Major changes and new features from TiMidity++ v1:

* Auto voice reduction
  If there is not enough machine power to play the MIDI in real time,
  TiMidity automatically reduces the voice.

* Audio buffering
  TiMidity has internal audio buffer.  TiMidity doesn't start to play until
  this buffer is filled.

* Support Windows GUI.

* Support Sherry WRD.

* Support NAS (Network Audio System) output.

* Support EsounD output.

* Enable to play the stream MIDI in real time.  The stream format is same
  as OSS /dev/sequencer stream.  See the protocol in interface/server_c.c

* Enable to assign patch from part of soundfont.

* Lower CPU load average in trace playing.
  v1 is implemented with busy-wait-method to synchronize audio and
  interface, but v2 is not.

* Reverb and chorus effect is enabled by default.  You can disable their
  controls with command line option.

* DEFAULT_VOICES is now 64.  You can change this value to modify
  `timidity.h'.

* Modified reverb optimization implementation (the algorithm is not
  changed).

* Added sun_a.c which is modified from the original of TiMidity.  Removed
  audriv_sun.c because it is not used in this version.

* Implemented function to change the audio sample rate for 'n' interface.

* `altassign' of drumset 0 is used by default.

* Add -EF option for effects.
  -b, -n, -EC, -Ec, -ER, -Er, and -r options are now obsolete.
  Use -EFdelay=?  instead of -b?,
    -EFns=?     instead of -n?,
    -EFchorus=0 instead of -EC,
    -EFchorus=1 instead of -Ec,
    -EFreverb=0 instead of -ER,
    -EFreverb=1 instead of -Er,
    -EFreverb=2 instead of -r.

* Portamento control time is completely adjusted.

* Expanded playlist files (.m3u, .pls, and .asx).

* s/win32/w32/gi

* Full support module midi files.