This adds a feature to auto-correct CSBK control messages when the system is using 'alternate' CRC mask values. Based on what I'm seeing in the example recording that you sent, this CapMax system is using different CRC mask values in each timeslot, and even within the same timeslot there were 6x different alternate mask values used. I guess one could ask 'How can you be sure?' that the alternate mask value is valid? Statistically, there were over 300 identical messages with identical CRC checksums for each of the 6x different alternate mask values that I saw. It seems like they're using different CRC mask values for each distinct message content. On a control channel the messages are repeated continuously, so you can statistically track and auto-correct these values. However, on lower frequency messages, this becomes problematic to collect enough examples to say definitively that the alternate mask value is correct. This all makes me scratch my head and wonder how they're actually implementing this. A puzzle to ponder for sure.
Anyway, these changes may not resolve the original issue that @CDTbolt10033 reported. I'll need some more bitstream recordings that contain voice calls so that I can further test this.
Denny