Guide 06 — Digitising
Scan them yourself, or pay someone?
This is a genuine arithmetic question with a genuine answer, and it is almost never the answer people expect. It depends on three numbers: how many photographs you have, what resolution you actually need, and what an hour of your time is worth to you on a Sunday.
Three ways, and what each is really for.
| Route | Speed | Quality | Use it for |
|---|---|---|---|
| Phone camera | Fastest — 15–30 seconds each | Fair, with perspective correction and flat light. Poor for small prints and glossy surfaces | Contact-sheet records, sharing, triage, albums you cannot disassemble |
| Home flatbed scanner | 60–90 seconds each including handling | Best available for prints | Small prints, anything you will reprint, your final archive copies |
| Mail-in or shop service | Days to weeks, plus shipping both ways | Consistent and high, automated dust removal, handles batches | Large collections, negatives, slides, anything delicate |
The important thing about the phone route is not that it is worse. It is that a phone capture is a photograph of a photograph, not a scan, and the difference matters later when you try to restore from it. More on that in section 05.
Five hundred photographs, three ways.
Take a realistic box: 500 prints. Here is what each route actually costs, counting your time at something honest — say $25 an hour, which is well below most professional rates and generous to the DIY case.
| Route | Time | Cash | Total realistic |
|---|---|---|---|
| Phone, 25s each | ~3.5 hours | $0 | About 3.5 hours, and a mediocre archive |
| Flatbed, 75s each incl. handling and editing | ~10.5 hours | Cost of a decent scanner, often $100–$300 | Two or three weekends, plus the hardware |
| Mail-in service, roughly $0.30–$1.50 per print | An hour of packing | Roughly $150–$750 | About an hour, plus shipping risk and waiting |
The crossover is roughly 100 to 150 photographs. Below that, DIY is usually right because the setup cost dominates. Above that, the arithmetic starts favouring a service — not because your time is more valuable than money, but because almost nobody actually finishes a ten-hour scanning project. The failure mode of DIY scanning is not slowness. It is abandonment at photograph 200, which leaves you with a half-scanned collection and a scanner in a cupboard.
Honest adjustment: if you enjoy the work, or you have a scanner already, or the collection is small and precious, DIY wins on more than arithmetic. The numbers above assume you are scanning unwillingly.
The DPI you actually need, by purpose.
Most people scan at the wrong resolution in one of two directions: 300dpi for everything and then discover they cannot reprint, or 4800dpi for everything and fill a hard drive with files nobody will ever open. Match the resolution to the job.
| Purpose | DPI | Notes |
|---|---|---|
| Screen, sharing, family group chats | 300 | Plenty. Do not bother higher |
| Reprint at original size | 600 | The sensible default for a family archive |
| Enlargement, or small prints (under 4×6in) | 1200 | Small originals carry fine detail worth capturing |
| 35mm negatives and slides | 2400–3200 | A 35mm frame is tiny; it needs the resolution |
| Large format negatives, glass plates | 1200–2400 | The physical area is already large |
Those are working numbers for a family collection. The formal benchmark is higher and worth knowing: US federal regulation 36 CFR 1236.50 sets a floor of 400 ppi for photographic prints, mapped to the FADGI three-star category, and archival practice for 35mm film sits above the range in this table. Scan above the floor if the time is available; scan at the floor rather than not at all.
Scan as a TIFF or a high-quality JPEG, not as a screenshot, and never resize down before archiving — the small version is easy to make later and impossible to recover.
Negatives and slides are a different job entirely.
This is where DIY usually fails and where paying is usually correct. A negative is a transparent original: it needs light passed through it, which means a transparency adapter, a dedicated film holder, and usually software that inverts and colour-corrects the result. A flatbed without a transparency unit will produce a black rectangle.
Two things follow. First, if you have the negatives, they are frequently a better source than the prints — they are first-generation, usually less faded, and often include frames that were never printed. If a family collection contains a shoebox of negatives alongside the prints, that shoebox is the more valuable object and most people do not know it.
Second, film scanning rewards proper equipment. A service with a dedicated film scanner will beat a home flatbed on 35mm by a wide margin, and the cost per frame is low because the handling is fast. For anything more than a handful of strips, pay.
A photo of a photo is not a scan.
The most consequential mistake in this whole process happens before any restoration. Someone holds a phone over a print, takes a picture, and that file becomes the thing everything else is built on. It has four problems baked in that no amount of restoration can remove:
- Perspective. The print is never perfectly square to the lens, so the image is a trapezoid. Correction helps but resamples every pixel.
- Glare and reflection. A glossy print lit by a window or a ceiling bulb has a hotspot over part of the image — often over a face. The information there is gone.
- Curvature. Old prints are rarely flat. A curled print has parts out of focus that no sharpening recovers.
- Moiré and softness. The camera resolves the paper texture and the print's own dot structure, and the result is a soft, patterned image at a fraction of the detail a scanner captures.
So: if the only version you have is a photo of a photo, the correct first step is not restoration. It is getting a proper scan. Restoration from a bad capture is expensive, and the result will be invented detail layered on missing detail. Our triage tool will tell you this outright if you select it, because it is the most common way people waste money on restoration.
If a proper scan is genuinely impossible, the acceptable middle ground is a phone capture done properly: the print flat on a table, camera parallel and directly above, diffuse daylight from a window with no direct sun, flash off, and a scanning app that handles perspective and colour. It will not equal a flatbed, but it is not a snapshot either.
The short version.
Pay a service when: you have more than about 150 photographs; there are negatives or slides; the albums cannot be disassembled; the material is delicate, mouldy or pre-1900; you have a deadline — a funeral, an anniversary, a birthday; or you know you will not finish it yourself.
Do it yourself when: the collection is under about 100 prints; you already own a scanner; you only need screen-resolution copies; or you want to handle the prints yourself because they are fragile or you do not want them to leave the house.
Never: send irreplaceable originals by post without insurance and tracking, or accept a "we scan it and keep the originals" arrangement from an unclear provider. Ask what happens to the paper and get it in writing.
Once you have the files, the rest of the work is sorting and storage — and if some of what you scanned is damaged, here is what restoring it costs.
Questions
The things people ask next.
Is it cheaper to scan old photos myself or pay a service?
What DPI should I scan old photos at?
Can I scan old photos with my phone?
How do I scan negatives and slides at home?
Should I send my only copies to a scanning service?
Where these figures come from
Not our estimates. If a number on this page contradicts one of these, the number is wrong and we will fix it.
- US National Archives and Records Administration — 36 CFR 1236.50: Requirements for digitizing permanent paper and photographic print recordsA 400 ppi minimum for photographic prints, the FADGI three-star benchmark
Scanned, and now wondering whether to restore?
Most damaged photographs are not worth a professional restoration — and some are worth far more than you would expect. Thirty seconds to find out which.