Simplify the definition of many build options.

The old formula calculates the output buffer size with "regular"
rounding (to the nearest integer), however sometimes, that is
insufficient and the last sample cannot be resampled.  This causes
audible distortions.  By changing the formula to consider the worst
case (always round up), this problem is eliminated.

Use the "_format" parameter instead of the (uninitialized) "format"
attribute.

Report unsupported format while opening the filter, not later when the
first conversion takes place.

Fixes assertion failure in destructor by not assigning
{src,dest}_format when an error occurs.

The standard has been renamed since the early draft that was
implemented in MPD.

Necessary for "final" on gcc 4.6.

Unused.  Bit reversing is done in the decoder.

Obtained from the Mercurial repository at
https://code.google.com/p/dsd2pcm/

For the range-based loop on a ConstBuffer object.

Convert to 31 bit first, then right-shift with rounding to 16 bit.

Avoid multiplication.  This is a speedup of 20%.

This is nearly 4 times faster than the "portable" algorithm.

The previous commit eliminated the need for that.

.. instead of reusing FloatToInteger<S24> and converting from S24 to
S32 in-place.

Refactor the conversion functions to classes and pass an instance to
the new function AllocateConvert().

Kludge to work around DSD playback failure because the
ReplayGainFilterPlugin is always in the filter chain.

Use the existing PcmDither library.