Hi Alex,
You are totally correct. I was misunderstanding the definition of the epoch between the two streams, i.e. the fact that the python stream is epoching the raw data for averaging, then "epoching the epoch" if you will for the noise covariance matrix (which is what I thought the tmin/tmax, bmin/bmax parameters were accomplishing in the .cov description file for mne-c and they are indeed NOT). Compared to the mne-c stream, the epoching for the averaging and noise covariance estimation is entirely separate according to their individual description files and the bmin/bmax options are just to baseline correct the epoch being defined by tmin/tmax in the description file before the noise covariance is computed over the entire epoch defined by tmin/tmax in the description file.
Thanks for helping me sort that out.
Cheers,
Cody