NearIMG

2026-09-26

Why a QR code stops scanning after you edit it — and what the FTC's 2026 warning is actually about

The FTC's own September 2026 consumer alert opens with an unusually blunt title: “See a QR code parked somewhere? Don't scan it…yet!” It describes something the agency first warned about back in December 2023 and keeps re-flagging — scammers sticking a fake QR code directly over a legitimate one on a parking meter, so a driver who just wants to pay for parking ends up on a lookalike payment page instead. The FBI's Internet Crime Complaint Center (IC3) had already warned about the same tampering trick in a January 2022 public service announcement, and followed up in July 2025 with a separate alert about QR codes arriving inside unsolicited packages. None of that is really about image editing — it's about which QR code you choose to trust enough to scan at all, and that's a judgment call no image editor can make for you (more on that below). But a much more common, far less newsworthy problem sits right next to it: a completely legitimate QR code — a Wi-Fi password on a printed flyer, a payment handle screenshotted to text a friend, a ticket cropped out of a longer confirmation email — that simply stops scanning after someone edits it. Here's what actually breaks a QR code during an edit, and what doesn't.

What a QR code needs to keep working

A QR code isn't a photograph with slack built in — it's a fixed grid of black-and-white squares (modules) plus a required blank border, and the scanning math depends on both staying intact. Two facts from Denso Wave — the company that invented the QR code and still maintains its official specification pages — explain most of what goes wrong:

RuleWhat it actually says
The quiet zoneDenso Wave's own specification page states it plainly: “QR Code requires a four-module wide margin at all sides of a symbol.” That blank border isn't visual padding — it's how a scanner first locates the code's three corner squares, before it ever tries to read the data inside.
Error correctionThe ISO/IEC 18004 standard QR codes are built on defines four error-correction levels — L, M, Q, and H — commonly documented as tolerating roughly 7%, 15%, 25%, and 30% damage respectively. Denso Wave's own page says Level M is “most frequently selected” as the everyday default. That budget is what lets a code survive a little dust, a fold, or a small logo — it isn't unlimited, and it isn't a substitute for the quiet zone.

Where an edit actually breaks it

  • Cropping into the quiet zone. The single most common mistake: cropping a screenshot down tight to “just the square part” to make it look neater, which eats into that required four-module margin. The code underneath can be otherwise perfect and still fail to scan, because the scanner never manages to find it in the first place.
  • Re-saving as a low-quality JPG. JPEG's compression works in blocks and is built for photographs, where a slightly shifted pixel is invisible. A QR code has no tolerance for that kind of shift right at a module's edge — push the quality slider too low and the compression artifacts can flip enough modules to eat into the error-correction budget above, especially on a code that was already using a low correction level.
  • Shrinking it too far. Resize a QR code down enough and individual modules end up smaller than a single output pixel, so neighboring squares blur into each other. A phone camera scanning from arm's length needs each module to still read as a distinct square, not a gray smear.
  • Screenshotting a screenshot. Every re-save adds its own rounding and, if it's JPEG, its own new compression pass on top of the last one. A code that's been forwarded, re-screenshotted, and re-sent a few times has usually taken several small hits before anyone notices it stopped scanning.
Same QR code, before and after four common edits Only one of these keeps its required quiet zone and error-correction budget intact Original: margin intact still scans Cropped into the margin scanner can't find it Over-compressed JPG artifacts flip modules Shrunk too far modules blur together Repeated re-saving stacks all of this screenshot → text it → recipient screenshots the message → forwards it again each hop can crop the margin a little tighter and re-compress it a little harder, so a code that scanned fine the first time can fail by the third or fourth re-share
The quiet zone and the error-correction budget both have to survive an edit — a QR code that looks fine to a human eye can still fail a scanner that can no longer find or read it.

Editing one without breaking it

None of NearIMG's tools know a QR code is a QR code — they treat every image the same way, which is exactly why it helps to know what to avoid:

  • Crop generously. Leave visible white space around the black square rather than trimming to its exact edge — when in doubt, leave more margin, not less.
  • Convert to PNG rather than JPG when the source allows it. PNG's compression is lossless, so it can't introduce the block artifacts a low-quality JPG can.
  • Compress only if the file size genuinely needs to come down, and keep the quality slider high — then actually scan the result with a phone before sending it, rather than assuming a small size reduction was safe.
  • Resize down cautiously, and re-scan afterward. The AI-upscale option under Resize is built to add plausible photographic detail to a photo — it has no way to reconstruct a QR module that a heavy downscale already merged with its neighbor, so it isn't a fix for a code that's already too small or blurry to read.

What this honestly can't do

  • It can't tell you whether a QR code is safe to scan. That's exactly the problem the FTC and FBI alerts above describe — a tampered sticker on a parking meter, or a code mailed in an unsolicited package, can be printed at perfect resolution with a perfect quiet zone and still lead somewhere malicious. Checking the destination link most phone cameras preview before you tap it — not the image quality — is what catches that.
  • It doesn't generate or decode QR codes. This is about editing an image that already contains one — a screenshot, a saved flyer, a photo of a sign — not creating a new code from a link or a password.
  • It can't guarantee any specific phone will scan it. Camera quality, lighting, and distance all vary; staying well inside the quiet-zone and compression margins above gives a code the best odds, not a guarantee.

The short version

A QR code needs a four-module blank margin on every side — Denso Wave's own specification says so directly — and depends on an error-correction budget that a low-quality JPG or a too-aggressive downscale can eat into. Crop generously, keep it as a PNG when possible, compress lightly and re-scan the result, and don't expect a resize (AI-upscale included) to rescue a code that's already too small or blurry. None of that, however, answers a completely different question the FTC has spent since December 2023 warning drivers about — whether the QR code in front of you is the real one at all.

Sponsored
← NearIMG

This page shows ads only if you consent.