QR Code is capable of encoding four types of data, as listed below. Note that the interpretation of byte characters is subject to the ECI, which is assumed to be ISO8859-1 if no ECI is defined in the symbol.

QR Code defines several encoding modes, with the numeric set the most efficient and the Kanji set the least. Internally, the encoder switches the modes automatically based on the nature of the data, with the goal of minimizing the symbol size required. The algorithm produces the optimal results in most cases.

When space is a concern, it is desirable to adjust the character set of data encoded so that they can be more efficiently encoded. For example, the alphanumeric character set in QR Code does not include lower case letters. If the string encoded is case insensitive, it reduces space requirement by converting all letters upper case.

QR Code also defines several advanced features that can only be encoded with escaped format, referred as tilde codes throughout the manual. These advanced features include:

Three options affect the final shape of the symbol generated:

The Center Overlay Block feature, also referred as feature 50 in this manual, allows user to specify a bit pattern block that is overlaid on the center of the QR Code. This feature is required to implement QR Code for Swiss payment, where a Swiss logo is required to overlay on the center of the barcode.

Because QR Code has error correction capability, overlaying a small image won't affect the reading. However, care must be taken to avoid situations that the image is too big to cause the whole barcode unreadable.

The following are the two examples of center overlay block. In the first one, the overlaid bit pattern is expressed as rows, cols and pattern. In the second one, :swiss7 indicates to overlay 7 mm x 7mm swiss logo when the overall symbol size is 46mm x 46 mm (the size required by Swiss payment standard). Both starts with tilde code 50, so this feature is also referred as Feature 50.

~50[7,7,007c6c446c7c]

~50[:swiss7]

For more information about Center Overlay Block, see Appendix E, Center Overlay (Feature 50) Support.

QRCode can encode Kanji text natively. A Kanji character consists of two bytes and has a value either between 0x8140 and 0x9ffc, or between 0xe040 and 0xebbf. Kanji characters are inputted using Kanji block, which starts with ~4 following by the byte values of kanji characters enclosed in square brackets. Byte values can be escaped using ~dnnn format. For example, the following code is the input for text

89~4[~d147~d250~d150~d123]

The Kanji characters have hex values 0x93fa and 0x967b - whose corresponding decimal values are 147, 250, 150 and 123.

ECI was introduced to allow mixed character sets in a QRCode symbol, and to enable different interpretations on the output data stream.

An ECI can be any number between 000000 and 999999. The tilde code sequence ~7nnnnnn is used to enter an ECI value. The code sequence can appear at any places of the input, but there must be exact 6 digits following ~7. For example to start an interpretation of 10, enter ~7000010.

In addition to ~7 (ECI basic), ~8 and ~9 can be used to represent Nesting-Begin and Nesting-End. The Nesting-Begin ECI does not turn off the current ECI in force until Nesting-End is encountered at the current nesting level. This allows multiple ECIs (for example, Greek language plus encryption) effective in a data segment.

The table below lists common ECI values and their corresponding character encoding. Note that some character encodings may have multiple ECI values. For example, you can use either 1 or 3 for ISO8859-1 encoding.

The structural append feature allows up to 16 symbols in a structure. A capable reader buffers the contents of each symbol until all symbols are read.

To encode structural append, supply the three required fields for each symbol:

The tilde code sequence is expressed in the following format:

~2[SI][TS,P]

For example, tilde code sequence ~2[1][6,89] indicates that the current symbol belongs to a group that has 6 symbols in total. The parity value for input is 89.

The structural append block must appear at the beginning of the input.

Structural append is not supported in Micro QR code.

FNC1 (first position) is used when encoding GS1 structural data (GS1-128) in QR Code symbols. The leading FNC1 symbol is required to indicate the data followed is a GS1 structured data.

When encoding GS1 structured data, it is user's responsibility to ensure that the data supplied is correct (data length and checksum), and a separator is present after a variable length data item. QRCode standard requires % character to be used as separator, instead the FNC1 character as in the case of Code 128.

For example, to encode GS1 structural data (01)04912345123459(15)970331(30)128(10)ABC123, which denotes GT IN number 04912345123459, Sell By date 97/03/31, Quantity 128 and Lot number ABC123, the input should be:

~101049123451234591597033130128%10ABC123

FNC1 First Position is not supported in Micro QR code.

FNC1 (second position) is an AIM extension to encode structural data similar to GS1. The encoder needs to know the AI in order to create an encoding. Therefore, it is required to put AI in brackets immediately after tilde code ~3.

FNC1 Second Position is not supported in Micro QR code.

QR Code 2005 does not support unicode natively. The default character set is ISO8859-1. Theoratically the support could become available using ECI; however most scanners and applications do not have ECI support builtin.

Recently in mobile industry the trend is to use UTF8 to encode Unicode string into QR Code. Such feature requires support of the reader. We decided to add Unicode support in a way compliant with Android/IPhone barcode reader. Therefore, you need to have compliant reader in order to decode properly.

This feature is added in 5.1, released in May 2012. Enhancement is made for components that take UTF16 string input, such as QR Code ActiveX control and Word Add In. If all characters in the input string falls into ISO8859-1, the resulted barcode remains the same. Otherwise the input is converted to UTF8 with BOM first, and then encoded. In Font encoder DLL, additional function QRCodeEncode2W is added to perform such operation. More information is available at Appendix D, Unicode String Encode Support.

Creating barcodes using fonts involves two distinct processes: encoding and rendering. Rendering is to choose the appropriate font and font size and format the encoding results. Encoding is to convert the data into a special string, refereed as Barcode String, which becomes a barcode after being formatted with the font.

The font-based approach is easy to understand, and in some occasions the only way to add barcode printing functionality, such as in Crystal Reports.

The drawing below illustrates how you create a QR code using font-based approach. You first create the barcode string (which is an array of text lines). Format the barcode string with MRV QRCode font you get the barcode.

The font characteristics are listed as below:

You can export barcodes as standard image files to be consumed in other word processing and imaging programs. The QRCode Encoder GUI supports exporting images in four formats via a GUI interface. The Encoder DLL and QRCode ActiveX Control also provide programming interfaces to export images in those formats:

Special attention should be paid when the target printer has low resolution, such as a thermal printer or fax machine. When generating vector graphics files, the software allows you to specify target dpi which you should set to the one that matches your printer. It is generally not recommended to scale the image down or up when working with low-resolution printers.

For environments that support ActiveX Controls (also called OLE controls), you can insert the QRCode ActiveX Control into documents, spreadsheets, or invokes the control at the background in your program. With ActiveX Control you are able to change the barcode image by editing its properties.

Using ActiveX Control in Microsoft Office programs is straightforward. Programming knowledge is required in order to integrate the control with your custom programs.