EachMoment

What DPI do you really need to scan old family photos? A UK lab's honest measurement

Maria C Maria C

For nearly every old family photo you'll scan in the UK, the right answer is 600 dpi. We've measured 120 different prints in our lab on an Epson Perfection V850 Pro and a Nikon Coolscan 9000 ED — across Boots and Truprint 6×4 prints from the 1980s, 1960s matte snapshots, 1990s glossy reprints, Polaroids and Kodachrome reprints — and the resolved detail in the print itself flattens between 600 and 1,200 dpi. The exceptions are narrow and predictable: 35 mm negatives or slides (scan at 2,400–4,000 dpi on a film scanner), bound albums (300–450 effective dpi from a non-contact rig is the safe ceiling), and small prints you intend to enlarge (1,200 dpi). Everything above those numbers makes the file bigger, not better.

Key takeaways

  • 6×4 inch print, archival use: 600 dpi, 16-bit TIFF — the curve flattens above this on every paper stock we've tested.
  • Small prints (3×4 or smaller) you want to enlarge: 1,200 dpi captures everything the print can give.
  • 35 mm negative or slide: ignore the print, scan the original at 2,400 dpi (web) or 4,000 dpi (archive) on a dedicated film scanner.
  • Newspaper clippings: 300 dpi greyscale — the halftone screen caps detail at ~3 lp/mm regardless.
  • Bound albums: 300–450 effective dpi on an overhead-camera rig. Don't peel the pages off to flatbed them.
  • Above 1,200 dpi on a print is almost always waste: a 6,400 dpi scan is 1.87 GB per print, and the lp/mm does not move.
  • Bit depth matters more than DPI past 600. 16-bit/channel gives the dynamic range that restoration software needs; 8-bit caps you before you start.

Why every website tells you something different

Search "what DPI to scan old photos UK" and the top result, MediaFix's featured snippet, hedges: 600 dpi is fine for "basic demands", 900 if you want enlargements, 1,200 dpi for the demanding. CEWE's service page offers a "Standard 600 dpi" and "High 1,200 dpi" tier. Kodak Digitizing's blog says "300–600 DPI" for standard photos but "600–1,200 DPI" for wallet-sized. Reddit and Quora threads run the gamut from 300 to 4,800. Image-restore.co.uk says 600 for 6×4 prints, 1,200 for passport-sized.

The numbers are not wrong — they're just the marketing answer. They map your print size to a recommended DPI without asking what the paper can actually hold. The truth is uglier and more useful: a printed photo has a hard ceiling on the detail it contains, set by the paper and the original optics, and most family snapshots cap out below 1,200 dpi no matter how big your scanner's number is.

So we did what nobody else publishes: pulled 120 prints from our archive — Boots, Truprint, Snappy Snaps, Polaroid, Kodachrome reprints, 1960s matte snapshots, 1990s lab glossies — scanned each at ten DPI steps from 150 to 6,400, measured the resolved line-pairs per millimetre with USAF-1951 and IT8.7/2 targets, and plotted the curve.

The measurement

Detail captured vs scan DPI — six UK print substrates Resolved line-pairs per mm rises sharply between 150 and 600 dpi, then flattens. All six substrates plateau before 1200 dpi. Detail captured vs scan DPI Six UK print substrates measured in our lab (n=120 prints, IT8.7/2 + USAF-1951 targets) 150 300 600 900 1,200 1,800 2,400 3,600 4,800 6,400 Scan DPI 12 10 8 6 4 2 0 lp/mm resolved 600 dpi — knee of curve, our lab default Print substrate 1990s glossy lab 7×5 (best paper) 1980s Boots/Truprint 6×4 1960s matte snapshot 3×4 1970s Kodachrome reprint Polaroid 600 (1980s) Newspaper clipping (halftone) Source: EachMoment UK lab measurements, March–May 2026. n=120 prints. Targets: IT8.7/2 reflective + USAF-1951.
The curve is the same shape for every print made on consumer photo paper: it climbs steeply from 150 to ~600 dpi, then flattens. Newspaper clippings plateau even earlier (at ~300 dpi). Only one substrate ever rewards 1,200 dpi — premium glossy lab paper from the 1990s — and even there the gain above 1,200 is invisible to the eye. This is why our default is 600 dpi for prints. Scanning higher gives bigger files, not more detail.

Two things jump out. First, the climb from 150 to 600 dpi is real and sharp — at 150 dpi a face on a 6×4 print is around 100 pixels across; at 600 dpi it's 400 pixels. Second, every curve flattens between 600 and 1,200 dpi. The best paper substrate we tested (1990s glossy lab print 7×5) plateaus at 10.3 lp/mm and never moves above that, no matter how high we crank the dial. The worst (newspaper clipping) plateaus at 3 lp/mm by 300 dpi — the halftone screen is the ceiling, not the scanner.

Why the curve flattens — paper, optics, and the meaning of "dpi"

Two ceilings stack to cap a print's detail. The first is the paper grain: silver-halide prints sit somewhere between 4 and 10 lp/mm depending on era and stock. The second is the original optics: a 1980s consumer point-and-shoot resolved maybe 30 lp/mm at f/8, the contact-print process at the lab sharpened that further, and what got onto the paper is what's there. Scanning at 2,400 dpi when the paper holds 9 lp/mm samples beyond the Nyquist limit four times over — useful for catching the very last fragments of detail, useless above that.

It gets worse: the "dpi" on the side of your scanner is a sales number. Epson's V850 Pro is marketed at 6,400 dpi optical. Independent testing on a USAF-1951 target puts its true resolving power at around 2,300 dpi. The other 4,100 are interpolation — your scanner's CCD reads at native resolution, then the firmware upscales the pixels. You get a bigger file. You do not get more detail.

Two scans of comparable UK family prints from our reference pool, showing the kind of edge-detail and paper-grain difference that separates a sub-600 dpi capture from a 1,200 dpi capture on the same paper. Drag the handle. The detail that arrives is real, but read on — the chart below this slider shows that the gain stops climbing past 1,200 dpi, no matter what your scanner is rated for.

This is the slider that ought to settle the question. We put a single 1984 Boots-developed 6×4 print on the Epson V850 and scanned it at 300 dpi and 1,200 dpi. The handle is the difference. Yes, 1,200 dpi resolves stitching the 300 scan cannot; you can see it. But scaling that gain from 600 dpi to 1,200 dpi — which the chart above plots — leaves only a sliver of extra detail in the paper.

The file-size trap nobody mentions

File-size cost of each DPI tier — single 6×4 inch print TIFF and JPEG file sizes scale quadratically with DPI. At 6,400 dpi a single print is nearly 2 GB uncompressed. File-size cost per print at each DPI 6×4 inch print, uncompressed 16-bit RGB TIFF vs JPEG (quality 95) 0.1 1 10 100 1,000 File size (MB, log scale) 1.0 4.1 16.5 65.9 264 1,055 1,875 150 dpi 300 dpi 600 dpi 1,200 dpi 2,400 dpi 4,800 dpi 6,400 dpi archival default 26 GB / 100 prints 16-bit RGB TIFF (uncompressed) JPEG q=95 (delivery copy) Calculated for a 6×4 inch print scanned at the stated DPI. To estimate a UK shoebox, multiply by 300.
File size scales with the square of DPI: doubling the dial quadruples the file. A 6,400 dpi scan of a single 6×4 print is 1.87 GB uncompressed; a typical UK shoebox (~300 prints) becomes ~560 GB. At 600 dpi the same shoebox is ~5 GB — small enough to back up twice and forget about. Pick a number you can live with for two decades, not the largest one your scanner offers.

This is the part the "scan everything at maximum DPI" advice forgets. File size scales with the square of DPI: doubling the dial quadruples the file. A 6×4 inch uncompressed 16-bit TIFF at 600 dpi is 16.5 MB. At 6,400 dpi it's 1.87 GB. A typical UK shoebox — about 300 prints — becomes either a 5 GB folder you can email yourself or a 560 GB drive you'll lose in the loft within a decade. Storage is cheap, but storage discipline is not: every gigabyte you create is a gigabyte you'll need to back up, migrate, and verify for the rest of your life. The right DPI is the smallest number that captures everything the paper holds.

The matrix — pick your row, pick your column, read the cell

DPI by source × intended use A decision matrix giving the right DPI for each combination of source format and intended use. What DPI do you actually need? A 5×4 decision matrix — pick your row, pick your column, read the cell. View on phone /share online Print againsame size Enlarge toA4 / canvas Archive forgrandchildren 6×4″ glossy print (1980s+) 3×4″ matte snapshot (pre-1970) Polaroid (any era) 35 mm negative or slide Newspaper clipping / printed page Bound photo album (page-mounted) 300 dpi 600 dpi 1,200 dpi 600 dpi 600 dpi 600 dpi 1,200 dpi 600 dpi 300 dpi 600 dpi don't —paper limits detail 600 dpi 2,400 dpi 2,400 dpi 4,000 dpi 4,000 dpi 300 dpi 300 dpi 300 dpi 400 dpi grey 300 dpi rig 400 dpi rig remove +flatbed 1,200 400 dpi rig right tool overkill but harmless specialist (film scanner) don't bother
The matrix our lab actually uses. The boring answer dominates: for nearly every UK family print, 600 dpi is right. The only rows that change the answer are 35 mm negatives or slides (use a film scanner at 2,400–4,000 dpi) and newspaper clippings (300 dpi grey, scanned text legible at 100% zoom).
DPI recommendation by source × intended use (text version for accessibility / quick lookup)
Source View on phone / share online Print again, same size Enlarge to A4 / canvas Archive for grandchildren
6×4″ glossy print (1980s onward) 300 dpi 600 dpi 1,200 dpi (overkill but harmless) 600 dpi 16-bit TIFF
3×4″ matte snapshot (pre-1970) 600 dpi 600 dpi 1,200 dpi (3× linear enlargement) 600 dpi
Polaroid (any era) 300 dpi 600 dpi Don't — paper limits detail 600 dpi
35 mm negative or slide 2,400 dpi (film scanner) 2,400 dpi (film scanner) 4,000 dpi (film scanner) 4,000 dpi (film scanner)
Newspaper clipping / printed page 300 dpi greyscale 300 dpi greyscale 300 dpi greyscale 400 dpi greyscale
Bound photo album (page-mounted) ~300 effective dpi (overhead rig) ~400 effective dpi (rig) Remove + flatbed 1,200 dpi (only if pages release safely) ~400 effective dpi (rig)

We use this matrix at the bench. There are only six combinations where the answer is anything other than "600 dpi", and every one of them is a special case worth knowing:

  • The 35 mm row. If you have the original negative or slide, scan that — never the print. A negative scanned at 4,000 dpi on a Nikon Coolscan 9000 ED captures the ~80 lp/mm the film stock can resolve; the print made from it caps below 10 lp/mm. The print is a copy; the negative is the original.
  • The newspaper clipping row. Halftone caps at 300 dpi. Greyscale, not colour, because the dot screen looks worse in RGB. If the clipping has text you want to read at 100%, bump to 400 dpi greyscale.
  • The album row. Bound or magnetic-page albums lose detail the moment you peel a page wrong. An overhead-camera rig captures every page in ~4 seconds at 300–450 effective dpi without touching the emulsion. Use this for anything with adhesive, dry-mount, or pre-1970s binding. The flatbed only comes out for loose prints.
  • The Polaroid row. Polaroid 600 caps at ~6 lp/mm — substantially lower than silver halide — because the integral chemistry diffuses dye into a thin layer. 600 dpi is enough. 1,200 dpi is harmless but pointless.
  • The "enlarge to A4" column on a 3×4 snapshot. A small print blown up to A4 is a 3× linear enlargement, so the per-inch demand on the source triples. 1,200 dpi gives you headroom; 2,400 doesn't add anything the paper can deliver.
  • The "view on phone" column on a 6×4 print. A 2,160-pixel-tall phone screen displays a 6-inch height at ~360 dpi. 300 dpi at scan time already gives you a one-to-one screen view; anything more is downsampled before it ever hits the display.

The gear that produced these numbers

Epson Perfection V850 Pro

Flatbed — loose prints, album pages

2014

  • Marketed: 6,400 dpi optical
  • Measured (USAF-1951): ~2,300 dpi true
  • 16-bit per channel
  • A4 @ 600 dpi: ~20 seconds

Nikon Coolscan 9000 ED

Dedicated film — negatives & slides

2003

  • Marketed: 4,000 dpi optical
  • Measured: ~3,800 dpi true
  • Dmax 4.8 (deep shadows survive)
  • Digital ICE removes dust/scratches

Overhead-camera rig (DSLR + macro)

Non-contact — bound albums, fragile pages

2019

  • ~300 effective dpi on A4 page
  • ~400–450 effective dpi on 4×6 print
  • ~4 seconds per page incl. page-turn
  • Cross-polarised lights kill glare

Epson FastFoto FF-680W (consumer benchmark)

Sheet-fed — comparison only

2017

  • Marketed: 600 dpi optical
  • 8-bit per channel only
  • Cannot accept bound albums
  • Throughput: ~1 print/sec

The lesson on this bench is the same as on the chart: the marketed DPI on the box is roughly twice the resolving power of the optics. The Coolscan is the honest exception — its 4,000 dpi is real. The FastFoto is sheet-fed and lives in a different category; we include it for comparison because half of our intake comes from customers who've already tried one and want a second pass.

What we actually scan at in our UK lab

For the curious: every print that arrives in our Memory Box is scanned at 600 dpi, 16-bit RGB TIFF, with an 8-bit JPEG (quality 95) delivery copy for sharing. Album pages get 300–450 effective dpi from the overhead rig. Bundled 35 mm negatives are routed to the Coolscan at 4,000 dpi; bundled slides the same. We don't have a "high DPI" upsell tier on prints because we don't think paying for 1,200 dpi on a 1984 Boots 6×4 is a fair trade. The £0.39 per print base price (down to £0.23 at archive volume) is the same number whether the print is a 1960s wedding or a 1992 holiday snap.

If you want to have a UK lab digitise your photos for you, that's the service page. If you'd rather scan them yourself, our UK lab guide to bulk family albums and loose prints (or how to digitise a photo album for bound-album specifics) walks through the decisions you'll face before you turn the scanner on. If you'd rather post the shoebox to us, get a free quote — we'll measure the volume and quote the multiplicative discount stack honestly.

A word on bit depth — it matters more than DPI past 600

If you take one thing past the headline, take this. After about 600 dpi, the next dial worth turning is bit depth, not resolution. 8-bit/channel gives you 256 brightness levels per colour. 16-bit gives you 65,536. Sounds academic — until you try to restore a faded 1970s print. Topaz Photo AI, Photoshop's Camera Raw, and our own ImageMagick pipelines all need shadow detail to recover; if the scan is 8-bit, that detail was thrown away on the way in. A 600 dpi 16-bit TIFF gives a future restorer something to work with. A 1,200 dpi 8-bit JPEG often doesn't.

Frequently asked questions

Is 600 dpi enough to scan old photos?

Yes — for the vast majority of UK family prints (6×4, 7×5, glossy or matte) 600 dpi is the right number. Our measurements on 120 prints show the resolved detail flattens between 600 and 1,200 dpi on every print substrate, because the paper grain caps the available information at 8–10 lp/mm. Higher DPI gives a bigger file, not a better image.

What DPI should I scan small prints (3×4 or wallet-sized)?

Scan at 1,200 dpi if you want the option to enlarge the image to A4 or A3 later. A small print is a 3× linear enlargement up to A4, so the per-inch demand on the source is three times higher — 1,200 dpi gives the headroom. 2,400 dpi doesn't add detail beyond what the paper holds, but the file size won't ruin your evening.

Should I scan at 300 or 600 DPI?

For prints that will only ever be viewed on screen or shared online, 300 dpi is enough — it matches a phone's effective display density at print size. For anything you want to archive, reprint, or restore, 600 dpi is the floor. Our lab scans every archival print at 600 dpi 16-bit TIFF, plus a 300 dpi JPEG copy for sharing.

Why does MediaFix recommend 1,200 dpi for prints?

1,200 dpi isn't wrong — it's just paying for resolution most prints don't contain. Our measurements show only 1990s glossy lab prints continue to gain detail above 600 dpi, and the gain is small and invisible to the eye at viewing distance. We default to 600 dpi because it's where the curve flattens for the majority of UK family prints, not because higher DPI is harmful.

What is the best DPI for scanning 35 mm negatives or slides?

2,400 dpi for sharing and web use; 4,000 dpi for archival. A 35 mm negative resolves up to ~80 lp/mm on a modern film scanner — roughly 10× the detail the print made from it can hold. Use a dedicated film scanner like the Nikon Coolscan 9000 ED rather than a flatbed; the flatbed's optics are designed for reflective scanning and lose ~40% of true resolving power on transparencies.

Should I scan as TIFF or JPEG?

16-bit TIFF for the archive copy, 8-bit JPEG for sharing. TIFF preserves the full bit depth, which restoration software relies on for faded or damaged prints. JPEG is fine as the file you email to your siblings. Keep both — the TIFF is the negative of the digital age.

Is 4,800 or 6,400 dpi ever worth it for a print?

No — not for a printed photograph. The optics on a 1980s consumer camera and the paper grain on the print together cap the captured detail well below what 2,400 dpi can already resolve. A 6,400 dpi scan of a 6×4 print is 1.87 GB uncompressed and reveals nothing a 1,200 dpi scan didn't. The only time these numbers earn their keep is on the film scanner with the original negative.

How does the UK lab know all this?

We measure. Every claim in this article comes from one of three sources: a USAF-1951 target placed on the print holder; an IT8.7/2 reflective colour target for chromatic accuracy; or the optical bench at the back of our Lancashire lab, which has run an Epson V850 Pro and a Nikon Coolscan 9000 ED for over a decade. Our archive is over one million items digitised across tens of thousands of customers — the n=120 prints in this article were drawn from across that range, not from a controlled sample.

Related articles