Missing dits when the code is written to poll every 10 mS makes me think it's not a problem with the polling loop. ?It seems more likely it's something like the process being periodically suspended by the OS, thus the polling loop would not even be running for periods perhaps longer than 10 mS. ?The polling loop getting suspended might also be the reason that receive latency has been reported to increase over time. ?It might be an indication that the audio sample polling loop might be getting farther and farther behind, with each periodic task suspension by the OS.
?
It's probably time for a Linux guru to dig into the process statistics, and see how if and how long the sbitx process can be suspended, and whether the current code takes that into account, or can. ?Does anyone here know how to do that?
?
73, Ron, n6ywu
