Generate audio of simulated CW

Discussion in 'Amateur Radio Software' started by N4VDI, May 22, 2021.

ad: L-HROutlet
ad: l-rl
ad: MessiPaoloni-1
ad: L-MFJ
ad: HRDLLC-2
ad: abrind-2
ad: Left-3
ad: Left-2
ad: L-Geochron
  1. N4VDI

    N4VDI Ham Member QRZ Page

    Is there an open source library that can be used to generate something like a .wav file (or even a raw byte/short/int array of PCM-encoded audio) simulating the audio output of a transceiver in CW mode tuned to someone, with various emulated filters, frequency shifts, types of noise, etc?

    Ideally, it should be capable of being "keyed" like a radio (ie, instantiate something like a MorseKey object with methods like keyDown(ms) and keyUp(ms), though as long as the source is available, I could deal with ripping away a higher-level interface (say, if it only has a method like sendText(outputString)) to implement the more "raw" interface myself).

    Why: I'm taking a stab at creating a new digimode that's designed to work with "cw-only" radios, and is conceptually "FT8, but implemented via on-off keying (CW/OOK) instead of SSB". At this point, I need a way to run simulations to ensure I'm actually capable of writing a decoder that could work, and to experiment with different encoding schemes (Manchester, NEC-remote, double-pumped, etc) without polluting the HF bands with test transmissions.
  2. N0NC

    N0NC Premium Subscriber QRZ Page

    "without polluting the HF bands with test transmissions". No that will happen after your pseudo cw disguised FT-8 type digital mode is accepted by all the hams who refuse to learn morse code and will fill the bands with undecipherable cw. With all the digital modes available these days, is there really a need for another one? The intent of this idea seems be to to allow operators who can't legally use digital modes of certain bands (shall we say technicians) to work digital modes where they only have cw priviledges, rather than put in a litlle hard work and upgrade thier license. Not a new idea, one of our local no-coders came up with the same thought. He has since upgraded to extra,has become a ve, and is working psk-31 ft-8 rtty, and ssb. Also he is learning morse code at this time.

    73, have a great day
    Last edited: May 22, 2021
  3. KA0HCP

    KA0HCP XML Subscriber QRZ Page

    Read: "Please help me in my project to circumvent the letter and intent of the FCC Band plans and license privileges."

    -I'm not keen on your reasoning or goal.
  4. K7MEM

    K7MEM Ham Member QRZ Page

    Years ago, when I was working on my code speed for the 20 WPM Morse test, I used a application called Morse Academy to produce simulated QSO for copy practice. I then used WaveGen (same URL as Morse Academy) to convert the WAV files to MP3 format. That way I could use a small MP3 player for copy practice. The files I created are available on my web site at Morse Code. There is about 120 of them. This was over 20 years ago but the apps are still available. Maybe you could look into the generation of the WAV files at those URLs.

    However, I do agree that we don't need another digital mode. But you are welcome to experiment.
  5. N4VDI

    N4VDI Ham Member QRZ Page

    Actually, the mode's target audience is "General-class and above hams who could legally use SSB modes, but only have CW-capable hardware available... hardware that's probably QRP, and quite possibly chained to a single frequency".

    Or, put another way, "A sophisticated, modern, and robust OOK/CW-modulated FEC-fortified digimode compatible with ancient & toy hardware."
    Last edited: May 22, 2021
  6. AE1N

    AE1N Ham Member QRZ Page

  7. N4VDI

    N4VDI Ham Member QRZ Page

    @AE1N, wow, MorseRunner is really impressive. I'm going to have to dig through the source a bit, but the author did an incredible job with it!
  8. N6YWU

    N6YWU Ham Member QRZ Page

    One good solution would be to just use standard Morse Code, sent as synchronous CW, but substitute the SKCC or contest number with "FEC12345678" somewhere in each exchange. That way each exchange would be completely human copyable (for those who can copy 60 WPM, or whatever), but still computer correctable in low S/N, and computer parseable for the automated generation "click-the-mouse/tap-the-iPad" replies.

    But, given what Google showed at the this weeks developer conference, there's not much to prevent a computer machine learning model from carrying out an entire normal sounding conversational long form CW QSO, without the extremely narrow limitations of the FT8 exchange formatting (which seem just as limited as contest exchanges).

    On the side topic, AFAIK, there's nothing in the regulations limiting Technician licensees to using only straight keys. I would not be surprised if the vast majority of big station contest exchanges already involve computerized keyer+logging automation.

    BTW, it's really not that hard (in any language, say Python) to write an on-off-keyed sine wave with additive Gaussian white noise (recording noise off the air is even better) to an array, and write that array out to a WAV file. If you can't do that, then coding FEC and ECC algorithms is going to be a real pain.
    Last edited: May 23, 2021

Share This Page