1. Koch method
Start with two characters. Add a new character only after two consecutive sessions reach at least 90% accuracy.
Progressive audio learning
Recognise the rhythm of each character directly, progress without counting dots and dashes, then prepare for your first CW QSO.
Koch + Farnsworth
The Koch–Farnsworth combination avoids learning artificially slow CW and limits bad habits that are difficult to correct later.
Start with two characters. Add a new character only after two consecutive sessions reach at least 90% accuracy.
Keep character shapes at 18–20 words per minute, but increase spacing between characters and words to obtain a more comfortable effective speed.
Avoid visual trees and “dot-dash” wording. The answer should come from the sound rhythm itself.
Practise for 10 to 15 minutes once or twice a day. Stop when concentration drops instead of accumulating errors.
Do not reduce character speed below about 15 words per minute. Very slow CW encourages counting elements instead of recognising the character melody.
Progression
Each phase consolidates one skill before introducing the next.
Two characters, then progressive additions in Koch order.
Copy random groups without anticipating the next character.
Recognise digits, prefixes and callsign suffixes.
CQ, DE, RST, QTH, QSL, QRM, QRN, 73 and common abbreviations.
Copy a structured exchange with callsign, report, name and QTH.
Gradually introduce QSB, QRM, different tones and variable speeds.
Web Audio API
Set character speed, effective speed and Koch level. Listen before revealing the answer.
Validation
The tests follow the useful skills in the order in which they should be consolidated.
25 characters from the current Koch level. Target: 90%.
40 characters in groups of five. Target: 90%.
30 random digits. Target: 85%.
Eight international callsigns. Target: 85%.
Ten common Q-codes or abbreviations. Target: 90%.
Copy a structured radio message.
Copy a CW exchange with callsigns, RST, name and QTH.
Pratique
The required equipment is simple. Consistency and listening quality matter more than hardware.
A clean tone between 550 and 750 Hz is usually less tiring. Keep the volume moderate.
Choose one medium at the beginning. Write one or two characters behind to avoid anticipation.
Begin with reception. Add sending practice once character recognition is already stable.
After the basics, listen to real beacons and QSOs, first in good conditions and then with noise and fading.
En complément
Aide-mémoire
Use it after the exercise, not while listening. Direct sound recognition remains the goal.
.-
-...
-.-.
-..
.
..-.
--.
....
..
.---
-.-
.-..
--
-.
---
.--.
--.-
.-.
...
-
..-
...-
.--
-..-
-.--
--..
-----
.----
..---
...--
....-
.....
-....
--...
---..
----.
.-.-.-
--..--
..--..
.----.
-.-.--
-..-.
-.--.
-.--.-
.-...
---...
-.-.-.
-...-
.-.-.
-....-
..--.-
.-..-.
...-..-
.--.-.
.-.-.
.-...
-...-
-.--.
...-.-
...---...