Why your old JPEGs look blocky (and how to fix them)
Last month, a user emailed me a photo from 2015. Compressed to death, saved and resaved a dozen times, downloaded from a forum, uploaded to WhatsApp, downloaded again. It looked like someone had pixelated it on purpose. She asked if we could help. That email landed the same week we were testing JPEG cleanup in Cleanr, and it crystallised why this matters.
The compression problem nobody talks about
JPEG compression is a marvel of maths. It lets you cram a photo into a fraction of its original file size by discarding information your eye supposedly won't notice. Works beautifully once. Save a JPEG, and you lose some detail. Save it again as a JPEG, and you lose more. Do it five times, and the image falls apart into visible blocks and colour banding. Your social media photos, archive shots, screenshots saved as JPEGs - they all suffer from this.
The technical term is "compression artefacts," but what users see is just a messy, degraded image. We discovered this was one of the most common complaints in our user feedback. People weren't asking for exotic features. They wanted their old photos to look crisp again.
How Cleanr actually tackles it
We use a median filter paired with noise reduction on the compressed image. A median filter works by looking at small clusters of pixels and replacing each pixel with the median value of its neighbours. This smooths out the blocky artefacts without blurring the entire image into mush. It's a surgical approach, not a sledgehammer.
The process happens in a single tap. You load your JPEG, tap "Cleanup," and within seconds you get back a cleaner version. No sliders to fiddle with, no preview toggles to master. The filter does the heavy lifting based on what it detects in the image. This was deliberate. We've watched people abandon tools that require too much thought. Simplicity wins.
When it's not just old photos
JPEG artefacts aren't just a museum problem. Mobile messaging apps crush images. Screenshots from videos are compressed twice. Downloaded product photos for small business owners are often low-quality JPEGs from suppliers. One Etsy seller told us her product photos looked fuzzy until she ran them through cleanup, then they were sharp enough for her listings. She was doing this manually in five different apps before Cleanr.
That's the pattern we kept seeing. The feature solves a real friction point. People had jobs they were already doing; we just made it fit into one place instead of scattered across their phone.
Why we didn't oversell it
Cleanup is one of 22 tools in Cleanr. It's not the flashiest. Sky replacement and generative fill grab attention. But the quiet features are often the most useful. We get daily emails from people who use cleanup, then object removal, then auto enhance, and suddenly a mediocre photo becomes one they're proud to post. The cleanup step is invisible to them. It just works.
The Free tier includes three cleanups a day. We wanted this accessible to everyone. A student digitising old family photos shouldn't hit a paywall after one attempt. Plus and Pro tiers unlock unlimited cleanups and batch processing for people who need to restore entire photo libraries at once.
What we learned in testing
Before launch, I expected the median filter to be too aggressive on some images and too mild on others. Testing showed the opposite. It's surprisingly solid across different kinds of compression damage. The real lesson was speed. Users don't wait. If cleanup took three seconds instead of one, the feature felt broken even though technically nothing had changed. We optimised until it felt instant.
We also discovered people use it on photos they've never thought of as "old." Compressed downloads from the internet, phone screenshots, blurry security footage someone asked us to enhance. It became a catch-all solution for photos that arrived damaged or degraded, not just a tool for digitising archives.
The next time you pull up an old photo on your phone and squint at the blocky mess, remember that's not damage you have to live with. But here's what I'd ask you: how many photos in your camera roll are sitting in the same state right now, waiting for someone to actually do something about them?