, then ordinary PDF (https://www.loc.gov/preservation/resources/rfs/text.html). For still images its order of preference is TIFF, then JPEG 2000, then PNG, then JPEG, and it asks for the highest resolution available, not rescaled (https://www.loc.gov/preservation/resources/rfs/stillimg.html). For a household or small business, the practical lesson is simpler: keep documents as PDF (PDF/A if your software offers it), keep original photos as the camera produced them, and do not throw away the originals after making smaller copies.

## What portals and inboxes actually ask for

Official upload systems usually state their rules, and they vary. Three examples, checked on the organisations’ own pages on 29 September 2026:

- US Citizenship and Immigration Services, for online filing: files must be PDF, JPG or JPEG (some forms also accept TIFF), and no larger than 12 MB each (https://www.uscis.gov/file-online/tips-for-filing-forms-online).
- Immigration, Refugees and Citizenship Canada: each document uploaded to an IRCC secure account has a 4 MB limit, and the IRCC Portal allows 5 MB (https://ircc.canada.ca/english/helpcentre/answer.asp?qnum=1123&top=23).
- UK visa applications: supporting documents can be uploaded as PDF, PNG, JPG or JPEG (https://www.gov.uk/guidance/uploading-evidence-as-part-of-your-visa-application), while a digital passport photo must be between 50 KB and 10 MB (https://www.gov.uk/photos-for-passports).

Email has limits too. Gmail allows attachments of up to 25 MB for personal accounts (https://support.google.com/mail/answer/6584) and Outlook.com has the same 25 MB limit (https://support.microsoft.com/en-us/outlook/sending-limits-in-outlook-com). A PDF of twenty phone photos at original quality can easily pass that, so either send fewer pages per message or rebuild the PDF with smaller images. When a limit applies, reduce the pixel size first (about 2000 to 2400 pixels on the long side keeps an A4 page printable) and the JPG quality second, and check the smallest text before sending. For a PDF you cannot rebuild from the original images, [Compress PDF](/tools/compress-pdf) recompresses the pictures inside it the same way, leaves the text as text, and shows the size before and after.

## DPI for scanning and printing

DPI (dots per inch) or PPI (pixels per inch) is how many pixels represent each inch of paper. For scanning documents, the US National Archives’ guidance for textual records sets 300 ppi as the minimum for colour and greyscale scans and recommends 400, and allows black-and-white scans of clean printed type at 300 to 600 ppi (https://www.archives.gov/records-mgmt/initiatives/scanned-textual.html). For printing photos, 300 pixels per inch is the usual target; Adobe calls it a go-to option for good print material (https://www.adobe.com/uk/creativecloud/photography/discover/pixels-per-inch-ppi-resolution.html).

**Pixel size of an A4 and a US Letter page at common DPI settings**
| DPI | A4 (210 × 297 mm) | US Letter (8.5 × 11 in) | Good for |
| --- | --- | --- | --- |
| 72 | 595 × 842 | 612 × 792 | Thumbnails, previews |
| 150 | 1240 × 1754 | 1275 × 1650 | Reading on screen, email, slides |
| 300 | 2480 × 3508 | 2550 × 3300 | Printing, OCR, most portals |
| 600 | 4961 × 7016 | 5100 × 6600 | Small print, drawings, archiving |

Doubling the DPI quadruples the number of pixels, and roughly quadruples the file size. A page scanned at 600 DPI is not twice as good as 300 for ordinary text; it is four times as large. The [screen resolution PPI calculator](/calculators/screen-resolution-ppi-calculator) shows the same relationship for screens, if you want to match an image to a display.

## Which format for which job

**A quick decision table**
| Task | Use | Why |
| --- | --- | --- |
| Scan a signed contract | PDF, 300 DPI, all pages in one file | Keeps order and page size; prints correctly; accepted everywhere |
| Send ID documents to a portal | Whatever the portal lists, usually PDF or JPG, within its size limit | Portals reject other formats outright; keep all four corners visible |
| Share a photo | JPG | Small files, full photographic quality |
| Post a screenshot | PNG | Text and interface edges stay sharp; flat areas compress well |
| Send receipts for expenses | One PDF with every receipt | One file to attach, in order, with location data removed |
| Print a poster | PDF (vector) or a high-resolution image at 150 to 300 DPI at print size | Large prints are viewed from further away, but text must stay crisp |
| Archive records | PDF/A for documents, original camera files or TIFF/PNG for images | Formats archives recommend for long-term access |
| Put a page into a slide | PNG of that page at 150 DPI | Sharp text at slide size, small file |

## How to convert each way without losing quality

Images to PDF: use a converter that embeds JPG photos as they are rather than recompressing them, lets you set the page size (A4 or Letter for anything printed) and order, and removes photo metadata. Our [JPG to PDF converter](/tools/jpg-to-pdf) does all of that in the browser, and turns iPhone HEIC photos into PDF pages too. On a computer you can also open the images together in Preview on a Mac and use File, Print, then Save as PDF, or print them to Microsoft Print to PDF in Windows.

PDF to images: decide whether you want a picture of the page or the photos inside it. Rendering a page at 150 DPI suits screens, 300 DPI suits printing and OCR, and PNG keeps text sharp. To get the original photos out of a PDF, extract them instead: a good extractor saves embedded JPGs byte for byte, so nothing is lost. Before uploading a private document to any online converter, read our guide [Are online PDF converters safe?](/blog/files/are-online-pdf-converters-safe), which explains what happens to the file and how to check a site in a minute. More guides on scanning, converting and sending files are collected in [Files and PDFs](/blog/files).

Tool: [Turn PDF pages into JPG or PNG](https://dothecalculation.com/tools/pdf-to-jpg) — The free PDF to JPG converter renders pages at 72 to 600 DPI or extracts the original images, in your browser, with no upload.

## Frequently asked questions

**Should I scan documents as PDF or JPEG?**

As PDF, in most cases. A PDF keeps several pages in one file in the right order, carries the page size so it prints correctly, and is what offices, courts and portals usually ask for. Scan to JPEG only for a single photo or when the recipient specifically asks for images. If your scanner or phone produced JPEGs already, combine them into a PDF before sending.

**Can a PDF contain a JPG?**

Yes. A PDF can hold JPG data directly, and a good converter puts your photo into the PDF byte for byte, so the picture inside is identical to the file you started with. In our measurements the PDF was only about 1,200 bytes larger than the JPG it contained. The same photo can later be extracted from the PDF unchanged.

**Is PDF better quality than JPG?**

It depends what is inside the PDF. Text and vector drawings in a PDF are sharp at any zoom, which no image can match. A photo inside a PDF is exactly as good as the image that was put in: a PDF made from a quality-70 JPG is a quality-70 JPG in a container. A PDF is not a way to add quality, only to avoid losing it.

**Does converting a PDF to JPG lose quality?**

Some, in two ways. The text stops being text and becomes pixels at a fixed resolution, so it cannot be searched or zoomed without blurring. And JPG compression adds small smudges around sharp edges. Render at 300 DPI and save as PNG to keep the page looking as good as possible, or extract the embedded images if what you need is the photos rather than the pages.

**Why do offices and portals prefer PDF for documents?**

Because a PDF looks the same on every device, keeps multi-page documents in one file in order, preserves the paper size for printing, can be searched when the text is real, and can carry signatures and form fields. Images of pages can be shuffled, cropped or shrunk in transit, and several files are harder to track than one.

**Is PNG or JPG better for screenshots?**

PNG. Screenshots have sharp text and flat areas of colour, which PNG stores exactly and compresses well, while JPG blurs the text and can make the file larger. Use JPG only when a screenshot is mostly a photograph, such as a capture of a photo gallery, and size matters.

**What DPI should I scan at?**

For ordinary printed documents, 300 DPI in colour or greyscale; that is the minimum in the US National Archives’ guidance, which recommends 400 for colour and greyscale. Use 600 DPI for small print, fine drawings or records you are archiving. For a document that only needs to be read on screen and must be small, 150 to 200 DPI is enough.

**What format is best for printing?**

PDF, for documents: it carries the page size and keeps text and lines as sharp vectors. For photos, a JPG at high quality with about 300 pixels per inch at the printed size, or whatever format the print shop asks for. When printing a PDF, choose Actual size or 100 percent rather than Fit to page so nothing is shrunk.

**Is PDF/A different from a normal PDF?**

Yes. PDF/A is a restricted form of PDF, standardised as ISO 19005 for long-term archiving. It requires fonts to be embedded and forbids encryption and features that depend on anything outside the file, so it can still be displayed correctly decades later. Many office suites can save as PDF/A in their export options.

---

_Source: [Do The Calculation](https://dothecalculation.com/blog/files/pdf-vs-jpg-vs-png). Quote freely with attribution and a link to this page._