* Threading lock structure is very slow if it becomes starved of CPUs
* Fix "!!!" points
* LADSPA plugin has too much "artificial latency"
* LADSPA plugin probably doesn't want to go any higher than about +2 octaves
* implement+test, or remove, reset()
* sweeps & tones
* ensure default options don't produce garbage at any extreme

DONE * Rationalise naming further (e.g. use of "lock" for both peak phases
  and resynchronisation at transients; use of block vs chunk vs frame
  vs window)
DONE * Vamp plugin argument no longer properly reflect internal parameters
DONE * 192kHz
DONE * save/restore FFTW wisdom if using FFTW
DONE * Sort out resynthesis gain
DONE * LADSPA plugin could do with switches for transient/phaselock
DONE * Solve "for very long stretches" issue in calculateSizes()

DONE Rejig crispness options:

0 - no transients, no peak lock, long window
1 - no transients, no peak lock, normal window
2 - no transients, peak lock, normal window
3 - band-limited transients, peak lock, normal window
[NO] 4 - transients, no peak lock, normal window  * - is this a good option to have?
5 - transients, peak lock, normal window
6 - transients, no peak lock, short window