@k5okc Yes, you are right. A correction algorithm is overshooting badly at the moment. There should be some improvements very soon. There’s a lot of learning going on in two of our labs this week for both university students and mentors.
@k5okc <sends empathy about Python> We’re looking at Gardner, Muller/Muller, and Zero Crossing to learn. But, the current #OpulentVoice general processor implementation currently does a brute force correlation on every frame with a 16 symbol sync pulse, and then assumes 40mS later the sync pulse is findable. This works but has some obvious fragility.