Making short words long, if you know what I mean (and more)

[lawapa]

If you have an audio editor like SoundForge there would be several ways to pull that off. The most practical would be to loop that part you want to sustain with the rest as a release sample. But then that does require you to play back the looped sample in say something that could translate the loop for you. Most softsamplers will read a looped sample. But for the most convincing manipulation without that granulation thing you would need to isolate the ooo or aaa sound fade the beginning and end to simulate a good cross fade and have the snippets overlap themshelves. you could do this in Vegas or any other multitrack app. But it is a major time consumer Also a small program like Granulab can pull this off but takes some tweekin to get it right. Time streaching just that part of the sample you wish to sustain would introduce artifacts that stand a little to proud of the sound. Cross fading grains while not natural don't sound that bad. And last but not least. try doing it the old fashond way. Sing or speak it that way to begin with.