Constraining frequency floor and ceiling with ksvF0 will crash R on Windows

[rpuggaardrode]

Hi!

Since my latest update of R and wrassp I've found that my R crashes much more often than it used to. I've narrowed down at least one source of this problem to pitch and ceiling settings in ksvF0. If I bulk process files with default floor/ceiling of 50 and 600 Hz, it's generally fine. If I use settings close to those, e.g. 75 and 500 Hz, that's usually also fine. But if I set more constrained settings of say 120--250 Hz, then R will almost consistently crash at some point during the processing. This isn't actually reproducible because it usually doesn't happen with the same file; sometimes it'll happen at 13% progress, sometimes at 81%, sometimes at 56%. But the more constrained floor/ceiling, the quicker it crashes. I tried messing around with all the other parameter settings too, and they all seem fine. Any idea what's going on here?

This is wrassp 1.0.4 on R 4.3.2 from Windows 11 x64. In this particular instance I was running ksvF0 from emuR::get_trackdata, but it's also been happening when using wrassp directly so I doubt it's an emuR issue.

[rpuggaardrode]

Forgot to mention -- I wanted to see if I get the same issue with mhsF0. That doesn't seem to be the case, with mhsF0 I can use whichever floor and ceiling values I want.