QR Code Scanner Online

How QR decoding works

QR Code Format Variants and Compatibility Limits

The QR Code standard (ISO/IEC 18004:2015) defines 40 symbol versions, ranging from 21×21 modules (Version 1) to 177×177 modules (Version 40). Version 40 stores up to 7,089 numeric characters, 4,296 alphanumeric characters, or 2,953 bytes of binary data. Most smartphone cameras max out reliable decode at Version 20 under real-world conditions. Micro QR codes (ISO/IEC 18004 Annex E) use versions M1–M4 and cap at 35 numeric characters; scanners that handle standard QR often fail on Micro QR entirely. QR Code Model 1 (the original 1994 spec) is structurally different from Model 2 and lacks the alignment patterns present from Version 2 onward - scanners built against the Model 2 spec will misread or reject Model 1 codes silently.

Error correction operates at four levels: L (7% recovery), M (15%), Q (25%), and H (30%). A code printed at level L with a physical tear covering more than 7% of its area produces an unrecoverable decode failure. Payment QR codes in India's UPI system mandate level M minimum; GS1-compliant retail codes require level M by GS1 General Specifications v24.0 section 5.5.3.

Image Quality Thresholds That Cause Decode Failures

JPEG compression is the single most common cause of scan failures on uploaded images. At a JPEG quality setting below 60 (out of 100), blocking artifacts corrupt the module boundaries in the quiet zone and finder patterns. PNG is lossless and always preferable for QR uploads. Minimum recommended resolution for reliable software decode is 5 pixels per module; a Version 5 code (37×37 modules) therefore needs at least 185×185 pixels of source image. Below 72 PPI effective resolution, ZXing-based decoders (the engine behind most browser QR tools) report a 40–60% failure rate on codes with dense payloads.

Contrast ratio matters independently of resolution. The ISO 15415 print quality standard sets a minimum symbol contrast of 70% between dark modules and the background. Codes printed on kraft brown stock or metallic foil frequently fall below this threshold. Rotating the image 45° before upload resolves roughly 12% of scan failures caused by perspective distortion in photographs of printed codes - the decoder's finder-pattern search runs more effectively against a skewed grid than a near-vertical one.

Data Payload Types and Parsing Risks

Raw QR payloads are unstructured byte strings. The encoding mode byte at position 0–3 signals the data type: 0001 = numeric, 0010 = alphanumeric, 0100 = byte (UTF-8 or Latin-1), 1000 = kanji. A payload beginning with SMSTO: triggers an SMS draft on most Android devices without user confirmation beyond the messaging app opening. A WIFI: payload can auto-join a network. Payloads beginning with javascript: or data: are rejected by all major browsers when scanned natively, but third-party apps vary. For static images where you cannot verify origin, scan without a camera and inspect the raw string before following any embedded URL.

GS1 structured payloads use Application Identifiers (AIs) wrapped in FNC1 mode. AI (01) is a 14-digit GTIN; AI (17) is an expiry date in YYMMDD format. Misreading a GS1 code as plain text produces a string like ]Q3010123456789012817261231 - the ]Q3 prefix is the symbology identifier, not payload data, and most generic parsers discard it incorrectly, shifting every AI offset by three characters and producing corrupted field reads.