Adding a "sick bunch of reverb" and "just looping it - you get the right bit and it won't sound looped" are good starting points. But once you get annoyed with the difficulties of getting good loops while retaining the human voice tonality or start to hate losing all that definition just to extend a voice sample, you'll finally get see why "granular sampling" is the key.
Granular sampling will take your voice loop, cuts it into lots of tiny pieces and pastes them beack together together, a bit like making a collage. Instead of your vocal loop fitting together like bricks in a wall (end to end to end), using a granular system you can take pieces and overlap them, smear the edges between parts, make each section of sound more human by adding in tiny amounts of pitch and timing randomness (which is essentially what adding "sick reverb" does, but without all that smearing in the time domain) until you just can't tell where the join is.
Get hold of a good sampler - not Simpler - and try doing it granular.
Bookmarks