Question 1

What does the check under the preview actually do?

Accepted Answer

It draws the symbol again at the size the PNG will be exported, reads the middle of every module straight off those pixels, and runs what it finds back through a decoder, then compares the result against what you typed. That is a different thing from the contrast grade beside the colour pickers, which is arithmetic on two hex values and has never looked at the image. One detail is worth knowing: inverting a barcode leaves every run length identical, so Code 128, Code 39, ITF and Codabar all decode from a negative exactly as they do from the real thing. Reversed colours therefore cannot be caught by decoding, and the quiet zone is measured separately to settle which end of the range is the background.

Question 2

What is GS1-128, and why does it need those brackets?

Accepted Answer

A GS1-128 is an ordinary Code 128 that opens with an FNC1 and carries its data as element strings, each beginning with a two to four digit Application Identifier saying what follows it: (01) is a GTIN, (17) an expiry date, (10) a batch number. The brackets are how the data is written on paper and in documentation, and they are never encoded. What is encoded instead is an FNC1 separator after every element whose length is not predefined, which is what tells a reader where one field ends and the next begins.

Question 3

Why does the tool refuse an Application Identifier it does not recognise?

Accepted Answer

Because the alternative is silently wrong data. Whether an element needs an FNC1 after it is decided by the first two digits of its AI, and there is no way to work that out from the AI itself: (402) carries a fixed seventeen digits and still needs a separator, because the prefix 40 is not one of the predefined-length prefixes. Guess that wrong in either direction and the symbol still scans, it just hands the receiving system fields that have run into each other. So an AI whose length rule has not been checked is refused with a message rather than encoded on an assumption.

Question 4

How do I print a whole sheet of labels?

Accepted Answer

Switch to Batch, paste one value per line, and open the printable sheet. It lays the symbols out two, three or four across in a new tab, each sized in millimetres at 300 dpi so it comes off the printer at the width the size table gives, and the print dialog will also save it as a PDF. One thing to avoid: the fit-to-page option rescales everything, which changes the physical width of the narrow bar and is the quickest way to turn a good sheet into an unreadable one.

Question 5

Which barcode format should I pick?

Accepted Answer

If the code is yours and only your systems read it, Code 128 is the right default: it holds the whole ASCII set, packs digits two to a symbol so numeric codes stay short, and every scanner made in the last thirty years reads it. If the product is going to be sold through a retailer, the format is not a choice: EAN-13 outside North America, UPC-A inside it, and the number itself has to be one your national GS1 office issued. ITF-14 is for the shipping carton rather than the item, Code 39 turns up in older industrial and defence systems that still specify it, and Codabar survives in libraries, blood banks and photo labs.

Question 6

Can I make up my own EAN or UPC number?

Accepted Answer

You can generate a valid symbol from any digits, and this tool will happily draw one, but a made-up number is not a usable retail barcode. The first digits of an EAN or UPC identify the company that owns the number, and those prefixes are issued by GS1 member organisations for an annual fee. A retailer scanning an invented code either finds nothing in their catalogue or finds somebody else's product. For a code that only moves inside your own warehouse, use Code 128 and skip the whole question.

Question 7

What is the check digit, and why did my number change?

Accepted Answer

Every EAN, UPC and GTIN ends in a digit computed from the ones before it: multiply the digits alternately by 3 and 1 counting from the right, add them up, and take whatever gets you to the next multiple of ten. It exists so a scanner that misreads one bar produces a number that fails the arithmetic instead of a number that quietly belongs to a different product. Paste twelve digits into EAN-13 and the thirteenth is added; paste thirteen and it is checked, which is why a typo shows up here as an error rather than as a barcode.

Question 8

Why does my barcode fail to scan after I print it?

Accepted Answer

Three causes cover almost all of it. The quiet zone has been cropped, so the scanner cannot tell where the symbol begins. This tool draws that margin into the image, so leave the white space that comes with the file alone. The symbol has been printed too small, which the size table underneath the preview is there to catch: below the 0.264mm GS1 sets as the point-of-sale floor, ink spread starts closing the gaps. Or the code was scaled non-uniformly, which happens when an image is dragged by a corner handle in a layout tool and stretched more one way than the other.

Question 9

Can I make the barcode any colour?

Accepted Answer

A scanner reads a barcode by measuring reflected light, so what matters is not hue but how far apart the bars and the background sit in brightness. Dark bars on a light background is the only combination that is reliably safe, and ISO/IEC 15416 grades a printed symbol partly on that gap: 40% is the floor most readers are specified for. The panel here works out the figure as you pick the colours and says which grade you have landed on. Light bars on a dark background look striking and are refused by most readers outright, since laser scanners in particular are built to expect the opposite.

Question 10

How tall should a barcode be?

Accepted Answer

Tall enough that a scanner sweeping across it at a slight angle still crosses every bar. The rule of thumb the standards work from is about 15% of the symbol's width, and EAN and UPC have a nominal height fixed by GS1 rather than left open. Truncating a barcode to a thin strip to fit a package is a common request and a common cause of slow checkouts: the code still reads, but only when the beam happens to cross it squarely, so the person at the till ends up passing it over the glass three times.

Question 11

What do the three Code 128 subsets do?

Accepted Answer

Code 128 has three character sets sharing one set of bar patterns. A covers capitals, digits, punctuation and the ASCII control codes; B swaps the control codes for lower case; C encodes nothing but digits, and packs two of them into each symbol, which makes a numeric code roughly half as wide. Auto is what you want almost always: it starts in whichever set fits, drops into C whenever four or more digits come along, and switches back for the letters, so it produces the shortest valid symbol without you thinking about it. Choose a subset by hand only when the system reading the code insists on one.

Question 12

What is the bearer bar on an ITF-14 for?

Accepted Answer

ITF encodes digits in pairs, half in the bars and half in the spaces, and it has no per-character structure that tells a reader where a character ends. That makes it uniquely vulnerable to a short scan: a beam crossing the symbol at an angle can leave partway through, and what it collected can still look like a complete, valid, shorter number. The bearer bar is the frame around the symbol, and a scan that exits through it is discarded rather than reported. On a corrugated carton, where print quality is poor and the label is scanned in motion, it is not decoration.

Question 13

Should I download the PNG or the SVG?

Accepted Answer

PNG for anything with a fixed size: a label printer, a spreadsheet, a screen, a slide. SVG whenever the symbol will be resized afterwards, because it stores the bars as geometry rather than as pixels and stays exact at any scale. That matters more for barcodes than for most graphics: a raster barcode enlarged in a layout tool gets soft edges between bar and space, and a scanner reading the edge of a bar in the wrong place is exactly what the standards' print quality grades are measuring.

Question 14

What is the difference between UPC-E and UPC-A?

Accepted Answer

UPC-E is a UPC-A with a run of zeros squeezed out, so it fits on small packaging where a full UPC-A will not. Which zeros come out is decided by the sixth digit, and the process reverses exactly, so a scanner reading a UPC-E hands the till the full twelve-digit UPC-A. Two details surprise people: only numbers with the right pattern of zeros can be compressed at all, and the check digit belongs to the expanded number rather than to the six digits printed on the label, which is why it cannot be worked out from what you see.

Question 15

Can a barcode hold a website address or a name?

Accepted Answer

Code 128 and Code 39 hold text, so a short URL fits, but a linear barcode gets physically longer with every character and a scanner has to see the whole width at once. A forty-character link becomes a symbol too wide to read comfortably from a phone. Retail formats cannot do it at all: EAN, UPC and ITF hold nothing but a fixed count of digits. Anything longer than a short code belongs in a QR code, which stores data in two dimensions and stays compact.

Question 16

Why does Code 39 print asterisks around my text?

Accepted Answer

The asterisk is Code 39's start and stop character, and it is part of every symbol rather than an option. Older scanners transmit it along with the data, which is why software reading Code 39 often has to strip it, and why the printed line under the bars conventionally shows it. Code 39 also has an optional modulo 43 check character, off by default here because a lot of systems that use Code 39 do not expect one and will read it as an extra character of data.

Printed at	Symbol width	Narrow bar
96 dpi	122.2 mm	0.53 mm
300 dpi	39.1 mm	0.17 mm · too fine
600 dpi	19.6 mm	0.08 mm · too fine

tooldura

Barcode Generator

About Barcode Generator

How to use Barcode Generator

Frequently Asked Questions