The padding must be constant and sample type aligned (e.g. some fixed multiple
of two bytes between the start of two consecutive frames for 16-bit output).
The intent is to always have the ability for stereo output with WASAPI even if
the device has some other unsupported configuration, as long as front-left and
front-right exist.
This simply omits the scale factor from the filter, similar to how up-sampling
does. The consequence of this is less smooth transitions when ramping the
pitch while down-sampling, but otherwise behaves fine.
This takes advantage of the fact than when increment <= 1 (when not down-
sampling), the scale factor is always 0. As a result, the scale and scale-phase
deltas never contribute to the filtered output. Removing those multiply+add
operations cuts half of the work done by the inner loop.
Sounds that do need to down-sample (when played with a high pitch, or is 48khz
on 44.1khz output, for example), still go through the normal bsinc process.
No other effect depends on context or listener properties, so reverb being the
only exception for speed of sound and meters per unit was putting extra work on
the effect engine for no real reason. Especially since the reverb decay time
should be the time actual time to decay irrespective of other settings.
