Confused by DPI, bit depth, and TIFF vs JPEG? This plain-English reference tells you exactly what to choose and why.
You have dug out the old flatbed scanner, gathered a stack of faded family prints, and opened the scanner software — only to be confronted with a bewildering grid of settings: 300 DPI, 600 DPI, 1200 DPI, 24-bit colour, 48-bit colour, TIFF, JPEG, BMP, unsharp mask, dust removal, auto-exposure. Which ones actually matter? Which ones do you simply leave alone? This guide strips away the jargon and gives you a definitive, practical answer for every setting.
DPI stands for dots per inch — in scanning, it means how many pixels the scanner captures for every inch of the original print. A 4×6 inch print scanned at 300 DPI produces an image 1200×1800 pixels. The same print at 600 DPI produces 2400×3600 pixels — four times as many pixels and roughly four times the file size.
More pixels sounds better, but there is a point of diminishing returns. Most consumer flatbed scanners have an optical resolution of 600–1200 DPI. Beyond their optical limit, the scanner uses interpolation (mathematical guessing) to generate additional pixels — which adds file size without adding real image detail. The key insight: matching DPI to your intended use, not maxing it out, gives you the best result.
Fine for photos you will only ever view on screens or share via messaging apps. Files are very small (100–400 KB each).
Not recommended for archiveThe sweet spot for most photo scanning. Produces sharp, detailed files that print beautifully up to 4×6 at original size. 1–3 MB per scan.
✓ Best for most usersUse for small prints (wallet size, passport photos) or when you need to enlarge prints beyond their original dimensions. 4–10 MB per scan.
Use for small printsOnly meaningful for scanning film negatives, slides, or very small miniature prints. Produces enormous files (20–80 MB) with no benefit for standard prints.
Avoid for prints💡 Rule of thumb: Scan standard 4×6 or 5×7 prints at 300 DPI. Move to 600 DPI only when the original print is smaller than 3×3 inches (wallet photos, passport photos, strip photos). Never exceed 600 DPI for standard consumer prints — you are just generating scanner noise, not additional detail.
Your scanner software will offer several colour modes. Here is what each one means:
| Mode | Bits per Pixel | Colours Captured | Use For |
|---|---|---|---|
| 24-bit Colour (RGB) | 8 bits per channel | 16.7 million | All standard photo scanning ✓ |
| 48-bit Colour (RGB) | 16 bits per channel | 281 trillion | Professional film scanning only |
| Greyscale (8-bit) | 8 bits total | 256 shades of grey | Line art, text documents |
| Black & White (1-bit) | 1 bit total | 2 (black or white) | Pure text documents only |
The most important rule here: always scan in 24-bit Colour (RGB), even for black-and-white photographs. Here is why: vintage black-and-white prints are rarely truly neutral — they often have a warm sepia tone, a cool silver-blue cast, or subtle colouration from the paper. Scanning in greyscale permanently discards all that tonal information. Scanning in 24-bit colour preserves it. The file size difference is minimal (a 24-bit grey scan is only slightly larger than an 8-bit greyscale scan) and the preservation quality is meaningfully better.
Skip 48-bit colour for standard photo scanning. While the extra bit depth is valuable for film scanning and professional retouching workflows, it doubles file sizes and most photo viewers cannot display the full tonal range anyway. 24-bit is the universal standard for photo archives.
This is the question that generates the most debate among digitization enthusiasts. Here is an honest, practical breakdown:
JPEG uses lossy compression — it discards some image data to achieve small file sizes. At quality settings of 85% or above, the difference from the original is virtually invisible to the human eye. A 300 DPI scan of a 4×6 print saved as JPEG 90% will be about 1–3 MB. JPEG is the best choice for photo archives where you want a balance of quality, compatibility, and storage efficiency. Every device, app, cloud service, and photo viewer on earth understands JPEG.
TIFF uses lossless compression (or no compression at all) — every pixel is preserved exactly. A 300 DPI scan of a 4×6 print saved as TIFF will be 15–25 MB — roughly 10–15× larger than the JPEG equivalent. TIFF is the professional standard for archival master files. If storage space is not a concern and you want the absolute pixel-perfect preservation of each scan, use TIFF. If you are scanning hundreds of photos and will be storing them on cloud services with storage limits, JPEG at 90%+ is the pragmatic choice.
PNG is a lossless format like TIFF but is generally not used for photo scanning — PNG files are typically larger than JPEG but smaller than TIFF, and they lack the rich metadata support that TIFF provides for professional workflows. PNG is excellent for screenshots, graphics, and web images. For scanned photographs, choose JPEG or TIFF instead.
If storage permits, save both formats. Use TIFF as your archival master — put these on an external drive. Use JPEG 90% as your working copy — upload these to Google Photos, share with family, and use for everyday viewing. If you ever need to reprocess a photo (crop it differently, do heavier colour correction), you have the lossless TIFF to work from. For most home archivists who are storage-constrained, JPEG at 90% is perfectly sufficient and will look indistinguishable from the TIFF in everyday use.
Auto-exposure seems helpful — it automatically adjusts the brightness so your scan looks good. The problem is that it adjusts per scan, meaning a batch with one dark photo and one light photo may have inconsistent brightness from one scan to the next. More critically, auto-exposure can clip (permanently overexpose) highlights in bright prints or crush shadows in dark ones. Turn it off and apply any brightness corrections later in a photo editor where you have full control.
Auto-crop attempts to automatically detect and crop individual photos from a batch scan. Scanner software auto-crop is notoriously unreliable — it frequently misses photos, merges adjacent ones, or crops them incorrectly. Turn it off entirely and use PhotoSplit Studio instead. PhotoSplit's OpenCV.js detection engine is vastly more accurate than any scanner driver's auto-crop and works across all scanner brands with a single consistent interface.
Technologies like Epson's Digital ICE and Canon's FARE use infrared scanning to detect and digitally fill scratches and dust. These work well for film scanning (negatives and slides) but should be turned off for reflective print scanning. On prints, dust removal tends to soften fine details, blur texture in the photo paper, and can incorrectly identify image content as scratches. Clean the scanner glass with a microfiber cloth instead — that eliminates 95% of dust-related issues without any image quality compromise.
In-scanner sharpening applies a sharpening filter to the scan during capture. This is destructive — it permanently bakes a sharpening artifact into the file. Turn it off and apply sharpening non-destructively in your photo editor or when resizing for a specific output (print, web, etc.) after the fact.
If you have 35mm or medium format film negatives to digitize, the settings above change significantly. Negatives require a scanner with a transparency adapter (a backlit scanning bed) since the image is captured from light transmitted through the film rather than reflected off a print surface. For film scanning you should use 1200–2400 DPI (the entire frame is only 35mm wide), 48-bit colour is beneficial, and Digital ICE or FARE can be enabled. PhotoSplit Studio's Convert Negative feature then handles inverting your scanned negative frames to produce natural positive photographs — including automatic orange mask removal for colour C-41 negative film.
🚀 Ready to split your scans? Once you have your batch scan files, drag them into PhotoSplit Studio to automatically detect, crop, and save every individual photo in seconds — free, private, and right in your browser.