QR Code

What is a QR Code?

A QR Code (Quick Response Code) is a two-dimensional matrix barcode that stores data in a grid of black and white modules. Originally developed by Denso Wave in 1994 for automotive parts tracking, QR codes have become ubiquitous because any smartphone camera can read them.

QR codes can encode up to 7,089 numeric characters or 4,296 alphanumeric characters, and they include Reed-Solomon error correction that allows them to be read even when partially damaged. For GS1 supply chain applications, the most important development is the GS1 Digital Link - a standard that structures GS1 identifiers as web URLs inside a QR code.

Used for: Consumer engagement (product information, marketing), supply chain traceability, retail point-of-sale (via GS1 Digital Link), authentication, and serialisation.

GS1 Digital Link

GS1 Digital Link is a standard that encodes GS1 identifiers as structured web URLs. Instead of carrying a raw number like a traditional barcode, a Digital Link QR code contains a URL where the path segments map directly to GS1 Application Identifiers (AIs).

URL format

The path segments /01/ and /21/ correspond to the same Application Identifiers used in GS1-128 barcodes. This means a GS1 Digital Link URL carries the same structured data as a traditional barcode - but in a format that is also a valid web address.

This dual nature is the key innovation:

• A standard QR reader (any smartphone) opens the URL, which can resolve to a product page, nutritional information, recall notices, or marketing content - enabling direct consumer engagement.
• A GS1-aware POS scanner parses the URL structure and extracts the GTIN , serial number, batch, and other AIs - functioning exactly like scanning a traditional barcode at retail checkout.

How QR Code / Digital Link relates to EPC/RFID

A GS1 Digital Link QR code and an EPC RFID tag can carry the same underlying identifiers. The GTIN and serial number encoded in a Digital Link URL map directly to the fields in SGTIN EPC schemes.

In supply chains that use both technologies, a product might carry a GS1 Digital Link QR code on its consumer-facing label and an SGTIN EPC tag on its RFID inlay - both encoding the same GTIN and serial number, serving different read scenarios.

Sunrise 2027

GS1 Sunrise 2027 is a global initiative positioning GS1 Digital Link QR codes as the eventual successor to traditional EAN/UPC barcodes at retail point of sale. By 2027, GS1 expects retailers worldwide to be able to accept 2D barcodes - including Digital Link QR codes - at checkout.

The transition does not mean EAN/UPC barcodes will disappear overnight. Instead, Sunrise 2027 establishes 2D readiness as a baseline expectation: POS systems should be capable of reading both legacy 1D barcodes and new 2D symbols. Over time, brands can migrate to QR codes that serve a dual function - marketing and supply chain in a single symbol - rather than printing separate barcodes and QR codes on the same package.

For RFID-enabled supply chains, this convergence is significant. As the same GS1 identifiers flow through Digital Link QR codes, GS1-128 barcodes, and EPC RFID tags, the three technologies become interoperable views of the same data.

Where QR Code / Digital Link is used

• Consumer engagement: Product pages, how-to videos, ingredient transparency, sustainability information, and loyalty programmes - all accessible with a smartphone scan.
• Retail POS (Sunrise 2027): As retailers upgrade to 2D-capable scanners, Digital Link QR codes will replace or supplement traditional EAN/UPC barcodes at checkout.
• Traceability: Encoding batch, lot, and serial data in the URL enables farm-to-fork and pharmaceutical track-and-trace without requiring specialised scanning hardware.
• Authentication: Serialised Digital Link QR codes can be verified against a cloud database to detect counterfeits.
• Healthcare: Pharmaceutical packaging increasingly uses 2D barcodes for serialisation under regulations like the EU Falsified Medicines Directive and US DSCSA.

Construction

New to these terms? module , quiet zone , X-dimension , symbology - the linked pages cover each one from first principles.

Versions and module count

QR Code (Model 2, the universal version) defines 40 versions. Each version is a fixed-size square; the module count per side grows linearly from version 1 (21 modules) to version 40 (177 modules):

modules_per_side = 21 + 4 × (version - 1)

The Annex 1 Micro QR variant defines its own four smaller sizes (M1–M4: 11×11 to 17×17) that do not follow this formula and are encoded differently - see the Micro QR page for details.

Symbol structure

Every QR Code is composed of fixed function patterns plus the variable data region:

• Finder patterns - three nested 7×7 concentric squares placed in the top-left, top-right, and bottom-left corners. They have a distinctive 1:1:3:1:1 dark-to-light ratio when scanned through their centre line, which the decoder uses to locate the symbol and determine its rotation.
• Separators - a single-module-wide light border around each finder pattern.
• Alignment patterns - 5×5 concentric squares placed at fixed coordinates that depend on the version. Version 1 has none; version 2 has one (centred at module (18,18)); higher versions add a grid of alignment patterns. They let the decoder correct for perspective distortion when the symbol is photographed off-axis.
• Timing patterns - one row and one column of alternating dark/light modules running between the finder patterns at row 6 and column 6. The decoder uses them to count modules across the symbol regardless of physical size.
• Format information - a 15-bit BCH-encoded value duplicated in two places around the finder patterns, declaring the error-correction level (L/M/Q/H) and the mask pattern applied to the data region.
• Version information - an 18-bit BCH-encoded value carried in two locations for versions 7–40; versions 1–6 derive their version from the module count alone.
• Data and EC modules - everything else. Encoded payload + Reed-Solomon error-correction codewords, placed into the remaining modules by an interleave/zig-zag rule.
• Quiet zone - ≥ 4× X-dimension on all four sides. Notably larger than the 1× quiet zone of Data Matrix.

Capacity by version

Pick a version below to see its module count, total codewords, and maximum payload at each error-correction level for all three character-set modes. Toggle the full 40-version table underneath to compare versions side-by-side.

Source: ISO/IEC 18004:2015 Table 7 (capacity per mode and EC level) and Table 9 (codewords + block structure), cross-checked against the canonical encoder tables in @rfid-me/epc-parser (VERSION_EC_TABLE / BYTE_CAPACITY in src/barcode/encoders/qr-tables.ts; epc-parser ships exact tables for versions 1–10).

Encoding modes

The data stream is split into one or more segments, each prefixed by a 4-bit mode indicator and a character-count field. Switching modes mid-stream is supported and is used by encoders to minimise the total bit length.

GS1 Digital Link payloads (HTTPS URLs) are encoded in Byte mode. The FNC1 mode indicator (0101) at the start of the symbol switches the decoder to GS1 QR Code behaviour, where the data stream is interpreted as concatenated GS1 Application Identifiers (the same payload model as GS1-128 and GS1 DataMatrix). A QR symbol can carry GS1 data either as a Digital Link URL (Byte mode) or as raw AIs (FNC1 mode) - both are valid and both decode to the same logical AI set.

Error correction

QR Code uses Reed-Solomon error correction (see the Reed-Solomon guide for how the algorithm computes parity symbols and recovers from damage). The codeword arithmetic happens in the Galois field GF(256) with primitive polynomial 0x11D - the same field as Data Matrix uses, but with a different polynomial (Data Matrix: 0x12D). A codeword in this context is the 8-bit unit - one symbol of the Reed-Solomon code - that the data region is broken into.

Four user-selectable error-correction levels trade payload capacity for damage tolerance:

• L (Low) - ~7% of codewords recoverable
• M (Medium) - ~15%
• Q (Quartile) - ~25%
• H (High) - ~30%

For larger versions the data + EC stream is split into one or more blocks (independent Reed-Solomon code words) and the bytes from each block are interleaved before placement - byte 0 of block 1, byte 0 of block 2, byte 1 of block 1, byte 1 of block 2, etc. Interleaving spreads any localised defect (a coffee stain, a finger smudge, a torn label edge) across fewer codewords per block, so each block stays inside its individual recovery threshold. Block parameters are version- and EC-level-specific - ISO/IEC 18004 Table 9 lists all 160 combinations.

Data masking

After encoding and Reed-Solomon, the data region is XORed with one of eight fixed mask patterns (mask 000 through 111). Each mask is a fixed black-and-white grid defined by a Boolean formula over the module's row (i) and column (j) coordinates - modules whose coordinates satisfy the formula are flipped.

Masking has nothing to do with security or scrambling: it exists to break up unfortunate coincidences in the bit pattern that would confuse a scanner - for example, a long horizontal run of dark modules, or a fragment that happens to look like a finder pattern (the 1:1:3:1:1 ratio). The chosen mask index is recorded in the 3-bit mask portion of the format information field (a 15-bit BCH-encoded value carried twice around the finder patterns - see "Symbol structure" above).

The eight mask patterns

Source: ISO/IEC 18004:2015 §7.8.2 Table 10. i is the row and j is the column, both numbered from 0 at the top-left of the symbol. ⌊x⌋ means floor (integer division).

Penalty score (mask selection)

The encoder applies each of the eight masks in turn, scores the result against four penalty rules, sums the four rule scores, and picks the mask with the lowest total. The four rules each target a different scan-failure mode:

Source: ISO/IEC 18004:2015 §7.8.3.1. The encoder evaluates the sum N1 + N2 + N3 + N4 for each of the eight masks (so eight totals) and writes the index of the lowest-scoring mask into the format information field.

X-dimension and quiet zone

GS1 Sunrise 2027 sets the minimum X-dimension for retail-POS scanning at 0.396 mm with up to 0.991 mm for general distribution; healthcare DPM applications permit smaller. The 4× quiet zone on each side is mandatory - reducing it below 4× causes intermittent decode failure even when print quality is otherwise fine.

Key standards

The base QR Code symbology is defined by ISO/IEC 18004 - its Annex 1 covers the compact Micro QR variant used when the standard symbol is too large for the available label area.

Implementation in printer languages

Related EPC schemes

Source

GS1 Digital Link Standard, GS1 2D Barcodes initiative, and the GS1 General Specifications. The mapping between Digital Link AIs and EPC schemes is defined in the GS1 EPC Tag Data Standard, Release 2.3.