A bug in preprocessDataSub?

[HiroMiyawaki]

In the high pass filtering loop in preprocessDataSub.m,
reading frame progresses by [NT-ops.ntbuff] sample points,
but [NT] sample points are written to temp_wh.dat,
so the filtered signal becomes longer than the original dat.

Is it intentional?
I thought it's a bug, but somehow final results seem fine.

[HiroMiyawaki]

I think it is intentional.
temp_wh.dat includes buffers used later in learnAndSolve8b.m. Right?

[marius10p]

Correct, it is intentional to reduce artifacts at transitions between batches.