How to Restore and Animate Old Photos: A Complete Workflow

Why Old Photos Deserve More Than a Drawer

Somewhere in your home, there is a box of old photographs. Maybe they are in an attic, tucked inside a Bible, or pressed between the pages of a forgotten album. These photos represent real people — your great-grandparents, distant cousins, ancestors whose names you may not even know. And most of them are fading.

According to the Library of Congress, photographic prints from the early 20th century have a lifespan of 50 to 100 years before significant degradation occurs. Color photos from the 1960s through the 1980s fare even worse, with dye-based prints losing vibrancy in as little as 30 years.

The good news: AI has made it possible not only to restore these photos but to animate them — to see your ancestors blink, smile, and turn their heads as if captured on video. This guide walks you through the complete pipeline, from scanning a damaged photo to watching it come alive.

The Complete Workflow: Scan, Restore, Enhance, Animate

Think of the process as four distinct stages. Each one builds on the last, and skipping a step usually produces noticeably worse results.

Step 1: Scanning the Original Photo

The quality of your scan determines the ceiling for everything that follows. A blurry, low-resolution scan cannot be fully rescued by AI — garbage in, garbage out.

Best practices for scanning old photos:

• Use a flatbed scanner at 600 DPI minimum. For prints smaller than 4x6 inches, go to 1200 DPI.
• Clean the scanner glass with a microfiber cloth before every session.
• Handle photos by the edges only. Oils from your fingers accelerate degradation.
• Scan in TIFF format if your scanner supports it. TIFF is lossless, while JPEG introduces compression artifacts.
• If the photo is behind glass in a frame, remove it carefully. Scanning through glass adds reflections and reduces sharpness.

No scanner? A smartphone can work in a pinch. Use a dedicated scanning app like Google PhotoScan or Microsoft Lens rather than the default camera. These apps stitch multiple exposures to eliminate glare and correct perspective.

Method	Resolution	Glare Control	Best For
Flatbed scanner (600+ DPI)	Excellent	None (no glass)	Loose prints, documents
Smartphone scanning app	Good	Automatic correction	Quick jobs, photos in albums
DSLR on copy stand	Excellent	Manual control	Large or fragile originals

Step 2: Restoring Damage

Old photos accumulate damage over decades: scratches, creases, water stains, mold spots, torn edges. AI restoration tools have become remarkably good at reconstructing missing detail.

Recommended free and affordable tools:

• Remini — The most popular AI photo restoration app. The free tier gives you several restorations per day. It excels at reconstructing faces, which is critical for ancestor photos. Available on iOS, Android, and web.
• MyHeritage Photo Enhancer — Specifically designed for genealogy use cases. It handles faded photos well and includes colorization. Free with limited usage; subscription for full access.
• GFP-GAN (open source) — If you are comfortable with Python, this face restoration model from Tencent produces results that rival commercial tools. Completely free.

Tips for dealing with specific types of damage:

• Fading: Most AI tools handle this automatically. If the photo is severely faded, increase contrast slightly in a photo editor before running it through AI restoration.
• Scratches and creases: Remini and similar tools fill these in automatically. For deep scratches that cut across a face, you may need to run the photo through restoration twice.
• Water damage and mold: These are the hardest to fix. If large portions of the image are destroyed, AI can sometimes hallucinate plausible detail, but the result may not be accurate to the original person. Be transparent with family when sharing these restorations.
• Torn or missing sections: Tools like Adobe Generative Fill can reconstruct missing portions of a photo. The free alternative GIMP with the Resynthesizer plugin can do similar work.

Step 3: Enhancing Resolution and Detail

After restoration, the image may still be soft or low-resolution. Upscaling tools use AI to add realistic detail that was not in the original scan.

Recommended tools:

• Topaz Gigapixel AI — The industry standard for photo upscaling. It can enlarge images up to 6x while adding convincing detail. One-time purchase, no subscription.
• Real-ESRGAN (open source) — A free alternative that runs locally. Quality is close to Topaz for most use cases.
• Upscayl (open source, GUI) — A user-friendly desktop app built on Real-ESRGAN. No command line required. Free for Windows, Mac, and Linux.

For best results, upscale to at least 1024x1024 pixels before animating. Most AI animation models perform significantly better with higher-resolution inputs. In our testing, animations generated from 512x512 images showed noticeable artifacts around the eyes and mouth, while 1024x1024 inputs produced smooth, natural movement.

Step 4: Animating with AI

This is where the magic happens. An AI video model takes your restored, enhanced photo and generates a short video of the person moving naturally — blinking, breathing, turning their head, even smiling.

What to look for in an animation tool:

• Face quality: The face is where viewers focus. Models trained specifically on human portraits produce far better results than general-purpose video generators.
• Natural motion: Subtle movement looks more realistic than dramatic motion. The best tools add gentle breathing, slight head turns, and natural blinks.
• Resolution: Look for at least 720p output. Lower resolutions look noticeably soft on modern screens.

Incarn uses Seedance 2.0, which is currently the leading model for portrait animation. You can try it directly on the homepage — upload a photo and see the result in about 60 seconds, no account required. For tips on getting the best results, see our complete guide to animating old photos.

The Full Pipeline in Practice

Here is what a typical workflow looks like end to end:

1. Scan the photo at 600 DPI using a flatbed scanner. Save as TIFF.
2. Restore damage using Remini (free tier). Run it twice if scratches persist.
3. Enhance resolution using Upscayl. Upscale to 2x or 4x, targeting at least 1024px on the shortest side.
4. Animate using Incarn. Upload, wait about a minute, download or share the result.

Total time: 10 to 15 minutes per photo. Total cost: potentially zero if you use free tools for every step, plus one free animation credit on Incarn.

Common Mistakes to Avoid

• Over-processing: Running a photo through too many AI tools can make faces look plastic or unnatural. One round of restoration and one round of upscaling is usually enough.
• Ignoring the original: Always keep the unedited scan. AI tools improve rapidly, and you may want to re-process the original with better tools in a year.
• Animating a group photo: Most animation models work best with a single face in frame. If your photo has multiple people, crop individual faces before animating.
• Using outdated tools: The technology has advanced rapidly. If you have been using Deep Nostalgia, check out our comparison of modern alternatives to see what is available now.
• Using JPEG at every step: Each time you save as JPEG, you lose quality. Use PNG or TIFF for intermediate steps.

Frequently Asked Questions

Can AI restore a photo that is almost completely destroyed?

It depends on how much of the face remains. If at least 60-70% of the facial features are visible, modern AI tools like Remini and GFP-GAN can produce a convincing restoration. If the face is mostly gone, the AI will generate a plausible face, but it may not resemble the actual person. For severely damaged photos, consult a professional photo restorer who can combine AI tools with manual retouching.

How much does the full workflow cost?

You can do it entirely for free. Remini offers free daily restorations, Upscayl is open source, and Incarn gives you a free animation to try without creating an account. If you want to process many photos, Remini's subscription is around $10/month and Incarn's credit packs start at a few dollars.

Will animating a restored photo look worse than animating a high-quality original?

There will be some quality difference, but it is smaller than you might expect. The restoration and enhancement steps bring the image close enough to a modern photo that animation models handle them well. The biggest factor is face resolution — make sure the face is at least 256 pixels wide after enhancement.

Is it legal to use AI to restore and animate photos of deceased relatives?

In most jurisdictions, there are no legal restrictions on restoring or animating photos of deceased family members for personal use. Copyright in a photograph typically belongs to the photographer, but for personal family photos, this is rarely an issue. If you plan to use the animations commercially or publicly, consider the photographer's rights and your jurisdiction's personality rights laws.