vp9000 and vsynth sample optimization

[ozy]

Hi.

I have finally managed to spend some quality time [= making my own samples] on my vp9000, but I have hit a snag.

Premise: I use the vp9000 for playing "scat" from vocal samples (mostly female, sometimes ethnic, always in solo mode, always robot switch, always playing legato). The vp9000's ram is fully expanded, out put is monoaural, I play one voice at a time.

I have been starting from very clean vocal samples (no chorus, no reverb), monoaural, *.wav files from 700k to 3 megabytes size each.

Problem is, even without transposing the samples (i.e. playing them in their natural register, on a range of 12, top 18 notes),

I can't get rid of some "flanger-like" artifacts.

"ahhh" and "ehhh" formants are harsh and, well, sound "chorused" or "flanged" (I'd say "chorused".

If everything hadn't been processed in modo, starting fron mono files, and on a mono outputs, I'd say it's a phase problem somewhere in the audio chain, but that's not the case.

Now my questioins are:

1) Any ideas?

2) is this common on the vp9000?

3) is this common on various variphrase tools?

4) anybody using the roland vsynth xt can tell me if the quality is better from this point of view?

5) any suggestion for optimizing the wav files before importing them in the vp9000?

Thanks and regards