Morse Code: How to Encode and Decode It in Twelve Alphabets

By Ali · Washington, DC

A message in Morse code looks like noise until you know three things: which signals exist, how much silence separates them, and which alphabet the sender had in mind. Get those right and a line like .... . .-.. .-.. --- / .-- --- .-. .-.. -.. reads as easily as plain text. Get the third one wrong and the same dots turn into a different language entirely.

This guide covers the system as it is actually sent, then shows the working mechanics of encoding and decoding with a browser tool that carries twelve published alphabets.

What the signals are, and what the gaps mean

Morse code has two signals and three gaps, all measured against one base unit.

That is the whole protocol at the signal level. The international version was standardised in 1865 and is now published as ITU-R M.1677-1, which fixes the letter and numeral assignments and the timing ratios above.

The spacing is not cosmetic. Inside a character the signals are separated by a single unit, so ... is three dits and reads as S, while . . . with element gaps between them reads as three separate E characters. Copy that distinction loosely and the message changes meaning without a single wrong dot.

Why some letters are short and some are long

The Latin table was designed around printing frequency, in the era of hand-set type. E, the most common letter in English, is a single dit. T is a single dah. A, I, M and N, all common, use two symbols. Rare letters like Q, Y and Z run to four.

The practical consequence is speed. A trained operator sending ordinary English needs fewer key presses per word than the raw letter count suggests, because the frequent letters are cheap. This is also why Morse survives as a human skill: the code is optimised for hands and ears, not for machines.

Numerals are uniform at five symbols each, which makes them slower per character but impossible to confuse with letters at speed. Punctuation uses six symbols and a small number of mixed patterns; the full stop is .-.-.-, the comma --..--, the question mark ..--...

Timing, WPM and the PARIS reference

Speed is measured in words per minute, and the word is notional. The standard reference word is PARIS, fifty units long when its letters and gaps are counted. That single number defines the arithmetic:

• 1 WPM equals 1.2 dots per second;
• 12 WPM equals 14.4 dots per second;
• 20 WPM equals 24 dots per second;
• 25 WPM, the usual amateur radio target, equals 30 dots per second.

Manual sending usually lands between 5 and 40 WPM. Below about 12 WPM operators tend to count symbols individually; above 20 the letters start arriving as shapes, which is why learning methods switch to whole-character recognition early.

Farnsworth timing is a training arrangement, not a different standard. The characters keep their full speed, but the gaps between them are stretched so a beginner has time to write each one down. A learner at 18 WPM character speed with 10 WPM effective spacing hears crisp letters and generous pauses.

Twelve alphabets, one tool

A dot-dash pattern is a code, not a drawing of a letter. Four dots are H in Latin, Х in Cyrillic, and the numeral 5 in a numeric context. Every translator therefore needs to know which table is in play before it can decode anything.

The converter on this site carries twelve sets, and their sizes are not equal:

The mappings follow different logic. Greek assigns codes by resemblance and sound, so alpha takes Latin A and beta takes Latin B even though beta sounds like V. Cyrillic maps letters to the Latin letters that look closest, which is why Привет encodes as . --. .-. .. .-- . - and comes back as ПРИВЕТ. Arabic and Persian build codes per letter without regard to Latin shapes. Japanese uses the Wabun table for katakana. Korean sends individual jamo, so a syllable block is written as its component letters rather than as one unit.

Encoding a message step by step

Encoding is the simple direction, and three habits prevent most errors.

1. Choose the alphabet before typing. Text in Cyrillic encoded under the Latin table loses its meaning even though the output looks valid.
2. Keep words separate. The tool inserts a slash with a space on each side, matching the seven-unit gap. Removing those separators turns a sentence into one long string that nobody can split correctly later.
3. Read the result back once. Entering hello world gives .... . .-.. .-.. --- / .-- --- .-. .-.. -..; decoding that line returns the original text, which is the cheapest possible check.

For a distress call, SOS is ... --- ..., three letters chosen because the pattern is unmistakable by ear: three short, three long, three short, with no gap between the letters beyond the standard three units.

Decoding, and where it goes wrong

Decoding needs a stricter input format than people expect. Symbols inside a character must be adjacent, characters separated by single spaces, words by a slash. A line pasted with double spaces still decodes, but one pasted with no spaces at all does not, since the receiver cannot tell where one character ends.

Four failure modes account for almost every broken decode.

• Mixed alphabets in one line. A message that starts in Latin and continues in Cyrillic has to be decoded twice, changing the selector between passes.
• Missing separators. ...---... is a single unrecognised group; ... --- ... is SOS.
• Unsupported characters. A hash in the output means the character has no entry in the active table. Cyrillic ё, accented Greek vowels, and Hebrew final letter forms are the usual cases; the unaccented or standard form of the letter is normally present.
• Numerals confused with letters. In the Latin table, five dots are the numeral 5, not a letter. Decoding a numeric group under the letter table returns a plausible but wrong character.

The last one is the reason the alphabet selector exists at all. A decoder that ignores context will always produce something; the question is whether the something is what the sender meant.

Where Morse code is still in service

Commercial and military telegraphy retired the code decades ago, but three uses remain current.

Aviation navigation beacons identify themselves in Morse, so pilots hear a three-letter code repeated under the carrier. Amateur radio operators use it as the CW mode, which still reaches further per watt than voice because the signal occupies a band only 50 to 150 Hz wide. Maritime and emergency services keep it as a fallback that works when voice links fail, and the SOS pattern remains the international distress signal.

There is also a quieter use: puzzles, games, and teaching material. A short encoded line is a self-contained hint that needs no key, and the reference table doubles as an answer sheet.

Getting the spacing right when you copy by hand

Anyone writing Morse by hand runs into the same problem: element gaps and letter gaps look identical on paper. Two conventions solve it.

The first is to write the letters as separate groups and never let them touch. Three units of silence inside a letter is a fact about transmission, not about the written form; on paper the symbols inside a character are simply adjacent.

The second is to mark words explicitly. A slash, a double space, or a line break all serve, as long as the reader knows the convention in advance. The tool uses spaced slashes because they survive copy and paste into chat windows and spreadsheets, where double spaces are often collapsed.

When a message has to travel through systems that mangle whitespace, the safer form is to replace the word separator with something visible that is not itself a Morse symbol, then note the substitution in the covering message.

Ali

Washington, DC · SO reputation 19990 · Badges: 26🥇109🥈143🥉 · SO member since 2010