> I'm using 22khz wavs because 16khz doesn't provide the quality i
need in
> the original wavs. I can't get 32khz wavs because that part of the
> project is finished and not redoable.
It's difficult to really respond as you've not provided any substantial
details in your questions, sample test files, and specific details of the
distortions discovered with using Speex. I'm not even clear about what
kind of audio files you're encoding. At 22KHz, this is well beyond
what's
needed for clear (human) speech, even for female speakers.
With the past fortnight, we had a visit from the Borg Hive who provided
sample test files and explicit details of the distortions encountered with
various encoding methods. It proved more useful and gave the list
something to look into.
If you provided a similarly detailed report, I imagine there will be more
of a useful response from the Speex development community.
Cheers,
=MB=