Morse Code Decoder

Give the decoder the boundaries it cannot infer

Enter dots and dashes with spaces between character groups. A slash divides words, and a line break is treated as another word divider. Leading or repeated word dividers are cleaned up rather than creating empty words. The converted text updates from the normalized groups, but it still depends on the boundaries you supplied. Ordinary spaces are not a signal for a word boundary in this notation.

Compare ..., . . . and . / . / .. The first is S, the second EEE and the third E E E. An uninterrupted stream is not uniquely decodable because Morse character patterns have different lengths. Even .- can be A or E followed by T if a missing gap is allowed. A decoder that chooses a plausible word would be guessing, not recovering information preserved in the input.

When a sender has supplied separate letters but no word separators, the letters can still decode, although word divisions are lost. Ask for the original grouping before adding them. If a long clue is only visually wrapped, remove editorial line breaks deliberately while retaining its true word separators. Do not merge everything into one stream; letter boundaries are more important than making the input compact.

Normalization fixes notation, not a damaged message

The decoder accepts periods as dots and hyphens as dashes. Middle dots and bullets become periods; en dashes, em dashes and mathematical minus signs become hyphens. A vertical bar or backslash is normalized to a word separator. Repeated whitespace and separators are made consistent. These changes are useful when copying code from a document with typographic substitutions.

Normalize spacing preserves unfamiliar characters inside a group. Thus .x- remains invalid instead of quietly becoming .-. That distinction matters: deleting x would create A and hide the original problem. The same applies to commas used as tiny dots, decorative symbols or an OCR error. Accepted normalization is not an optical-character-recognition service, and the tool does not interpret every glyph that resembles a mark.

Save the original before normalizing a transcription of archival material. A normalized working copy can be easier to decode, but it no longer preserves the exact typography, line layout or separators of the source. If context matters, note which changes were mechanical and which involved an editorial decision. For a classroom clue, this is an opportunity to distinguish changing a representation from changing its information.

Understand partial success and the [?] marker

... --- ... / .x- produces SOS [?]. The recognized groups are retained, and the invalid group is visible in the output and warning. [?] is a placeholder, not the result for an encoded question mark. A real question mark is ..--.. and decodes as ?. Keep that distinction when proofreading a clue with both uncertainty and punctuation.

Unknown groups are listed distinctly in the warning, so the list is not necessarily a count of every occurrence. Use the output to locate repeated placeholders and inspect the corresponding input groups. A single invalid group prevents a valid playback timeline for the whole Morse input. Removing the group would make the remaining sequence playable, but it would also remove information; only do that if you explicitly intend an incomplete excerpt.

A valid result can still be wrong. If ... was meant to be ...., the converter returns S rather than detecting the lost dot of H. If a character space is misplaced, two valid shorter groups may replace one intended letter. Meaning and source comparison therefore remain necessary even when no warning appears. Think of validity as 'recognized by the map', not 'confirmed by the sender'.

Resolve punctuation and procedure with context

Several joined signals overlap displayed punctuation. .-.-. normally decodes as +, but the prosign preference makes it <AR>. .-... normally decodes as the additional & pattern and becomes <AS> with that preference. The option does not modify the input marks or their audio; it selects the text representation for an overlapping group.

Joined ...-.- and ...---... have no ordinary punctuation entry in this map, so they decode as <SK> and <SOS> even with preference off. Separated ... --- ... remains the three letters SOS. Likewise, .- .-. remains AR rather than becoming a joined prosign. The spaces tell the decoder that the letters were separate.

ITU-R M.1677-1 gives procedural meanings and transmission rules, not simply a list of interchangeable labels. Its wait and end-of-work signals help explain the AS and SK conventions used here. It also includes an eight-dot error signal, understood signal and starting signal that are not supported tokens in this tool. An eight-dot group will therefore be unknown here even though it has a defined role in the document. Do not call every unknown group meaningless or every accepted label a complete operational transcript.

Read symbols without adding unsupported meaning

The decoder returns uppercase letters, digits, É and mapped punctuation. It does not preserve source case or translate a message into the page's language. A Russian-language explanation does not turn a Latin Morse group into a Cyrillic character. Different alphabets or national adaptations may use some of the same patterns; this tool applies the mapping displayed in its chart.

A written word separator / is distinct from the character group -..-., which decodes as a literal slash. For a fraction or date copied in Morse, preserve that group rather than replacing it with a separator. Multiplication shares -..- with X and decodes as X. The message context can suggest an intended multiplication sign, but the converter does not recover the original typography from an identical signal.

The punctuation view marks !, &, ;, _ and $ as additional patterns, not punctuation entries in the cited ITU table. A puzzle may deliberately use them, but another tool or operational convention may differ. If an accepted group yields surprising punctuation, compare the chart and source before rewriting it. Percentages and operational number formatting are not interpreted automatically; a readable transcription is not an evaluated expression or validated coordinate.

A repair sequence for plausible but incorrect output

Start with the smallest disputed passage and its source. First check whether the marks were copied correctly. Then inspect spaces between letters and the true word separators. Next normalize accepted typography. Finally examine procedural context and the prosign option. This ordering separates an input problem from a choice of representation. Re-running the same damaged line with different speed settings cannot repair text, because WPM affects playback only.

After an agreed correction, decode the complete message again and compare both the newly repaired portion and its neighbours. A separator change may affect more than one letter. If the result is intended for print, repeat the check on the final artwork rather than only on an intermediate text file. If it came from someone else, retain an indication of uncertain material instead of silently presenting a guessed word as confirmed.

Audio playback is an optional comparison once all groups are valid. For a visual check of S versus H, the chart is usually faster. For distinguishing A and N by rhythm, try the same character speed and pitch for both. Playback proves that a recognized written sequence can be generated as tones; it does not prove your transcription matches a recording, and there is no automatic audio-input decoder to make that comparison for you.

For a self-check, Hide answer is available in Morse-to-text mode; try reading the code first, then reveal the transcript. Random replaces the current input with a built-in practice word in the selected notation, so keep your original before using it. Copy and text download keep the converted transcript, including [?] placeholders. PNG and WAV instead represent the input Morse, with WAV requiring valid groups. Keep the original as a source record beside a corrected transcript when your edits need to remain auditable.

What did historical decoding require beyond a chart?

The Library of Congress's 1844 message tape shows a recorded message from Morse to Vail in 1844. Operators could work from a physical record, while a sounder such as the Smithsonian's telegraph sounder let them interpret electrical pulses through audible clicks. Both required a human to identify patterns and boundaries. The modern typed decoder begins after that identification: it receives explicit groups, not an ambiguous real-world signal.

The distinction also explains the value of the earlier maritime examples. The 1909 Congressional Record on Jack Binns records Jack Binns's CQD distress calls in 1909; the Science Museum's account of Titanic wireless describes how Titanic's calls and messages were received by other ships in 1912. A recognized distress pattern was only one part of an exchange, which also needed identifying information and attentive operators. These histories do not show that an isolated SOS translation contains everything needed for a useful communication.

For a visual example, the National Archives' Denton interview record documents the word TORTURE blinked during an interview in 1966. Human interpretation of a timed recording is different from entering separated groups here. When teaching with that episode, distinguish the historical evidence, your transcription and the converter's output. An invented transcript or a generated audio sample must not be presented as the original recording.

Frequently asked questions