Editor › Learn › Choosing the Right Image Format: TIFF, PSD, PNG, JPEG, WebP, PDF
Which image format should I use?
Match the format to the stage. Working master with layers: PSD (or layered TIFF). Print delivery: flattened TIFF with the profile embedded, or a print-ready PDF. Web photographs: JPEG (or WebP) in sRGB. Graphics with sharp edges or transparency: PNG. The classic mistakes are re-editing JPEGs (each save re-compresses), delivering print work in sRGB screen formats, and archiving masters in lossy formats.
The one question that decides most cases
Is the file a MASTER you will edit again, a PRINT deliverable, or a WEB deliverable? Masters need structure and losslessness: PSD or layered TIFF, profile embedded, kept forever. Print deliverables need exact pixels and correct colour meaning: flattened TIFF or print PDF in the destination space. Web deliverables need small size at acceptable quality: JPEG or WebP for photographs, PNG for interface graphics — in sRGB, because the unmanaged web assumes it.
Format properties that actually matter
| Format | Compression | Layers | Transparency | CMYK | 16-bit | Best at |
|---|---|---|---|---|---|---|
| PSD | lossless | yes — full structure | yes | yes | yes | working master, exchange of layered work |
| TIFF | lossless (LZW/Deflate) | optional | yes | yes | yes | print delivery, scans, archives |
| PNG | lossless | no | yes (alpha) | no | 16-bit exists, support varies | UI graphics, screenshots, sharp edges |
| JPEG | lossy | no | no | rare/legacy | no | photographs on the web |
| WebP | lossy + lossless modes | no | yes | no | no | web images where supported |
| Print PDF | wraps images | no (flattened) | n/a | yes | per content | final print hand-off |
The mistakes each row prevents
- Editing a JPEG repeatedly: every save re-runs lossy compression; artefacts compound. Convert to a lossless working format at first edit.
- Sending PNG or JPEG to a print house for colour-critical work: no CMYK, sRGB assumptions, and JPEG's artefacts sit exactly on the sharp edges print magnifies.
- Archiving only the flattened export: the day a client wants "the same but with the other headline", the master is the layers.
- Screenshot text saved as JPEG: block artefacts around every glyph. PNG exists for exactly this.
- Shipping wide-gamut files to the web: colour-unaware viewers oversaturate them. Deliver sRGB.
Where each format enters a print job
A typical professional flow touches four of these on purpose: camera or scan intake (often TIFF), working master (PSD or layered TIFF, RGB, profile embedded), print deliverable (flattened TIFF or print PDF, converted to the printing condition), and web derivatives (sRGB JPEG/WebP exported last, from the master, never the other way around). Each hand-off is one export from the master — the master itself never leaves the lossless world.
Try it in PhotoAIA
- Open one layered document and export it as TIFF, PNG and JPEG.
- Compare file sizes, then zoom to 800% on a sharp text edge in each export.
- Re-save the JPEG export twice more at the same quality and watch the edge artefacts compound.
- Reopen the TIFF: identical pixels — that difference is the whole lossless argument.
Questions
- Is TIFF better quality than PNG?
- Both are lossless — identical pixels survive either. They differ in ecosystem: TIFF speaks print (CMYK, 16-bit, profiles, multi-page); PNG speaks web/UI (alpha, universal browser support).
- Why do printers dislike JPEG even at maximum quality?
- Maximum-quality JPEG is visually clean on screen, but its 8×8 block artefacts and chroma subsampling sit on the sharp, high-contrast edges print reproduces mercilessly — and every resave compounds them. Lossless costs little at print resolutions.
- Should I use WebP for everything on the web?
- For photographs it typically beats JPEG at the same visual quality. Keep PNG for graphics needing perfect edges, and keep your masters lossless regardless of what the site serves.