Decoding Morse code from audio

Decoding Morse code from audio

Any channel plugin supporting audio can be used but most probably the SSB demodulator is going to be used for true CW (on-off keying). Decoding from audio can be performed using the MorseAngel companion software that is based on Morse code recognition from audio using a Deep Neural Network. Please check the readme of the MorseAngel project for details on its usage. Alternatively any decoding program based on audio can be used.

Connection with SDRangel is done simply through the audio device. If you do not want to use the audio mix because there is something else going on there you can create a dedicated virtual audio device and use it as the audio output device (check the plugin documentation for details). With MorseAngel it is recommended to use a 8 kHz (8000 S/s) sample rate. Details on how to create a virtual audio device and use it in another software is covered in the WSJT-X Wiki. Things work similarly for any external program piping audio from SDRangel.

It is also recommended not to use AGC with MorseAngel. AGC is convenient for voice SSB but may distort the CW envelope too much.

MorseAngel will tune on the signal with highest power peak so make sure you filter the signal of interest correctly in SDRangel. Most audio based decoding software will also require proper filtering of the signal of interest.

When a dedicated virtual audio device is used it may be handy to listen to the signal. This can be done with Pusleaudio using this command:

pacmd load-module module-loopback latency_msec=25

Then with the pavucontrol GUI you can select which interface is monitored via the loopback.