Photography Tips
JPEG Artifacts Removal: What Works and What Doesn't
Image quality can influence how people evaluate a listing, but there is no universal performance uplift that applies to every property. Measure clicks, inquiries, bookings, or sales on the individual listing.
By Hospitality Lens · · 15 min read
JPEG Artifacts Removal: What Works and What Doesn’t
You can reduce JPEG artifacts, but you cannot get back the detail the file threw away. Modern tools estimate plausible detail from the pixels around it, so they can calm blockiness and mosquito noise, and they cannot restore what was never saved. The best fix is almost always to go back to the original file and export it again. When there is no original, a careful AI cleanup is a reasonable second choice, as long as it stays truthful to the room.
This guide is written for property managers and co-hosts who inherit compressed photos: owner-submitted files sent through messaging apps, cleaner-taken shots pulled from a group chat, or old listing images downloaded back from a booking site.
HospitalityLens
Clean Up Listing Photos, Keep Them Real
HospitalityLens enhances existing property photos with careful cleanup while preserving the real room, a practical option when a new photoshoot is not needed.
Try a photoTL;DR
- JPEG artifacts (blocks, ringing halos, smeared texture) come from lossy compression, and the lost detail is gone for good. Tools can only estimate it.
- If you have the original camera file or a lossless master, re-export it. That beats any repair tool.
- One-click AI filters work best on a photo compressed once. Photos that were saved, uploaded and downloaded repeatedly are much harder to clean.
- Banding in gradients and color fringing on high-contrast edges are separate problems and need their own fixes.
- Over-smoothing makes a room look plastic. A little natural grain is more trustworthy than a waxy wall.
Table of Contents
What JPEG artifacts actually are
A JPEG does not store every pixel. It splits the picture into small 8x8 pixel blocks, runs a mathematical transform on each one, and then rounds off the fine, high-frequency information to make the file smaller. Researchers describe this as independent compression of 8x8 blocks followed by coarse quantization, which causes signal loss that cannot be fully recovered.
That single fact explains the three artifacts you will see most often in property photos.
Blockiness. Smooth areas such as a ceiling, a painted wall or a blue sky break into visible little squares. It shows up most when a file was shrunk hard to fit under an upload limit or a messaging app’s size cap.
Ringing, or mosquito noise. A fuzzy halo or speckle appears around sharp edges: a window frame against bright daylight, a black chair leg on a white floor, text on a sign.
Blurring. Fine texture such as wood grain, linen weave, tile grout or foliage turns into a soft smear. The research literature lists blocking, ringing and blurring together as the typical products of lossy compression.
None of these is a flaw in your property. They are flaws in the file. But a guest scrolling quickly will not make that distinction. A blocky bedroom photo reads as a neglected bedroom.
The most useful rule in this whole topic is about prevention. Exporting once from the original source at a sensible quality setting is always better than repairing a heavily compressed file, because reconstruction tools can only approximate lost detail. The same guide suggests a JPEG quality somewhere around 75 to 85 for a good balance of size and cleanliness.
So before you open any filter, ask three questions.
Does the photographer, owner or cleaner still have the original? Phone galleries often hold the full-size file even after it was sent through a messaging app. Ask for the file itself, not a screenshot and not a forwarded copy. Many chat apps shrink and recompress images on the way through.
Is there a lossless master? If your team shoots on a camera, keep RAW files or export TIFF or PNG masters, then create JPEGs only when a platform needs them. Artifacts also compound with every save, because re-saving a damaged JPEG at low quality encodes the damage and adds new errors.
Did the photo come back down from a platform? If a file was uploaded to a social site or booking platform and then downloaded again, it has been compressed at least twice. Cleanup tools do their best work on single-compressed images, because double compression leaves overlapping patterns that are harder to separate.
If the answer to the first two is yes, stop here and re-export. That costs nothing and gives you a cleaner result than any filter.
The one-click options, and where they stop
When no original exists, the quickest route is a purpose-built filter. Adobe Photoshop includes a JPEG artifacts removal Quick Action that analyzes the image and reduces compression distortion with one click, and you can refine the strength afterward. Photoshop Elements offers a similar Quick Actions option that examines patterns across the picture to find and soften the damage.
There are also free browser-based tools that promise a result in seconds. They are fine for a quick test, but check two things before you rely on them: the maximum image size they accept, and what they do with your file after upload. Some cap input at roughly 2 megapixels, which is a downgrade if your source is larger.
The academic work behind these tools is encouraging. Deep convolutional networks can reduce blocking without the heavy blur older methods produced. But the framing matters: this is an ill-posed problem, meaning many different original images could have produced the same compressed file. The software picks the most plausible one. For a flat painted wall, plausible is good enough. For a patterned rug or a tiled backsplash, you should zoom in and check that the result still looks like your rug and your tile.
A practical cleanup order
If you are repairing a photo by hand, work in this order and stop as soon as it looks right.
First, run the automated artifact filter at the lowest strength that removes the visible blocks. Low, medium and high settings exist for a reason, and high often goes too far.
Second, inspect at 100 percent zoom, not fit-to-screen. Look at skies, white ceilings, window edges and any textured surface.
Third, fix color and tone after the cleanup, not before. Brightening a dark, compressed photo first stretches the damage and makes blocks in the shadows much more obvious. This is why dark owner-submitted photos are the hardest to rescue.
Fourth, add back a trace of texture if the result looks waxy. Which brings us to the biggest risk.
The plastic look: how over-cleaning backfires
It is tempting to throw a strong blur, a median filter or a dust-and-scratches filter at a noisy file. Practitioners warn that indiscriminate use of those filters gives an unnaturally smooth, plastic appearance and strips out texture the room needs.
That matters more for lodging than for most photography. A guest compares your photos to what they find at check-in. Linen that looked like smooth plastic online, or a wooden floor with no grain, creates a quiet mismatch. Slight natural imperfection is more believable than a perfectly airbrushed surface.
A good rule: if you cannot tell what material a surface is made of, the cleanup went too far. Back the strength off.
Banding: the gradient problem
Banding looks like stripes or steps in a smooth gradient: a sunset sky, a painted wall washed by a lamp, the glow around a pendant light. It tends to appear in 8-bit files, and it often gets worse after a platform recompresses your upload.
One common fix is a light Surface Blur on a duplicate layer, with a radius in roughly the 8 to 15 pixel range, followed by a small amount of monochromatic Gaussian noise, about 1 to 3 percent. The noise breaks up the hard steps and helps keep them from returning when the file is compressed again.
One photographer who has written about this points out that every banding method adds another artifact, such as noise, blur or texture, so it is important to apply the fix locally and mask out the parts of the image that matter. In a bedroom photo, that means treating the wall gradient and leaving the bedding, art and furniture alone.
Chromatic aberration: the colored edge
Color fringing is a different animal. It appears as a purple, green or blue outline along high-contrast edges, such as dark window frames against a bright sky. It is a lens effect, not a compression effect, although compression can make it look worse.
In Lightroom, the first thing to try is the Remove Chromatic Aberration checkbox in Lens Corrections. The same author notes that the manual Defringe tool can cause color shifts elsewhere in the picture, especially around metal objects and blue sky, so a masked brush is safer for stubborn spots.
One caution from photography editors: even after a general cleanup, ringing and color fringing often persist on high-contrast edges and may need targeted defringe or lens correction. In practice, window views are where this shows. Check them last, and check them closely.
Which problem are you looking at?
| What you see |
Likely cause |
Best first move |
| Small squares in walls, ceilings or sky |
Heavy compression, often a shrunken file |
Re-export from the original; otherwise use an artifact filter at low strength |
| Fuzzy halo around window frames and furniture edges |
Ringing from compression |
Artifact filter, then inspect edges at 100 percent |
| Soft, smeared texture |
Compression plus downsizing |
Re-export; sharpen lightly only if needed |
| Stripes in a gradient |
8-bit banding, often after platform recompression |
Local blur plus light noise, applied by mask |
| Purple or green outline on bright edges |
Chromatic aberration from the lens |
Lens correction checkbox, then masked defringe |
| Everything is dark, soft and blocky |
Underexposed file compressed more than once |
Ask for a new photo from the owner or cleaner |
How to stop creating artifacts across a portfolio
If you manage a dozen listings or a hundred, most of the pain comes from workflow, not from any single file. A few habits help.
Collect originals, not forwards. Give owners and cleaners a shared folder or upload link instead of asking them to text pictures. Tell them to send the photo at full size.
Keep one master per photo. Store the unedited original, and export listing versions from it. Never edit and re-save a JPEG repeatedly.
Use one export setting. Aim for a unified, high-quality export across all your listing channels so that no single platform’s compression becomes the weak link. Each channel also has its own size rules. Vrbo, for example, is reported to require at least 1024 x 683 pixels, so uploading a tiny, already-compressed file leaves you little room. Check the platform’s own help page before you export.
If you want a fuller process, our guides on preparing photos for booking sites and standardizing vacation rental photos cover the portfolio side in more detail, and our photo standards guide for managers helps you write the rules down for owners and cleaners.
Our perspective: fix the photo, never the property
We see compressed, crooked, dim photos every week, usually sent by an owner who took them on a phone or a cleaner who snapped them between turnovers. Our view is that most of these are not hopeless, and that perfection is not the goal. The goal is a photo that looks like the real room on a good day.
That is why we draw a firm line. We improve lighting, color, perspective and clutter, and we polish the overall image while preserving the real property. We do not invent rooms, views, amenities or architectural features. Artifact cleanup follows the same principle: a calmer sky and cleaner edges are fine, a wall that has lost its texture is not.
We are also direct about limits. A heavily compressed, very dark file may not be worth rescuing, and a professional photographer is the better call for a flagship launch, architectural work, new angles or difficult capture conditions. Enhancement fits when you already have usable phone photos or existing listing images and a new shoot would be more than the job needs.
If you are deciding between a reshoot and a refresh, our comparison of HospitalityLens and an Airbnb photographer lays out when each makes sense.
— The HospitalityLens Team
When a refresh beats a reshoot
If your problem is a stack of inherited photos that are dark, crooked, cluttered or soft, a full photoshoot is not always the first answer. For multi-listing teams, enhancing what you already have is faster and cheaper to schedule, and it keeps the property true to how it looks.
Current HospitalityLens pricing (USD); one-time packs — The Hero: 1 photo for $12; The Refresh: 10 photos for $59; The Property: 25 photos for $129; The Portfolio: 100 photos for $399; monthly subscriptions — THE STARTER: 10 photos/month for $49/month; THE OPERATOR: 25 photos/month for $99/month; THE PORTFOLIO: 100 photos/month for $299/month; one-time pack credits do not expire; monthly subscription credits remain valid for up to 3 months; 2 free revisions per photo on one-time packs; 3 free revisions per photo on monthly subscriptions. See current pricing.
The simplest way to judge quality is to test it. Try one photo free: upload a file and you get a watermarked preview, typically in about 15 minutes, with no obligation. Pick your worst offender, ideally a compressed owner-submitted photo, and see whether the result holds up at full zoom. Our step-by-step guide walks through the upload, and our property managers page explains how it works for teams.
To be clear about scope: we focus on lighting and color correction, perspective and horizon correction, clutter cleanup, people or object removal, and overall polish. If a file is so degraded that no honest enhancement will help, we would rather you know before you commit to a bulk order.
A short checklist before you upload anything
Before any photo goes to a booking platform, run through these checks in order.
Is this the original file, or a copy that passed through a chat app? Ask for the original.
Does it look clean at 100 percent zoom in the sky, ceiling and window edges? If not, try a low-strength artifact cleanup.
Does any gradient show stripes? Treat that area locally.
Do bright edges show colored outlines? Apply lens correction and a masked defringe.
Can you still tell what every surface is made of? If not, you smoothed too much.
Does the photo still match the room a guest will walk into? If yes, it is ready.
For more on how images perform once they are live, see our notes on vacation rental image tips and hotel photo editing.
FAQ
What does it mean to remove JPEG artifacts?
It means reducing the visible damage that lossy compression leaves behind, such as blocky squares, halos around edges and smeared texture. The software does not restore the original pixels. It estimates plausible detail from the surrounding image, so results vary with how badly the file was compressed.
Can I fully restore a heavily compressed JPEG?
No. High-frequency detail is permanently discarded when a JPEG is saved, so any tool is reconstructing, not recovering. If the original camera file or a lossless master exists, re-exporting from it will always look better than repairing the compressed copy.
Why do my photos look worse after I upload them to a booking site?
Platforms recompress uploads to keep pages fast, and a file that was already compressed gets compressed again. Uploading a high-quality, full-size export gives the platform more to work with. Check the platform’s help page for current size and resolution requirements.
Does artifact removal fix color fringing and banding too?
Not reliably. Color fringing is a lens effect and usually needs lens correction or a defringe tool, while banding in gradients is usually handled with local blur and light noise. A general artifact filter may reduce both slightly, but you should check edges and gradients yourself.
Is an AI cleanup safe for listing photos?
It is, when it stays truthful. Keep the strength low, zoom in to confirm textures still look real, and never use a tool that adds or changes features of the property. A cleanup that makes a room look different from reality will cost you in guest reviews.
Sources
Recommended