fsk: FSK text with minimodem

Encodes text, or the contents of a file, as audio FSK with minimodem, and transmits that audio with rpitx’s sendiq. On the receiving side, minimodem --rx <baud> on the receiver’s audio output decodes it. Code: fsk.go and scripts/fsk.sh.

This is a script module. It needs minimodem and sox on the Pi, and sendiq in GORPITX_PATH.

Fields

JSON	Go field	Type	Required	Rules
inputType	InputType	string	yes	"file" or "text" (gorpitx.InputTypeFile, gorpitx.InputTypeText)
text	Text	string	when inputType is text	Not empty or only spaces
file	File	string	when inputType is file	Path to an existing file
baudRate	BaudRate	*int	no	Greater than 0; gorpitx sends 50 when it is not set
frequency	Frequency	float64	yes	Hz, 5000 to 1500000000

inputType decides which of text and file is used. If you send both, the other one is ignored, not rejected. The module appends a newline to the input before sending it.

50 baud is the default because it decoded cleanest in testing. Higher rates such as 110 or 300 work too and need a stronger signal.

Example

{
  "inputType": "text",
  "text": "HELLO DE N0CALL",
  "frequency": 144390000
}

Runs:

stdbuf -oL /tmp/fsk.sh 50 144390000

with HELLO DE N0CALL and a newline on stdin, and RPITX_PATH set to GORPITX_PATH.

What the script does

1. Reads all of stdin and encodes it with minimodem --tx <baud> into /tmp/fsk_<pid>.wav.
2. Converts the WAV with sox to raw 16-bit signed samples, 48 kHz, two channels.
3. Pipes that into $RPITX_PATH/sendiq -i /dev/stdin -s 48000 -f <frequency> -t i16.
4. Deletes the WAV file when it exits.

The whole input is encoded before anything goes on the air, so a long file takes a moment before the transmission starts. If minimodem or the sox and sendiq pipeline fails, the script prints an error to stderr and exits with status 1.