QR Code & Barcode Types Reference
QR codes and barcodes bridge the physical and digital worlds. This page summarizes common symbologies, capacities, use cases, QR error correction levels and the standard WiFi QR format.
Practical Notes
- QR error correction trades off against capacityLevel H (about 30% recovery) survives heavy occlusion and print wear, but holds noticeably less data at the same physical size than level L. Use H for posters where the code may be obscured, and M/L for on-screen use.
- Longer content means denser modules and harder scanningAt a fixed QR size, more characters means smaller modules, which defeats low-resolution cameras. Shorten long URLs with a link shortener or use shorter field names instead of just printing bigger — the available layout limits that anyway.
- 1D barcodes carry check digits; hand-built values often failEAN-13 and CODE128 include check digits or special start/stop symbols, so concatenating digits by hand yields an unscannable code. Always generate through a library, and error out on check failure instead of emitting silently.
Data Sources
- ISO/IEC 18004 (QR Code)
- ISO/IEC 15417 (Code 128)
- ISO/IEC 15420 (EAN/UPC)
Codes, names and definitions above come from public specifications; we claim no originality over the specification data itself. The arrangement, Chinese annotations and practical notes on this page are written by us.
How It Works
Barcodes and QR codes essentially encode data into machine-readable graphical symbols. A 1D barcode (e.g. CODE128, EAN-13) represents characters only horizontally with bars and spaces of varying width; it carries little data and needs a specific scan angle. A 2D code (e.g. QR Code) encodes in both directions, offering large capacity, error recovery and phone-camera readability at any angle.
A QR error correction level (L/M/Q/H, about 7/15/25/30 percent recovery) uses Reed-Solomon redundancy: redundant information is scattered across the pattern so the original data can still be recovered even when partly obscured, damaged or overlaid with a logo. Higher levels make the pattern denser and reduce capacity for the same payload, so trade capacity against robustness.
When to Use This Table
This table is for choosing a barcode symbology or a QR error correction level. It puts QR Code, Code 128, EAN-13, Code 39, Data Matrix and Aztec side by side by type, capacity and typical use, and explains the L/M/Q/H correction levels.
When not to consult it: you are designing the complete print and scanning scheme. Symbology choice depends on scanning distance, print process, whether the interaction is bidirectional (can scanning return data), and whether target devices support a symbology at all — that needs real testing rather than lookup.
Rule of thumb: decide on bidirectionality first. A one-dimensional code stores data and the scan simply ends; a QR code can carry a URL or command and supports error correction, which suits interactive use. Recording an identifier such as an order number is more efficient in a 1D code, while anything where a customer scans with a phone (open a menu, pay, view details) must be a QR code.
Mind the trade-off between correction level and capacity: level H tolerates 30% damage but shrinks capacity to roughly 55% of level M. Choosing H blindly for long content inflates the symbol until it exceeds the resolution of the printer or screen and stops scanning.
Boundaries and Common Pitfalls
Higher correction means less capacity, and level H can stop scanning for long content
Correction works by redundancy: about 7% for L, 15% for M, 25% for Q and 30% for H, so damage tolerance rises while the same content needs a larger symbol. With long content at level H the module density exceeds what a screen or printer can resolve, and scanning fails entirely at a distance or under glare. Pick the smallest level that suffices — L or M for short clean content, Q or H where print will wear — and switch to Data Matrix or Aztec when density is the binding constraint.
Static and dynamic barcodes are different product shapes
A static barcode prints the content directly on the label, one code per item, extremely cheap and suited to shelves and fixed assets. A dynamic barcode, or a QR pointing at a short link, resolves to a server-side record looked up at scan time, which makes the content mutable, measurable and revocable — a delisted or recalled item only needs a database change, not a label swap. The trade-off is that the server and link service become dependencies, so a label failure or expired link directly affects fulfilment.
A missing quiet zone is a frequent cause of unreadable codes
Every QR spec requires a quiet zone around the symbol, normally four module widths. It is not cosmetic: decoders use it to locate the symbol boundary, and without enough margin a scanner may report no code found or decode the wrong content entirely. Printing that crops edges, wraps a symbol onto a container corner, or butts it against text all destroy the quiet zone. Design files should include it in the canvas and leave real whitespace around the code.
How This Relates to Neighbours
To generate or decode QR codes, use the in-site QR tools; this page is a selection reference, not a generation pipeline.
Print processes (thermal, laser, inkjet) and scanner hardware selection are outside this table.
Designing the availability of a short link service behind dynamic codes is backend architecture and needs a fallback plan.
If the goal is debugging an unreadable code, measure the quiet zone and module width first, then contrast and damage, and only then suspect the wrong symbology.
Practical Snippet
Key points: (1) pick the smallest sufficient correction level by content length (L/M/Q/H give roughly 7/15/25/30% redundancy) — a long payload at level H pushes module density past what scanners resolve and stops reading; (2) always emit a full quiet zone (margin: 4), since decoders rely on it to locate the boundary, and cropping or wrapping onto a corner destroys it; (3) before printing, check the minimum module size in millimetres — at 300dpi keep it above 0.33mm; (4) choose between dynamic codes (short link to a server record, mutable and revocable but with a service dependency) and static codes (cheap and dependency-free but requiring a relabel when content changes) based on fulfilment needs.
// ① Pick the error-correction level by content length (higher is not always better)
function pickEcc(textLength) {
if (textLength < 100) return 'L' // about 7% redundancy
if (textLength < 300) return 'M' // about 15%
if (textLength < 600) return 'Q' // about 25%
return 'H' // about 30%; only for short content that must survive damage
}
// ⚠️ Choosing H for long content ⇒ module density exceeds the resolution limit ⇒ unscannable
// ② Always leave a full quiet zone (usually 4 modules)
import QRCode from 'qrcode'
const png = await QRCode.toBuffer('https://example.com/p/12345', {
errorCorrectionLevel: 'M',
margin: 4, // ← quiet zone; setting 0 makes many scanners miss the boundary
width: 512, // Physical size must be large enough
})
// ③ Self-check before printing: module count sets the real size of each module on a fixed label
function checkPrintable(svg, labelWidthMm = 40) {
const m = svg.match(/viewBox="0 0 (\d+) (\d+)"/)
if (!m) return { ok: false, reason: 'No viewBox, cannot compute module size' }
const modules = Number(m[2]) - 8 // Subtract the 2×4 quiet zone
const minMm = labelWidthMm / modules // Physical edge per module; should be ≥ 0.33mm at 300dpi
return { ok: minMm >= 0.33, modules, minMm: minMm.toFixed(2) }
}
// ④ Dynamic codes: short link + backend record (changeable and revocable)
const code = await QRCode.toDataURL('https://s.example.com/' + token, { margin: 4 })
// Delisting or recalling a product only changes a DB record; labels stay
// ⑤ Static codes: content encoded directly (cheap, no dependency)
const staticCode = await QRCode.toBuffer(JSON.stringify({ sku: 'A-123', lot: 6 }), {
errorCorrectionLevel: 'Q', margin: 4,
})
// ⚠️ Content is immutable ⇒ any change means new labels
// ⑥ Use 1D barcodes for pure identifiers (smaller label area)
import JsBarcode from 'jsbarcode'
// Code128 holds variable-length strings, EAN-13 only 13 digits
JsBarcode(svgEl, '1234567890123', { format: 'EAN13', margin: 10 })Common Symbologies
| Symbology | Type | Capacity | Best For |
|---|---|---|---|
| QR Code | Matrix 2D code | 7,089 numeric / 4,296 alphanumeric / 2,953 bytes | URLs, payments, contact cards, WiFi connections, mobile |
| Micro QR | Compact matrix 2D code | 35 numeric / 21 alphanumeric | Space-constrained parts, IC labels |
| CODE128 | Linear barcode | ASCII 128 charset, high density | Logistics, warehousing, serial numbers |
| EAN-13 | Linear barcode | 13 digits incl. checksum | Retail products, book ISBN |
| UPC-A | Linear barcode | 12 digits incl. checksum | North American retail products |
| CODE39 | Linear barcode | Digits, uppercase, some symbols | Industrial, defense, asset labels |
| ITF-14 | Linear barcode | 14 digits | Outer case packaging, logistics units |
QR Code Error Correction
| Level | Recovery | Suggested Use |
|---|---|---|
| L | ~7% | Short content, clean scan env |
| M | ~15% | Default, balanced capacity/recovery |
| Q | ~25% | Average print quality, higher recovery |
| H | ~30% | Logo overlay, harsh environments |
Frequently Asked Questions
Which QR Code error correction level should I use?
Default to M (15% recovery) for the best balance of capacity and redundancy. Use H (30%) when placing a logo in the center or printing on low-quality surfaces. Use L (7%) only for short content in clean scanning environments. Higher levels produce denser patterns for the same payload.
What is the difference between 1D and 2D codes?
1D barcodes store data in one direction only, carry little data (often tens of characters), require a specific scan angle and are suited for short codes like SKUs and tracking numbers. 2D codes store data both horizontally and vertically, offer large capacity and error recovery, and can be read by phone cameras at any angle, making them ideal for URLs, payments and contact cards.
What is the WiFi QR code format?
The standard format is WIFI:T:<auth>;S:<SSID>;P:<password>;H:<hidden>;;, e.g. WIFI:T:WPA;S:MyNetwork;P:MyPassword;;. Phones can scan it and connect automatically without typing credentials.
Can barcodes contain Chinese characters?
Common 1D barcodes (CODE128, EAN, UPC) only encode digits, letters or ASCII symbols and cannot represent Chinese directly. Use a QR Code for Chinese text because it supports UTF-8 encoding.