How to Enlarge Old Family Photos Without Making Them Look Worse
A practical workflow for scanning, cleaning and upscaling old prints and early digital photos so they can be printed larger or shared on modern screens.
By the ClearPix team · · 5 min read
Most families have a box of small prints, a few faded Polaroids and a folder of photos from a phone that was new fifteen years ago. These pictures matter, but they were never made for a 4K screen or a framed print on the wall. Blow one up and it turns soft, blotchy and grainy. The good news is that a careful workflow, done in the right order, gets far better results than simply dragging the corners of an image in an editor.
This guide walks through that workflow step by step. It applies whether you are working from paper prints or from early digital files, and it explains where AI upscaling genuinely helps and where it cannot.
Step 1: Get the best possible source
Every later step depends on the quality of the file you start with. An upscaler can rebuild edges and texture, but it works from what is in the file. A better source always beats a better tool.
If you have the paper print
- Use a flatbed scanner if you can. Phone "scanner" apps are convenient, but a flatbed gives even lighting, no perspective distortion and no glare.
- Scan at 600 dpi for normal prints (for example 10 × 15 cm or 4 × 6 in). For very small prints such as wallet photos or contact sheets, use 1200 dpi. Scanning higher than the print's real detail only makes bigger files, not better ones.
- Wipe the glass and the print gently with a microfibre cloth first. Dust shows up as white specks that an upscaler will happily make bigger.
- Save the scan as PNG or TIFF, not JPG. You want to avoid adding compression before you have even started.
If you only have a phone photo of the print
Photograph it in daylight near a window, not with the flash, and hold the phone exactly parallel to the print to avoid a keystone shape. Take several shots and keep the sharpest one. Crop away the table and the print's border before doing anything else.
If the photo was digital from the start
Look for the original file rather than a copy that was shared through a messaging app or social network. Those services shrink and recompress images heavily. The original from an old memory card, laptop backup or cloud archive is often two to four times larger and much cleaner.
Step 2: Decide how big the result really needs to be
Before upscaling, work out the target size from where the photo is going. This tells you which upscale factor to use and stops you from enlarging further than the image can support.
| Destination | Pixels needed (approx.) | Typical factor from a 1000 px source |
|---|---|---|
| Sharing in a chat or on social media | 1080 – 2048 px on the long side | 2x |
| Full-screen on a laptop or TV slideshow | 1920 – 3840 px | 2x – 4x |
| 10 × 15 cm (4 × 6 in) print at 300 dpi | 1800 × 1200 px | 2x |
| 20 × 30 cm (8 × 12 in) print at 300 dpi | 3600 × 2400 px | 4x |
| A3 poster viewed from a distance | about 3500 – 5000 px | 4x |
If you want the arithmetic behind the print numbers, our guide to print size, DPI and pixels explains it in detail.
Step 3: Upscale with the right settings
Open the AI Image Upscaler and drop in your scan or file. Three settings matter for old photos:
- Factor: choose the smallest factor that reaches your target size. 2x looks more natural than 4x, and 4x looks more natural than 8x. If the result is still too small, it is usually better to upscale a cleaner source than to push the factor higher.
- Denoise: old film grain and JPG blocks both read as "noise" to the network. For scanned prints, 30–50% is a good starting range. For early digital photos saved at low quality, try 40–60%. Push it too far and faces start to look waxy, so compare with the before/after slider.
- Colour boost: faded prints often benefit from 15–25%. Leave it at 0–10% for black-and-white photos, since there is little colour to lift and it can add a slight tint.
Step 4: Know what upscaling cannot fix
AI upscaling predicts plausible detail when an image is enlarged. It is very good at sharp edges, hair, fabric and foliage. It is not magic, and it helps to know its limits before you spend an evening on one picture.
- Out-of-focus or motion-blurred photos improve only slightly. The detail was never captured, so there is nothing to reconstruct.
- Tiny faces in a group photo (say, under 30 pixels across) may come out smooth or slightly generic. Crop the group into smaller sections and upscale those separately if one person matters most.
- Tears, folds and stains are not removed. Repair those in a photo editor before upscaling, otherwise the damage is enlarged too.
- Upscaling does not recolour black-and-white photos. Colourisation is a different kind of AI tool.
Step 5: Save and print the result properly
Download the upscaled image as PNG to avoid another round of compression. If a print shop or website insists on JPG, convert it with the PNG to JPG converter at 90% quality or higher. That keeps the file size reasonable without visible loss.
When ordering prints, choose the paper size first and check that your image meets the pixel guide in Step 2. Matte paper is more forgiving than glossy for older photos, because it hides a little softness and grain.
A quick checklist
- Find or scan the best original you can (600 dpi scan, PNG).
- Clean up dust, tears and crops before enlarging.
- Pick a target size, then the smallest factor that reaches it.
- Start with denoise 30–50% and colour boost 15–25%, then compare.
- Download as PNG and only convert to JPG at high quality if required.
Done this way, even a small, tired print can become a clean digital copy that looks good on a modern screen and holds up in a frame. It is also a good moment to back the scans up in two places, so the next generation does not have to start from the shoebox again.