Fix to AC temp normalization (previously not applied under certain conditions)

[asyates]

Seems temp normalization wasn't been applied to ACs if params.whitening=='N' and params.clip_after_whiten=='Y', hence the differences observed in Issue #257 when comparing ACs with whitening=='N' and whitening=='A' (which should have produced the same waveform if whitening of ACs is skipped in both cases).

Reason is because the conditional statement

if params.clip_after_whiten:
     logger.debug("Winsorizing (clipping) data after bandpass (AC)")
     tmp = winsorizing(tmp, params, input="timeseries")

was only within the block of code following 'if params.whitening == "A"' for the auto-correlations. Have now moved this outside of this block of code so that it will always be checked for ACs regardless of params.whitening value (as with SC and CC blocks).

I guess, if you're setting params.whitening == 'N', then it would make sense to have params.clip_after_whiten == 'N' (which would give desired behaviour)... but not obvious that this would be a necessary, at least wasn't to me!