¿ªÔÆÌåÓý

Re: Fixed: cw latency and wsjtx/flidigi sidetone


 

Executive summary:?? The pin change interrupts could/should implement Dit latch and Dah latch that should be processed before reading the state of the keys with key_poll.
?
In modem.c we have cw_tx_get_sample which implements the keying code.? That is called from
modems.c by function mode_next_sample and that is called by
sbitx.c function tx_process and that by sound_process in the same file.? Sound process is called from
sbitx_sound.c? in a function that ID's as the main processing loop.? This function processes blocks of audio of 1024 in size.
?
Although the functions are called 96000 times a second they are called in blocks? separated by 10.667 ms in time.? So in reality the keyer code is "run" every 10ms.
?
I could be wrong with some of the above but it should be close to how it works.??
?
Doing some of the math of 1200/wpm to find timing in milliseconds,? I think 30 wpm would not work well but 20 wpm should be ok.???? I am not sure if one can press and release a paddle in 10ms in order to miss sending a dit.?
?
But the pin change interrupts that actually read the paddles are asynchronous? to all of the above.? And if they latch key closures a dit will not be lost if one can close and open a paddle in 10ms.
?
I don't have a zbitx or a sbitx so I can't really go any further will this.
?
?

Join [email protected] to automatically receive all group messages.