Melismatic playing using samples

[barrabasito]

Let's say we'd like to recreate melismatic singing, which means gliding over different notes, using chromatically tuned samples of specific steady vowels, 'ah' for example. And let's consider each sample comes from the real source (voice), so they all sound natural.
If we try to play those samples in the same way we play a monophonic analog synth that seamlessly glides between notes there seems not to be any practical solution to make it possible. Current virtual samplers can't even automate sample looping.

However there are audio morphing real time applications. Could this approach be applied for the task i posed? Maybe any other kind of spectral interpolation?