1.4.5   Update Mac OSX build to use Intel 32-bit

1.4.4   Fixes to support sound output on Vista
	Make "Edit sequence" work on WinXP
	Default to 60-minutes for WAV-output if a length can't be established
	"Write to WAV (30 min)" option for Windows
	Stop status line wrapping on Windows

1.4.3	Fixed problem when playing 7+ hour sequences with -SE or -L
	Warns properly if the WAV file limit of ~7 hours is exceeded, and truncates

1.4.2	Handles playing sequences to the clock across a daylight savings time change
	Improved Linux /dev/dsp initialisation to get a good fragment size
	Fixed OGG setup bug that affected some Linux systems when playing river sounds

1.4.1	Added wakeup option to -p drop
	Permitted duration of -p drop session to be modified
	Added ReplayGain handling for OGG files
	Applied vorbisgain to river sound files
	Added support for writing to a WAV file through Windows right-click menu
	River sounds now distributed in separate archive, except on Windows

1.4.0	Added support for randomly looping OGG files with SBAGEN_LOOPER tag
	Added two standard looping river sounds OGG files to archive
	Added <carr>/<amp> for handling plain sine tones (no beats)
	Sequence (SBG) files may now contain all options, including -i and -p
	Added -d option to select a different /dev/dsp device
	Added -c option to compensate for low-frequency rolloff of audio equipment
	Added bail-out code so that thread terminates correctly if other one dies
	Added -p drop builtin predefined sequence (thanks to Jonathan Bisson for 
	  the first version of this code)
	Added -p slide builtin predefined sequence
	Added support to look for -m files in program directory

1.2.0	Added support for Ogg and MP3 files for mix input with -m switch
	Added mix/<vol> channel type for controlling level of mix input
	Added ring buffering for -m and -M input streams to reduce chance of dropouts
	Added short usage message, with longer usage help with option -h
	Added "Press return" message for Windows users on error
	Added dithering to improve output quality
	Output rate is now automatically picked up from input WAV/Ogg/MP3 file, 
	  unless overridden by -r
	Added support for command-line options embedded in sequence files (e.g. -m)
	Display ## comments from sequence file on playback
	Support multiple sequence files on the command line, to permit separate
	  files for tone-set libraries and sequences, for example
	Support automatic building with/without MP3/OGG according to availability
	Better information in error reports
	Miscellaneous clean-ups
	Documentation overhaul
	Using .sbg extension now, to improve user experience on Windows 
	Converted most t-* files to ts-*.sbg
	Created .sbg files for tone-sets in shell-scripts
	Updated all .sbg files to include ## comments and options where necessary

1.0.14	Fixed endianness problem on output for OSX users
	Added T_MSVC macro to handle compilation using MSVC
	Added mk-msvc.txt with instructions for building with MSVC

1.0.13	Increased number of channels to 16 (was 8)
	For mix-in files ending with .wav/.WAV, correctly skip header

1.0.12	NetBSD script added
	Added mix options: -m and -M
	New p-drop sequence script
	
1.0.11	Changed Mac support to use CoreAudio because Carbon calls failed on OS 10.1

1.0.10	Added support for Mac OS X using Carbon calls.
	Added -T option to start playback at a given clock time
	Cleaned up platform-specific code handling in source

1.0.9	Added support for Windows real-time output through direct Win32 calls
	Added experimental support for user-defined brain waveforms

1.0.8	Added new `spin' effect

1.0.7	Added -F and -R options to allow faster changes within the sequence

1.0.6	Fixed -q to work correctly in conjunction with -SE
	Fixed to use binary-mode when outputting files (for DOS)

1.0.5	Added code to handle output on non-ANSI terminals when NO_ANSI_TTY is defined

1.0.4	Added -W option to write WAV file headers to files or stdout
	Added -S and -E options to run sequence as a script
	Preprocessor lines added to support compilation without direct /dev/dsp output
	-L option now outputs exactly the right length, with either raw or WAV output

1.0.3	Allowed sequence file to be read from standard input
	Changed -v option to -D, and added options -o -O -L -Q
	Tidied up time + voice output, and time scanning
	Switched to using stderr for informational output, with quiet option
	Added facility to output to a file or pipe
	Added a time limit option
	Improved synchronisation with real-time
	Added bell sound

1.0.2	Made more friendly to other compilers
	No longer requires C++ part of GCC to build

1.0.1	Added 8-bit support
	Rationalised argument processing

1.0.0	Initial release

TO-DO	Add support on Mac OS X for CoreAudio sample-rate conversion
	
TO-DO	Maybe necessary to fix DOS output to pipe to use binary mode ?
	Allow multiple simultaneous sequences to run, overlaying one another
	Tone-set option to gently oscillate the beat frequency up and down
	Option to randomize the carrier frequencies/phases slightly
	Fade between tone-sets according to joystick port input
	Open /dev/dsp twice, and set buffers not-too-long and not-too-short
	Option to output a loop-able file of exactly the right length for -i option
	Allow keypresses to cause sequence to jump forwards/backwards to marked points
