1. International Morse reference
The core alphabet, numerals, punctuation, procedural signs, and relative timing rules follow Recommendation ITU-R M.1677-1. American Morse is a separate historical system and is not silently mixed into International Morse results.
2. Text conversion
Input is normalized to supported uppercase letters, numerals, and punctuation. Each character is matched to a fixed lookup table. A single space separates Morse characters in readable output, and a slash marks a word boundary. Unsupported characters are omitted rather than assigned an invented signal.
Decoding reverses that process: groups of dots and dashes are matched to the same table, with spaces treated as character boundaries and slashes as word boundaries. For example, HELLO becomes .... . .-.. .-.. ---.
3. Timing calculations
A dot is one timing unit and a dash is three. Silence within a character is one unit, between characters three, and between words seven. For standard PARIS-calibrated speed, the player uses approximately 1,200 ÷ WPM milliseconds for one dot unit. Frequency changes pitch, not the encoded message.
4. Audio, light, and vibration output
Browser audio schedules a tone for each dot and dash using the selected frequency and timing unit. Light playback shows the same on-and-off intervals on screen. Vibration requests the equivalent short and long durations from supported devices. Browser permissions, device vibration support, volume, screen brightness, and background throttling can affect the experience without changing the Morse pattern.
5. Image recognition status
The image decoder is a beta workflow. It currently accepts and previews browser-readable images, reports basic file progress, and provides verified sample signals. Automatic extraction of dots, dashes, and spacing is not yet complete for arbitrary uploads. Users should compare any recognized pattern with the visible source before relying on it.
6. Audio recognition status
The audio decoder is also beta. It accepts MP3 and WAV input and can request microphone access, while sample recordings demonstrate the intended tone-and-pause analysis. Background noise, clipping, speed changes, speech, music, and inconsistent keying can prevent reliable automatic decoding. An upload acknowledgement is not a decoded result.
7. Confidence and uncertainty
Confidence is shown only when the site has a known result or sample basis. A percentage is not proof that a message is correct. Zero confidence and an empty decoded result are used when an arbitrary image or recording has not been automatically interpreted.
8. Editorial review
Reference templates use the same canonical map so a correction can be applied consistently. Historical claims are checked against the source record. Material changes update the visible review date and are noted in the update history.