have you looked at the help file?
yeah, the quantiser is perhaps not as intuitive as it might be, I may change it one day...
the reason is, it treats the incoming pitch a relative to the scale/tonic/offset
i.e.
if all notes are on, then -63 in will given -63 out
BUT if only one note is on then -63 will give -51
i.e. a step value of 1 is now = 12 semitones, since only 1 note is active on the scale.
so an increment of 1.0 on the input, will increase note output depending upon the scale.
this obviously mean, like in your example, the output values become very high, very quickly, so the role of offset is to move it back into the range you want (see help file for example)
(btw: disp/note is useful for seeing what notes are being output)
I thought at the time this was reasonable i.e. the quantizer actually is acting as a scaler,
having though seen some using it, I think this is not quite what others wanted/expected.
I think what most expect is for the values to merely be quantised to the closest value.
e.g. if the scale is C F G B , and you give it E3, then if gives you F,3 if you give it F3 it gives you F3
so in the future i will add support for this, its pretty simple really... in fact simpler than its current implementation