The data appear to be corrupted mostly in the upper frequency channels. The problem can be demonstrated by running pretty much any dspsr command in single threaded mode and then again with 2 or more threads.
The bug is in the recently introduced FoldManager class.
In the end, the problem that this class was created to solve could also be solved in a simpler (but less efficient) way (see https://sourceforge.net/p/dspsr/bugs/90/) ... and the FoldManager class was never finished, so I've simply disabled it for now.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The bug is in the recently introduced FoldManager class.
In the end, the problem that this class was created to solve could also be solved in a simpler (but less efficient) way (see https://sourceforge.net/p/dspsr/bugs/90/) ... and the FoldManager class was never finished, so I've simply disabled it for now.