Have you tried not using presamp?
I've had trouble using presamp, so I quite honestly just manually tune it with TT.E.I.H while not using the presamp function.
It's difficult when you don't get the sample if you want, otherwise just time consuming.
Source:
I've done my part in Teto Talkloiding.
I've used Moresampler for this, and I will continue using Moresampler.
Just a tiny bit of testing, the closest approximation I could get within 5 minutes would be
[- ro][o k][ku]
I generally just find presamp functionality to be something I deem unnecessary.
And, yeah, it wouldn't do anything when beginning with r when I tried.
Either way, if the presamp thing doesn't work, simply do it manually.