Use the microphoneCapture a local recording, then decode when you stop.
Adaptive browser decoder
Morse Code Audio Decoder
Listen through your microphone or analyse an audio file. The decoder finds a prominent CW tone, estimates timing, and turns the signal into Morse code and readable text.
Processed locally in your browser. Recordings are not uploaded by this page.
Decode Morse audio
Automatic mode estimates frequency and speed. Switch either field to Manual when a noisy or irregular recording needs correction.
The selected alphabet changes how detected dot-and-dash groups are mapped into characters. Digits remain available across supported alphabets.
Analyse an audio fileUse a browser-supported audio file such as WAV, MP3, M4A, OGG, or AAC.
Ready. Use a clear CW/Morse tone for the strongest automatic decode.
Adaptive controls
Speed, tone, and volume window
Auto spacing estimate
dB
Discard quieter energydB
Clip louder energyTone on above threshold
Signal analysis
Spectrogram and Morse timing
Upload or record Morse audio to generate a real time-frequency spectrogram.
Range: 200 to 1800 Hz
How the adaptive decoder works
Read the signal, then expose the evidence
1. Find a prominent tone
The decoder scans the Morse/CW frequency range and selects the strongest candidate unless you lock the frequency manually.
2. Apply the volume window
Minimum and maximum dB controls discard energy outside the selected range. The threshold then separates tone-on frames from gaps.
3. Estimate timing
On/off runs are compared with likely 1-unit, 3-unit, and 7-unit Morse timings to estimate WPM, dits, dahs, character gaps, and word gaps.
4. Decode and diagnose
Detected groups are mapped through the selected alphabet. The spectrogram, timing barcode, confidence, and warnings let you inspect why a result succeeded or failed.
From sound to readable text
How to decode Morse code from audio
To decode Morse code from audio, start with the sound itself rather than manually typing dots and dashes. This Morse code audio decoder can analyse a supported audio file or a short recording captured through your microphone. It searches for likely CW tones, estimates Morse speed, separates marks from gaps, and converts the detected signal into Morse code and readable text.
Choose the sound source
Upload a recording or start the microphone. For live sound, place the microphone close enough to the speaker that the Morse signal is clearly distinguishable from room noise. A clean source is easier to analyse, but the decoder can also inspect recordings where frequency or timing is uncertain.
Run the automatic analysis
Auto mode searches likely tone frequencies and Morse speeds instead of assuming one fixed pitch or WPM value. It compares candidate results and selects the leading interpretation from the available signal and timing evidence. You can still lock frequency or speed when you already know them.
Read both outputs
Check the detected Morse sequence as well as the decoded text. Looking at both layers helps you distinguish an ambiguous character mapping from an earlier problem with frequency selection, timing, or gap detection.
Review the evidence
A readable result is not automatically a reliable result. Check the detected tone, WPM, Stability, ranked frequency candidates, character confidence, and the Accepted, Review, or Abstain decision before trusting the current text.
What audio can this Morse decoder analyse?
This Morse audio decoder accepts a microphone recording or a browser-supported audio file. Common formats such as WAV, MP3, M4A, OGG, and AAC can work when your browser can decode them. Actual format support varies by browser and operating system, so WAV is a useful fallback when a compressed format will not open.
WAVMP3M4AOGGAACMicrophone
The decoder is designed for audio that contains audible International Morse or CW-style tones. It does not require one universal pitch because automatic analysis can compare multiple likely frequency candidates. A recording with one dominant signal is generally easier to interpret than a mix of several simultaneous messages.
Why Morse audio decoding can fail
A Morse code sound decoder has to solve several problems before it can display readable text. It must decide which tone belongs to the message, identify where each mark begins and ends, estimate sending speed, distinguish dits from dahs, and determine whether a gap separates elements, letters, or words. An error at any stage can change the final decode.
- Wrong carrier: a louder whistle, harmonic, background tone, or second transmitter can pull automatic frequency selection toward the wrong signal.
- Wrong speed: an inaccurate WPM estimate can move dit, dah, letter-gap, and word-gap boundaries.
- Room echo or microphone tailing: reverberation can make marks appear longer, compress short gaps, or blur nearby boundaries.
- Noise, speech, or music: unrelated energy can fragment real marks or create false tone segments.
- Changing timing: irregular hand keying, changing WPM, clipping, or fading can make one global setting fit only part of a recording.
- Competing signals: two strong carriers can each produce readable-looking but different text.
Confidence is not a guarantee
What Accepted, Review, and Abstain mean
Audio can produce readable-looking characters even when the underlying evidence is weak. This decoder therefore separates the raw decoded text from the question of how much that result should be trusted, instead of treating every output string as equally reliable.
The leading text is supported by strong overall agreement across the tested signal, timing, and nearby-parameter evidence. It is still a machine decode, but no major conflicting candidate was detected.
The current text is the leading candidate, but some evidence is mixed. Different segmentation scales, nearby frequencies, timing settings, or uncertain characters may disagree, so inspect the conflict before relying on it.
The decoder keeps the raw text visible for inspection but does not endorse it as reliable. This can happen when nearby settings disagree strongly, signal separation is weak, too many groups are uncertain, or a strong competing carrier produces different text.
Improve the source before forcing the answer
Tips for better Morse sound decoding
A clear CW signal is easier to track than a recording with several simultaneous beeps, whistles, or transmitters. When multiple carriers are present, inspect the ranked candidates instead of assuming the first readable string must be correct.
Very loud distorted audio can flatten the signal envelope and damage the timing boundaries needed to separate marks and gaps. Reduce the source level when the audio is visibly or audibly distorted.
Background content adds changing broadband energy and can create false segments around the real Morse tone. A cleaner recording or quieter environment can improve the result.
Steady character speed and spacing make automatic WPM estimation more stable, especially in short recordings. Irregular hand keying can still be analysed, but uncertainty may increase.
For speaker-to-microphone decoding, improving the direct signal often helps more than forcing software thresholds. A nearby source can also reduce the relative effect of room echo and background noise.
If the result is marked Review or Abstain, compare ranked carriers, nearby WPM settings, Stability, and uncertain characters before manually correcting the final text.
Choose the right Morse tool
Morse audio decoder vs Morse code translator
Use the audio decoder for sound
Choose this page when your input is a recording, live microphone signal, CW tone, or Morse beeps. The processing path is sound → detected signal → timing → Morse symbols → readable text.
Use the translator for text or symbols
Choose the main translator when you already have normal text, dots and dashes, or written Morse notation. It converts directly between text and Morse symbols without recovering frequency or timing from a recording.
Open Morse Code TranslatorCommon questions
Morse code audio decoder FAQ
Can I decode Morse code from an MP3 or WAV file?
Yes. WAV and MP3 are common choices, and formats such as M4A, OGG, or AAC can also work when your browser can decode them. If a compressed file will not open, try converting it to WAV and analyse the new file.
Can I decode Morse code through my microphone?
Yes. Start Listen, allow microphone access, play or transmit the Morse sound, then stop and analyse the captured recording. For better results, keep the microphone reasonably close to the source and reduce unrelated background sound.
Why does the decoder return the wrong text?
Common causes include the wrong carrier frequency, an inaccurate WPM estimate, noise, clipping, room echo, compressed gaps, irregular hand keying, speech or music interference, or another strong signal. Use the spectrogram, Stability, ranked frequency candidates, timing evidence, and evidence decision to locate the likely conflict.
What frequency should I use for Morse audio?
There is no single pitch that fits every recording. Leave Frequency on Auto when the carrier is unknown, or lock the value manually when you know the source pitch and want the decoder to ignore a stronger unrelated tone.
What does WPM mean in Morse code audio?
WPM means words per minute and describes Morse sending speed. Leave WPM on Auto when the speed is unknown. A manual value can help when the sender speed is known and the automatic estimate is being affected by noise or irregular timing. The decoder also exposes Farnsworth WPM when character speed and wider spacing differ.
Is my recording uploaded to a server?
No. Selected audio files and microphone captures are processed locally in your browser for decoding. The decoder does not need to upload the recording to run its analysis.
What does Insufficient Evidence mean?
It means a raw candidate decode exists, but the available evidence is too weak or conflicting for the decoder to treat that string as reliable text. Nearby frequency or WPM settings may disagree, the signal may be poorly separated from noise, or another strong carrier may produce a different plausible result. Inspect the alternatives or improve the recording rather than assuming the raw text is correct.