Basics 2026-10-12 Updated 2026-10-30 9 min By Cornelious Fazal

QR Code with Logo Best Practices: Level H, Size, Contrast

Create a QR Code with Logo Free · No signup · Permanent

Quick Answer

A logo spends part of the code's error-correction budget, so use Level H, keep the logo small and centered away from finder patterns, back it with a light plate, and test a full-size printed proof on several phones.

QR code with logo best practices: set error correction to Level H, keep the logo near a tenth of the symbol area, stay clear of the corner finder patterns, and keep dark-on-light contrast strong. A logo works because ISO/IEC 18004 builds Reed-Solomon error correction into every symbol, so a decoder can rebuild a limited number of covered modules. A centered logo is damage you add on purpose, so leave room for real-world wear.

A branded code that scans well on a new phone in good light can still fail on an older handset, at an angle, or under glare. This guide explains the recovery budget a logo spends, how big the logo can be (and why published advice disagrees), where it can sit, how to build the light plate behind it, and how to test the result before a print run. No logo size or correction setting guarantees a scan, so the last step is always a printed proof tested on real phones.

How does a QR code with a logo remain scannable?

When the encoder builds the symbol, it adds Reed-Solomon parity codewords alongside the data. A decoder that finds modules missing under the logo treats them like dirt or damage and uses the parity data to work out what they should have been.

That recovery capacity is a fixed budget set by the error-correction level. The logo spends part of it before the code is even printed. Glare, scratches, curvature or a dim room spend more. Once the total goes past what the level can recover, the code stops decoding.

Two bars showing the Level H recovery budget: a modest logo using one third of it and leaving two thirds for real-world conditions, and a logo at the 30 percent ceiling using all of it
Original diagram: a logo and real-world wear draw on the same recovery budget. The split for the modest logo follows from 10 of the 30 percentage points; what the remainder absorbs is qualitative.

One detail makes the budget less predictable than a single percentage suggests. The figures refer to codewords, units of eight bits, and DENSO WAVE states that the rate is a ratio of total codewords including the parity ones. A logo covers a contiguous patch of the picture, and that patch can touch different parts of the data in different ways. This is why two codes with the same logo percentage can behave differently on the same phone.

Which error correction level should a logo QR code use?

Error correction levelApproximate restorationWith a logoGrid densityTypical use
Level L (Low)~7% of codewords✗ Not suitableLowest (largest modules)Clean digital screens without logos
Level M (Medium)~15% of codewords! Risky, little spare capacityMediumStandard print labels without artwork
Level Q (Quartile)~25% of codewords∼ Possible for a small mark; test carefullyHigherPackaging that may get dirty or scuffed
Level H (High)~30% of codewords✓ Usual choiceHighest (densest grid)Codes with an embedded logo

The restoration percentages are DENSO WAVE's published approximations for codewords, not for any area of the image. For a branded code, use Level H. Our QR code generator with logo switches to Level H automatically when you add one. The cost is density: DENSO WAVE notes that raising the level increases the amount of data and therefore the code's size, and one vendor help page gives a worked example of the same URL growing from 25×25 modules at Level L to 29×29 at Level H. At a fixed print size each module gets smaller. A short payload offsets that. The error correction levels guide compares all four in more depth.

How big can the logo be?

Level H can restore roughly 30% of codewords, but a logo that size leaves nothing for the conditions the code will actually meet. Published advice does not agree on a number, and none of it comes from the standard:

  • One vendor help page states a centered logo may take up to 30% of the surface at Level H, while also telling readers to test on different devices and readers.
  • Other vendor guides suggest a working limit in the range of 15% to 25% of the code area.
  • A common design guideline is about 10% to 12% of the symbol area.

Treat the smallest of those as a starting point and the largest as a ceiling you may never reach. Area grows with the square of the width, so a logo covering 10% of the area is about 31.6% of the width:

Max Logo Width = Total QR Width × √(0.10) ≈ Total QR Width × 0.316
  • A 500 × 500 px symbol allows a logo up to about 158 × 158 px.
  • A 1000 × 1000 px symbol allows a logo up to about 316 × 316 px.
  • A 50 mm (about 2 in) print allows a logo up to about 15.8 mm (0.62 in) wide.
Four schematic squares with a centered logo covering 5, 10, 20 and 30 percent of the area, which is 22, 32, 45 and 55 percent of the width, with a note on published guidance
Original diagram: the width of the logo at four area shares. Logos look smaller than the numbers feel, which is why the 30% figure is risky.

The grid itself matters too. A longer URL pushes the symbol to a larger version, which has more modules, and a fixed logo then hides a smaller fraction of them. A version 3 symbol has 29×29 modules (841 in all), so a logo at 10% of the area covers roughly 84 modules, a patch about nine modules across. A short link and a clean logo are the two cheapest ways to widen your safety margin. If the link is long, a dynamic QR code can encode a shorter redirect address instead, at the price of depending on a service.

Where should the logo go?

Put it in the center. That keeps it away from the structures a scanner needs to find and read the code, most of which sit near the edges and corners:

  1. The three finder patterns are the large squares in the top-left, top-right and bottom-left corners. Their 1:1:3:1:1 ring proportions let a scanner locate the symbol and work out its orientation. Covering even one of them can stop the camera from detecting the code at all.
  2. Timing patterns run between the finders as a line of alternating modules and tell the decoder the grid spacing.
  3. Format information sits beside the finder patterns and records the error-correction level and mask. If those modules are damaged, the decoder may not be able to interpret the rest.
  4. Alignment patterns are smaller squares that appear from Version 2 upward and help with perspective on angled or curved surfaces. From Version 7 upward one of them sits in the middle of the grid, directly under a typical logo, which is one more reason to test the finished code.
Labeled schematic of a 29 by 29 module QR code with finder patterns, timing lines and format information strips marked as keep clear, and a small logo on a light plate in the center
Original diagram: the logo lives in the middle of the data area. The square is a labeled schematic, not a scannable code.

Does the logo need a background plate?

Yes, in most designs. Put a solid light plate behind the logo so its edges do not run into the surrounding modules. A transparent logo that touches black modules makes it harder for the decoder to see where the data starts.

  • Leave a margin of one or two modules of plain white (or the code's light background color) around the logo.
  • Keep the data modules clearly darker than the background. Black on white is the safe baseline. A contrast checker screens out weak pairs but cannot certify that a print will decode; the QR code colors guide covers brand palettes.
  • Do not set a pale yellow, pastel or low-contrast icon straight over dark modules. Give it the light plate first.

The plate is not free either, since it is also missing data. Count the plate, not just the artwork, when you work out the logo's share of the area. DENSO WAVE's troubleshooting notes list overlapping pictures that reduce contrast between light and dark areas among the causes of unreadable codes, which is the same problem a busy logo edge creates.

How do you design and test a branded QR code?

Work through these steps before anything goes to mass print:

  1. Export the logo as SVG or a high-resolution PNG on a square or circular canvas, with a little white padding around the mark.
  2. Open the branded QR generator, enter your destination URL, set error correction to High (Level H), and upload the logo.
  3. Keep the URL short. Fewer characters mean fewer modules, and larger modules are easier to read.
  4. Download an SVG for print, or a 1000 px PNG for screens.
  5. Print a proof at 100% scale and scan it with at least three phones, for example an iPhone, a recent Android flagship and a budget Android. Try it in low light and at a steep angle of around 45 degrees, not just straight on at your desk.
Loop with five steps: prepare the logo, generate at Level H, print a full-size proof, test on at least three phones in good light, dim light and at an angle, then approve or shrink the logo and repeat
Original diagram: a failed scan sends you back to the generator, not forward to print.

For a stress test before printing, the QR scannability tester can apply blur and low-light conditions to your exported image, and the print size guide shows how large the code needs to be for the viewing distance. Test your own code in a browser with the online QR scanner as well as with phone cameras.

What mistakes break branded QR codes?

Consider a boutique that orders 2,000 shelf cards with a logo covering about a quarter of the code, because the preview on the designer's screen scanned fine. On the shop floor, the cards sit behind glossy lamination under spotlights. The glare adds damage the logo's share never allowed for, and several older phones fail to read them. The reprint costs more than a proof would have.

  • Sizing the logo to the maximum percentage instead of leaving room for wear.
  • Letting the artwork touch a finder pattern or the format strips.
  • Pasting a transparent logo straight on the modules without a plate.
  • Testing only the preview on one phone, and never the finished print.
  • Using a long link, so Level H pushes the grid to a dense version with tiny modules.

What matters most?

  • Use Level H, and treat the 10% area figure as a guideline to test, not a promise.
  • Keep the logo off the finder patterns, timing patterns and format information.
  • Keep the payload short so Level H's extra modules do not end up too small at print size.
  • Approve a full-size printed proof on several phones before a production run. For payments, access control or safety information, an unmodified code is the safer choice.

What can only a printed proof settle?

No source consulted gives a size at which every logo is safe, and scanner behavior varies by phone, app and lighting, so the numbers here are starting points that your own tests have to confirm. Brand rules may also limit how a logo can be placed on or near a code; check your brand guidelines and any platform rules for the channel where the code will appear. This guide is general information, not legal or compliance advice.

References

// FAQ

Frequently Asked Questions

A logo does not universally require Level H, although higher correction can provide more recovery capacity. The safe choice depends on payload density, logo size, placement, contrast, and print quality. Keep functional patterns untouched and test the exact finished code; error-correction percentages are not logo-size guarantees.

No standards-based percentage guarantees a safe logo size. A modest centered logo is less risky than a large or irregular overlay, but payload density and the covered modules matter. Begin small, preserve functional patterns and quiet zone, then test each export and production sample on representative devices.

Place the logo strictly in the center of the matrix. Never cover or touch the three large corner Finder Patterns, Timing Patterns, Alignment Patterns, or Format Information strips near the corners.

A solid knockout background plate (1 to 2 modules wide) prevents detailed logo artwork from merging into adjacent QR data modules, ensuring camera edge-detection algorithms can distinguish data boundaries.

There is no official minimum ratio for QR codes. Keep a strong difference in brightness between the dark modules and a light, solid background. Black on white is the safest, and a contrast checker (a 4:1 ratio is a common screening threshold) can catch weak combinations, but only a printed test shows whether a code scans.

No. One vendor help page says a centered logo may take up to 30% of the surface at Level H, other vendor guides suggest 15% to 25%, and a common design guideline is about 10% to 12%. None comes from the standard, so treat the lowest as a start and test.

The logo itself does not, but the Level H setting that usually accompanies it does. DENSO WAVE notes that raising the correction level increases the data amount and the symbol size, so the same link needs a denser grid. A short link keeps the modules large enough to print.

It is possible, but a plain unmodified code is the safer choice where a failed scan has real cost, such as payments, access control or safety information. If you do brand it, use Level H, keep the logo small and test the printed code on several phones.
Back to Knowledge Hub