Have you tried the other timestretching algorithms within FL? e3 Generic and Auto don't have any pitch-warping, as far as I know. You might get some artifacts, but you could try it.
Alternately, you can either put it in a sampler with some looping enabled or chop up the sample and duplicate out parts of it and crossfade it so it sounds like one long "Ooh". If looping it cuts off before the last consonants, you can just record out the sampler and put in the end of the original sample manually.