技術ガイド · 2026
QRコードの誤り訂正：レベル、Reed-Solomon、選び方
QRコードは、エンコード時に冗長データが加わるため、多少の傷、汚れ、部分的な遮蔽があっても読み取れます。誤り訂正は処理の一段階です。全体の流れはQRコードの生成方法をご覧ください。
QRコードの誤り訂正とは
QRコードの誤り訂正とは、記号の一部が損傷・汚損・遮蔽された場合にもスキャナーがペイロードを復元できるよう、エンコード時に追加される冗長データです。Reed-Solomon誤り訂正には、L（約7%）、M（約15%）、Q（約25%）、H（約30%）の4つの標準レベルがあります。上位レベルほど損傷に強くなりますが、マトリクスが密になり、通常はより大きな印刷サイズが必要です。
きれいなデジタル画面にはL、一般印刷にはM、包装やラベルにはQ、ロゴ重ねや屋外利用にはHを選びます。
QRコードの誤り訂正とは？
Why QR codes need redundancy
A printed QR code can pick up scuffs, ink loss, folds, glare, or a logo overlay. Redundancy gives the decoder a margin for these real-world defects instead of requiring every black and white square to be perfect.
Reed-Solomon in plain language
Reed-Solomon is the block error-correction algorithm specified for QR symbols by ISO/IEC 18004. It calculates extra codewords—units of data or redundancy—from the payload, then stores them in the finished symbol.
Generation and scanning are different stages
The generator adds redundancy during encoding. A scanner uses that redundancy during decoding to repair some missing or misread modules, the individual black or white squares in the QR matrix.
A standardized QR feature
ISO/IEC 18004 defines the four levels and their approximate recovery guidance. Denso Wave introduced QR Code in 1994; modern compliant encoders use the standardized error-correction structure rather than an app-specific recovery system.
QRコードでReed-Solomon誤り訂正はどう機能する？
Data codewords carry the URL, text, or other payload. Error-correction codewords are calculated from that data and occupy part of the symbol's finite capacity. More recovery codewords mean less room for payload data at a given QR version.
During a scan, finder patterns first help the camera locate and orient the symbol. The decoder samples modules, detects inconsistencies, and applies Reed-Solomon recovery before extracting the payload. Read how QR codes work for the wider scan-to-action flow.
Error correction cannot repair every failure. A cropped quiet zone, obscured finder pattern, extreme motion blur, severe glare, or poor contrast can prevent a scanner from locating or interpreting the symbol before recovery can help.
QRコードの誤り訂正レベル（L、M、Q、H）とは？
QRコード誤り訂正レベル表
| Level | Name | Approx. recovery | Data capacity | Module density | Best for |
|---|---|---|---|---|---|
| L | Low | ~7% | Highest | Lowest | Clean screens and controlled indoor print |
| M | Medium | ~15% | High | Moderate | Flyers, brochures, and cards without logos |
| Q | Quartile | ~25% | Moderate | Higher | Product labels, packaging, and moderate wear |
| H | High | ~30% | Lowest | Highest | Logo overlays, apparel, and outdoor signage |
復元率はISO/IEC 18004による近似値であり、記号内のどの位置の損傷でも読み取れる保証ではありません。
Level L: minimal redundancy
Level L suits a clean screen, PDF on a monitor, or controlled placement where physical damage is not expected. Its lower density can help when the physical symbol must stay relatively small.
Level M: balanced default
Level M is commonly used by compliant encoders as a general-purpose balance of capacity and tolerance. It is a sensible starting point for ordinary printed marketing materials without a logo overlay.
Level Q: practical protection
Level Q adds room for scuffs, partial ink loss, and handling variation. It is a strong fit for corrugated packaging, warehouse labels, and laminated restaurant menus.
Level H: maximum standardized tolerance
Level H is the usual choice for a meaningful centered logo or a harsh environment. It still needs an intact quiet zone, high contrast, adequate print size, and scan testing after the final proof.
どの誤り訂正レベルを使うべき？
- 1 Will the symbol stay on a clean digital display with no physical wear? If yes, Level L may suffice.
- 2 Will a centered logo or brand mark cover modules? If yes, choose Level H and test the finished design.
- 3 Will the label face weather, flexing, dirt, or regular handling? Prefer Level Q or H.
- 4 Is the payload long or the label very small? Shorten the URL first; higher EC makes the matrix denser and may need more print area.
- 5 Is this general-purpose indoor print without a logo? Level M is usually the balanced starting point, followed by proof scans.
| Use case | Recommended level | Why |
|---|---|---|
| Digital screen or PDF on a monitor | L | No expected physical wear |
| Flyers, brochures, indoor posters | M | Moderate print handling |
| Product labels and corrugated packaging | Q | Scuffs and ink variation |
| Business card with centered logo | H | Logo obscures some modules |
| Outdoor signage or apparel | H | Weather, fading, and flexing |
| Long URL on a small label | M + shorten URL first | Density limits the benefit of extreme EC |
誤り訂正はQRコードのサイズとデータ密度にどう影響する？
Long URLs and complex payloads compound this trade-off. A short redirect used by a dynamic QR can reduce matrix density before you raise error correction; see static vs dynamic QR codes and pricing for that separate redirect-and-analytics model. Use the QR code size guide for dimension and scan-distance guidance, and review types of QR codes when payload format affects density.
EC Level vs Size Trade-off
| Factor | Lower EC (L) | Higher EC (H) |
|---|---|---|
| Damage tolerance | Lower | Higher |
| Payload room at same version | More | Less |
| Matrix density | Lower | Higher |
| Minimum practical print size | Smaller may work | Larger is usually needed |
| Logo overlay suitability | Not recommended | Recommended with testing |
ロゴ重ねにはどの誤り訂正レベルが必要？
Error correction can help recover covered data modules, but it does not protect the quiet zone: the blank margin around the symbol. Do not overlap that margin or the three large finder patterns with a logo, border, or decoration.
Contrast still matters. A branded code with low contrast, glossy glare, or blurred module edges can fail before Reed-Solomon recovery is relevant. See custom QR code design for the branding workflow and examples.
スキャナーはコード読み取り時に誤り訂正をどう使う？
The order matters: a scanner must first find and orient the QR code. If the quiet zone is cropped, finder patterns are obscured, or glare prevents module sampling, recovery capacity cannot reliably rescue the scan.
Treat error correction as a margin of resilience, not a replacement for adequate size, contrast, a clear quiet zone, and a final device test.
実際のユースケース：誤り訂正の活用
Restaurant menus and table tents
Level M–Q
Use M for clean indoor print; move to Q for laminated cards that get wiped, handled, and exposed to food-service wear.
Warehouse and inventory labels
Level Q
Q adds a practical buffer for scuffs, dust, and partial print loss on labels that move through handling workflows.
Event badges and lanyards
Level Q–H
Choose Q for ordinary event handling and H when the badge bends, gets wet, or includes a logo that covers modules.
Corrugated or glossy product packaging
Level Q
Q helps account for ink variation, scuffs, and uneven substrates; approve it only after scanning the actual production stock.
Branded business cards
Level H
Use H for a centered logo, preserve the quiet zone, and test at the card's final dimensions before printing a run.
Screens, email, and digital-only campaigns
Level L–M
L can be appropriate when there is no wear risk; M adds a moderate buffer for screenshots or less-controlled viewing.
See more deployment examples in QR code examples.
QRコードの誤り訂正でよくある間違い
- Using Level L on packaging that will be scratched in transit.
- Assuming Level H makes a tiny code readable from any distance.
- Covering finder patterns or the quiet zone with a logo or design element.
- Raising EC without allowing a larger physical print size.
- Embedding a long URL in a small label and then demanding Level H.
- Skipping cross-device scans after printing the final proof.
- Confusing symbol error correction with dynamic QR redirect reliability or analytics.
QRコード誤り訂正のベストプラクティス
- Match error correction to damage risk instead of always choosing maximum redundancy.
- Use Level H for a logo overlay and test on iPhone and Android before mass print.
- Reduce payload length before compensating with extreme error correction on a small label.
- Keep finder patterns and the quiet zone clear in every layout.
- Export print artwork as SVG to preserve sharp module edges; use the SVG QR code generator when preparing vector output.
- Test at the intended viewing distance after the final proof, then use QR code best practices for broader deployment guidance.
開発者は誤り訂正レベルをどう扱うべき？
A production check should include cross-device scans, the final render format, and the actual print substrate. Do not assume a library's level setting alone guarantees a reliable deployment.
QR-Build produces spec-compliant QR symbols for static and dynamic campaigns. Error correction applies to the symbol at encoding time; a dynamic redirect layer and scan analytics operate after scanning. See the QR code generator for developers for programmatic evaluation guidance.
誤り訂正のガイダンスを検証する方法
Our guidance treats recovery percentages as a useful planning range, not an absolute guarantee. Damage location, lighting, focus, print quality, and scanner implementation change outcomes, so teams should verify their generator's current defaults and test their exact final deployment.
QRコード誤り訂正用語集
- Error correction
- Redundant codewords that let a scanner recover data from limited partial damage.
- Reed-Solomon
- The block error-correction algorithm used by QR symbols under ISO/IEC 18004.
- Codeword
- A unit of encoded payload data or redundancy in the QR bit stream.
- Recovery capacity
- The approximate portion of a symbol that may be damaged and still decode under a selected level.
- Level L / M / Q / H
- The four standardized settings with commonly cited recovery approximations of about 7%, 15%, 25%, and 30%.
- Module
- An individual black or white square in the QR matrix.
- Data density
- How much information is packed into a QR version; higher density makes modules smaller at a fixed print size.
- Quiet zone
- The blank margin around a QR symbol; it is essential for detection and is not protected by error correction.
- Finder pattern
- One of the three large corner patterns scanners use to locate and orient a QR symbol.
Related QR Code Guides
QRコードの誤り訂正に関するよくある質問
What is QR code error correction?
QR code error correction is redundant data embedded in the symbol using Reed-Solomon codes. It lets scanners reconstruct the payload when part of the code is dirty, scratched, or obscured within the selected level's recovery range.
What are the QR code error correction levels?
QR codes support L (~7%), M (~15%), Q (~25%), and H (~30%) recovery approximations. Higher levels add redundancy and damage tolerance while reducing payload room and increasing matrix density.
Which QR code error correction level should I use?
Use L for clean digital displays, M for general print, Q for packaging and moderate wear, and H for logo overlays, apparel, or outdoor signage. Match the level to the environment and test the finished code.
How does Reed-Solomon error correction work in QR codes?
An encoder calculates extra codewords from the payload and stores them in the QR matrix. During scanning, the decoder uses those codewords to reconstruct some missing or misread data when damage stays within the available recovery capacity.
Can a scratched QR code still scan?
Often, yes. Level H has more recovery capacity than Level L, but severe damage to finder patterns, the quiet zone, or overall contrast can still prevent scanning.
Does error correction affect QR code size?
Yes. Higher levels add recovery data, making the matrix denser for a given payload. Denser patterns generally need more physical size or a closer scan distance; see the QR code size guide.
What error correction level is best for a QR code with a logo?
Level H is recommended for a centered logo. Keep the logo modest, preserve contrast and the quiet zone, avoid finder patterns, and test the final artwork on multiple devices.
What is the difference between Level H and Level M?
Level M balances capacity and moderate tolerance for general print, while Level H offers the greatest standardized recovery capacity for logos and harsh conditions. Level H also creates a denser symbol and is not a substitute for adequate print size.
How much of a QR code can be damaged and still work?
Common ISO/IEC 18004 guidance is approximately 7% at L, 15% at M, 25% at Q, and 30% at H. Results vary with damage location, contrast, lighting, and camera quality.
Does a higher error correction level improve scan analytics?
No. Error correction affects symbol reliability. Scan analytics come from dynamic redirect and tracking infrastructure, a separate layer described in scan analytics.
Is error correction the same for static and dynamic QR codes?
Yes. Both static and dynamic QR symbols use Reed-Solomon error correction during encoding. A dynamic QR adds an editable redirect layer after the symbol has been generated.
What is the default error correction level in QR generators?
Many compliant generators commonly use Level M as a balance of capacity and tolerance, but defaults vary. Verify output for the final print size, payload length, and environment.
What is QR code error correction for packaging?
Packaging and product labels often benefit from Level Q to tolerate scuffs, handling, and partial ink loss. Test the actual printed substrate before a production run.
Does QR code error correction work for outdoor signage?
Level Q or H is usually appropriate for weather, fading, and dirt. It still requires adequate size, contrast, durable printing, and a test at the intended viewing distance.
What is ISO 18004 error correction?
ISO/IEC 18004 is the international QR symbology standard that defines the L, M, Q, and H Reed-Solomon error-correction structure and recovery-capacity guidance.
印刷環境に合わせたQRコードを作成する準備はできましたか？
QRコードを生成し、最終素材で校正して、人々が実際に読み取る場所でテストしてください。
QRコードを作成