=======================================================================
                      ALSA Sequencer Interface
          Copyright (c) 2000  Takashi Iwai <tiwai@suse.de>
=======================================================================

DESCRIPTION
===========

This document describes about the Advanced Linux Sound Architecture
(ALSA) sequencer interface.  The ALSA sequencer interface communicates
between ALSA sequencer core and timidity.  The interface receives
events from sequencer and plays it in (quasi-)real-time.
On this mode, TiMidity works purely as the software real-time MIDI
render, that is as a software MIDI synth engine on ALSA.
There is no scheduling routine in this interface, since all scheduling
is done by ALSA sequencer core.

For invoking ALSA sequencer interface, run timidity as follows:
	% timidity -iA -B2,8 -Os -q0/0 -k0
The fragment size is adjustable.  The smaller number gives better
real-time response.  Then timidity shows new port numbers which were
newly created (128:0 and 128:1 below).
       ---------------------------------------
	% timidity -iA -B2,8 -Os -q0/0 -k0
	TiMidity starting in ALSA server mode
	Opening sequencer port 128:0 128:1
       ---------------------------------------
These ports can be connected with any other sequencer ports.
For example, playing a MIDI file via pmidi (what's an overkill :-),
	% pmidi -p128:0 foo.mid
If a midi file needs two ports, you may connect like this:
	% pmidi -p128:0,128:1 bar.mid
Connecting from external MIDI keyboard may become like this:
	% aconnect 64:0 128:0

INSTALLATION
============

Configure with --enable-alsaseq and --enable-audio=alsa option.
Of course, other audio devices or interfaces can be chosen
additionally, too.

For getting better real-time response, timidity must be run as root
(see below).  Set-UID root is the easiest way to achieve this.  You
may change the owner and the permission of installed timidity binary
as follows:
	# chown root /usr/local/bin/timidity
	# chmod 4755 /usr/local/bin/timidity

Be aware that this might cause a security hole!


REAL-TIME RESPONSE
==================

The interface tries to reset process scheduling to SCHED_FIFO and as
high priority as possible.  The SCHED_FIFO'd program exhibits much
better real-time response.  For example, without SCHED_FIFO, timidity
may cause significant pauses at every time /proc is accessed.
For enabling this feature, timidity must be invoked by root or
installed with set-uid root.


INSTRUMENT LOADING
==================

Timidity loads instruments dynamically at each time a program change
event is received.  This causes sometimes pauses due to buffer
underrun during playback.  Furthermore, timidity resets the loaded
instruments when the all subscriptions are disconnected.  Thus for
keeping all loaded instruments even after playback is finished, you
need to connect a dummy port (e.g. midi input port) to a timidity port
via aconnect:
	% aconnect 64:0 128:0


RESET PLAYBACK
==============

You may stop all sounds during playback by sending SIGHUP signal to
timiditiy.  The connections will be kept even after reset, but the
events will be no longer processed.  For enabling the audio again, you
have to reconnect the ports.


VISUALIZATION
=============

If you prefer a bit more fancy visual output, try my tiny program,
aseqview.
	% aseqview -p2 &
Then connect two ports to timidity ports (assumed 129:0 and 129:1 were
created by aseqview):
	% aconnect 129:0 128:0
	% aconnect 129:1 128:1
The outputs ought to be redirected to 129:0,1 instead of 128:0,1.
	% pmidi -p129:0,129:1 foo.mid


COMPATIBILITY WITH OSS
======================

You may access to timidity also via OSS MIDI emulation on ALSA
sequencer.  Take a look at /proc/asound/seq/oss for checking the
device number to be accessed.
       ---------------------------------------
	% cat /proc/asound/seq/oss
	OSS sequencer emulation version 0.1.8
	ALSA client number 63
	ALSA receiver port 0
	...
	midi 1: [TiMidity port 0] ALSA port 128:0
	  capability write / opened none

	midi 2: [TiMidity port 1] ALSA port 128:1
	  capability write / opened none
       ---------------------------------------
In the case above, the MIDI devices 1 and 2 are assigned to timidity.
Now, play with playmidi:
	% playmidi -e -D1 foo.mid


BUGS
====

Well, well, they must be there..


RESOURCES
=========

- ALSA home page
	http://www.alsa-project.org
- My ALSA hack page (including aseqview)
	http://members.tripod.de/iwai/alsa.html

====================== DYNAMIC INTERFACE ======================

 Interface modules that are specified with `-i' option can be
loadable modules by using dynamic link module.
 Place the shared library

 if_<NAME>.so

on the SHARED_LIB_PATH directory. In this case <NAME> is the interface's
short name to be identified.
 TiMidity++ requires to this shared library the following function:
 (In this case <ID> is the interface's ID that are specified at `-i' option.)

ControlMode* interface_<ID>_loader(void)
{
	ControlMode* ctl;
	/* ... */

	return ctl;
}

 If the dynamic interface that are specified at -i<ID>, TiMidity++
loads if_<NAME>.so at the SHARED_LIB_PATH (this is the macro specified
in the Makefile) and calls a function in the shared library
interface_<ID>_loader().

 If a file

 if_<NAME>.txt

describing the information about the interface by 1 line is in the
SHARED_LIB_PATH, TiMidity++ displays this information when `-h'
option is specified.


WARNING:

 You shouldn't specify the interface you want to build as dynamic link
as enabled in Makefile. TiMidity++ searches its interfaces first from
interfaces statically linked with TiMidity++, and next from dynamic linked
interfaces.

m2m: mod->midi file converter for TiMidity++




BRIEF SYNOPSIS:

This adds the new -OM output mode to TiMidity++, which will read in a mod file
and output a midi file.  All parameters needed for the conversion are
contained in a .m2m file of the same base name as the mod.  If this file can
not be found, it will generate one for you.  Chord assignment and
transposition values can be very difficult and tedious to assign by hand.  It
is STRONGLY recommended that you let the program generate the initial .m2m
file, so that it can do most (if not all) of this for you.  You will still
need to assign drums, correct banks/programs, and tweak volume amps by hand.




EXAMPLE USAGE:

timidity -c c:\timidity\timidity.cfg -OM -V 2 -idt foo.mod

This will try to read in foo.m2m as the config file for the conversion, and
will output a file called foo.mid.  -V 2 tells it to generate midi for use
on a device that uses an X^2 volume curve (all GM/GS/XG hardware).  You can
use the -o flag to specify any other output name you wish.

If you don't have timidity installed, so that you don't have a valid
timidity.cfg file, just create a 0 byte file and use it instead of the real
thing.  Since you're playing mod files, timidity doesn't need to load any
midi instruments, so you don't need to have a set of patches or a real
timidity.cfg file :)




BACKGROUND:

MOD files are a lot like MIDI files.  Both formats are basicly a series of
events that control how notes get played with which instruments.  MODs
package the instruments along with the events into a single file, while MIDI
relies on external sources of instruments.  It is this fundamental
difference that creates the most difficulty in performing a mod->midi file
conversion.  The mod file does not need to know what pitch each sample is
tuned to, if it is a drum, or if it is a chord.  MOD players simply play the
packaged sample at the requested pitch, assuming all samples are tuned to
the same fixed frequency, whether they actually are or not.  Thus, if you
were to do a direct mod->midi event conversion, you would wind up with midi
instruments playing in the wrong keys, snare drums being treated as normal
melodic instruments, and single notes where there should be chords.
Transposition, drum related channel movements, and chord emissions are the
most noticable obstacles to overcome when performing an accurace mod->midi
conversoin.

Paolo Bonzini has already done half of my job for me.  He contributed a good
amount of code that turns TiMidity++ into a first rate mod player.  This,
alone, would not have helped me very much; it was how he implemented it.
Rather than handle the events like every other mod player known to man,
TiMidity++ converts them into standard midi events, loads the mod instruments
in as special patches, and then renders them just like it would any normal
midi file.  The mod event parsing, instrument parsing, and direct event
conversion was already done!  All I had to do was handle the problems I
mentioned above, along with many more minor ones I haven't mentioned, before
writing the internal TiMidity++ events out to a midi file.  See the comments
at the top of m2m.c if you are interested in some of the other issues that
needed to be addressed during the conversion process.  Although some of
these other issues were non-trivial to deal with, and pitch bends beyond 4
octaves may still sound a bit odd, they are nothing that the average user
needs to know about or keep in mind when trying to succesfully convert a mod
file.  The only thing you need to know is that, in order to address the
conversion problems disscussed above, some information about each sample in
the mod must be specified in a config file (.m2m) associated with each mod
file.  The format of this file is given below.




M2M CONFIG FILE FORMAT:

Comment lines must begin with a #.  Blank lines (no spaces or any other
character besides a newline or carriage return) are allowed.  All other lines
must specify ALL FIVE of the fields described below.  Each field is separated
by white space.



FIELD 1: Sample Number

This is the number of the sample that you are defining information for.
The first sample in the mod file is 1 (not zero).



FIELD 2: Bank/Program, drum flag, chord, silent flag

This field specifies several different properties of the sample.  Optional
paramaters are given surrounded by parentheses.  The format for this field
is:

(!)(bank/)program(chord)(*)

If the field begins with an exclaimation mark, ! , then no notes will be
issued for this sample.  This can be used to silence samples that you can
not assign to a general midi instrument, such as speech, complicated drum
tracks, or any sound effect that you can not create a close approximation to
using GS sfx banks.

The bank portion of the field specifies an optional bank selection.  This is
the number of the bank to use, followed by a / to separate it from the
program number.

The program number is the midi instrument you are assigning to the sample.
If the sample is a drum, this is the note that the drum is mapped to in the
drum set.

The optional chord field specifies what type of chord the sample is composed
of.  There are 4 types of chords, each of which has 3 subtypes.  The
supported chord types are (M)ajor, (m)inor, (d)iminished minor, and (f)ifth.
Each chord is specified by the letter surrounded by parantheses in the
previous line.  The subtype of the chord describes how much the chord is
"rotated" from a standard chord, which can be 0, 1, or 2.  As an example of
what I mean by "rotated", a major chord is composed of the following note
semitone offsets: 0,4,7.  If you were to rotate the chord one to the left, it
would be: -5,0,4.  Two to the left is: -8,-5,0.  If no subtype is given,
zero rotation is assumed.

The final part specifies if the sample is a drum.  Put a * at the end of the
field to indicate this.  Chord assignments will be ignored if the drum flag
is set.

Examples:
8/48M     bank 8, program 48 (Orchestra Strings), with a normal major chord
!8/48M    silence this sample
8/48M2    same as the first example, only the chord is rotated down twice
48        normal Marcato Strings in tone bank 0
16/38*    Power drum set, Snare1
38*       Snare1 on the regular drum set 0



FIELD 3: Transposition

This is how much to transpose the original note specified in the mod file.
If the sample is tuned at middle C (pitch 60), it will need to be transposed
+24 semitones for the midi instrument to play on the correct pitch.  Samples
marked as drums will not be transposed, since they are fixed to a single
note on the drum channel.  You must still enter a value for the
transposition field, even if it is ignored by the drums, so that the config
file parser will not crash.



FIELD 4: Fine Tuning

All pitch bend events for this sample will be adjusted by the given fraction
of a pitch.  This is sometimes necessary for highly out of tune samples.
Some MOD composers, instead of tuning their samples correctly, use pitch
bends to tune the samples.  When you play this music with correctly tuned
samples, these pitch bends detune the note and it sounds out of tune.  So the
fine tuning value is used to compensate for these detuning pitchbends.

It is also common to find out of tune samples that were NOT tuned with
pitchbends, so adding in a pitch bend adjustment would only make them sound
worse in a midi file.  To disable fine tuning, an optional ! can be placed
before the fine tuning value.  This is the DEFAULT SETTING in the automatic
config file generator.  If you find that a mod requires fine tuning for a
sample, simply delete the ! and redo the conversion.

This feature is not yet fully implemented.  Only existing pitch bend events
are affected, so no new pitch bend events are issued.  This is not usually a
problem, however, since most cases where this feature needs to be applied
involve mods that issue pitch bends before the affected notes, since they
were intended to tune the samples to begin with.  I plan to eventually
implement insertion of new pitch bend events, so that this will be a true
fine tuning feature.



FIELD 5: %Volume

Each sample can be amplified by scaling the expression events.  100 is the
default amount, which is 100% of the original volume.  50 would decrease it
to half of the original volume, while 150 would be 1.5 times the original
volume.  Don't forget that the maximum expression value is 127, so any
expression events that get scaled higher than this will cap off at 127 and
you won't hear any difference.  It is mainly used for quieting instruments
that are too loud in the midi file, or for amplifying instruments whoose
expression values are too low to begin with.

Any fields beyond the first 5 will not be parsed.  You can type anything
here that you want.  You do not have to place a # before comment text, but
it is conventional to do so.




FREQUENCY ANALYSIS:

So, how do you figure out how much to transpose each sample and what chord
it is?  Load it up in a program that can perform an FFT on the sample and
display the frequency peaks.  The first peak is usually, but not always, the
fundamental pitch of the sample.  If the sample is a chord, take the first 3
major peaks and assign the chord from these.  Then enter the appropriate
chord and transposition values in the .m2m file and see if it sounds correct.
It is VERY time consuming to do all of this by hand....  So, I wrote routines
to do all of the assignments for you :)  It is not 100% accurate, but it's
pretty darn close.  And when it does miss a pitch or a chord, it always
assigns it the correct LOOKING answer.  That is, if I were to visually
inspect the FFT data, I would pick the same pitch the algorithm does.  I'm
no expert at this, but after spending so many hours testing this on many
different difficult to assign pitches, I think I'm pretty good at it now :)
The only way I can see to improve it is to build in some sort of
psychoacoustical model that takes into account how the human ear percieves
the sound.  And I don't think I want to do that at the moment....  It does an
above average job at dealing with samples that have more than one pitch or
chord in them, but don't be surprised if a noisy or multi-tonal sample
doesn't get assigned correctly.  Garbage in, garbage out :)  The automatic
assignment is very good for the vast majority of samples and should DEFINATELY
be tried first before you start changing things by hand.  When it does mess
up, it's usually only off by a single semitone or an octave multiple, so it's
easy to tweak from there.

Before I wrote the automatic frequency analysis routines, I knew very little
about the field.  Pitch detection is a very old problem in the audio signal
processing literature.  I looked up references in the library dating from
the 1960's.  The stuff from back then is just as relevant as the later
literature, since the methods really haven't improved much since then.  The
two major camps on how to do this are "autocorrelation" and "cepstrum"
analysis.  It turns out that autocorrelation was not the answer to my
problems.  While it works well on "well behaved" samples, it breaks down
very quickly on synth instruments, noisy instruments, and instruments with
multiple fundamental frequencies.  A large number of samples encountered in
mod files exhibit these properties.  No matter what I did to try to tweak it,
and I tried a lot of good things, I just could not make it robust enough to
handle real world samples.  It's a good theory, but it falls apart in
practice.

Cepstrum analysis proved to be much more robust.  But even so, I had to do a
good deal of pitch filtering and peak weighting before I could get it to work
well.  The 2nd FFT analysis kept giving me frequency peaks that didn't exist
in the 1st FFT spectrum.  They were, however, very close to real peaks.  So I
throw away all frequencies that fall below a pitch peak area and maximum
magnitude filter, then force the cepstrum analysis to only choose pitches that
have made it through the filter in the 1st FFT spectrum.  I set a maximum
frequency based on zero point crossing analysis, going out two zero crossings
from the largest amplitude in the sample.  This was necessary to prevent
octave jumping errors.  I found that it is also important to weight the
cepstrum peak areas by the maximum magnitude within the corresponding pitch
peak in the 1st FFT.  This was a desperate attempt to get some especially
troublesome bass samples to assign correctly.  Surprisingly enough, it works
great, giving me a higher success rate on all my samples without inducing any
new misassignments!  The only catch is that the weighting only works well for
< 2 seconds of audio analysis.  Any larger than that and the FFT size gets
so big that the pitch peaks are too diffused, so the maximum magnitudes for
the pitches are too small, and the weighting starts to give wrong answers.
If anyone wants to analyze >= 2 seconds of data, which isn't neccessary for
assigning pitches to mod/midi instruments, it would be easy to implement a
sliding window average that calls the existing frequency assignment
function.

It appears to work better than any of the other sample analysis software I
have.  If you are interested in more details of how I did the cepstrum
analysis, try looking over the code in freq.c and/or email me for a more
complete description of the algorithm I wound up with.  The new FFT routines
are not mine, but are public domain.  From all the benchmarks I could find,
this is the best FFT implementation for doing what I need to do (and for
future effects processing, should they ever be added to TiMidity++).  See
fft4g.c for info on where to get the original FFT package.




SUGGESTIONS ?:

Feel free to email me with any suggestions you may have on how I can do a
better job of converting the mods, or how I can implement things on the TODO
or WISH lists in m2m.c.  I am considering turning this into a stand alone
program, but until I get more free time and energy, it's going to stay as
just an addon for TiMidity++.




LEGAL STUFF:

TiMidity++ is distributed under the GPL, and since my code is derived from
and makes use of it, I guess it's under the GPL too.  So blah blah blah,
legal stuff, blah blah blah, etc..  You know the drill.




Eric A. Welsh <ewelsh@ccb.wustl.edu>
Center for Molecular Design
Center for Computational Biology
Washington University
St. Louis, MO

======================================================================
MIDI Tuning Standard summary
======================================================================

Besides GS/XG scale tuning which adjusts the pitch of 12 tones in an
octave individually, TiMidity++ supports MIDI Tuning Standard in
Universal SysEx.  MIDI Tuning Standard has the following advantages
compared with GS/XG scale tuning:

  - Support for microtonal sound other than 12 tones
  - The pitch can be adjusted in 1/100 cent or less accuracy
  - Temperaments can be rationally setup based on the tonality

For details, please refer to the recommended practice.

(1) Bulk Tuning Dump Request (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 00 tt F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
00		sub-ID#2 = "bulk tuning dump request (Non Real-Time)"
tt		tuning program number (0 - 127)
F7		EOX
----------------------------------------------------------------------

(2) Bulk Tuning Dump (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 01 tt <tuning name> [xx yy zz] ... chksum F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
01		sub-ID#2 = "bulk tuning dump (Non Real-Time)"
tt		tuning program number (0 - 127)
<tuning name>	16 ASCII characters
[xx yy zz]	frequency data for one note (repeated 128 times)
chksum		checksum (XOR of all bytes excluding F0, F7, and chksum)
F7		EOX
----------------------------------------------------------------------

(3) Single Note Tuning Change (Real-Time)
----------------------------------------------------------------------
F0 7F <device ID> 08 02 tt ll [kk xx yy zz] F7

F0 7F		Universal Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
02		sub-ID#2 = "single note tuning change (Real-Time)"
tt		tuning program number (0 - 127)
ll		number of changes (1 change = 1 set of [kk xx yy zz])
[kk		MIDI key number
 xx yy zz]	frequency data for that key (repeated 'll' number of times)
F7		EOX
----------------------------------------------------------------------

(4) Bulk Tuning Dump Request (Bank) (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 03 bb tt F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI tuning standard"
03		sub-ID#2 = "bulk tuning dump request (Bank) (Non Real-Time)"
bb		tuning bank number (0 - 127)
		(described as 1-128 in MIDI Tuning Specification)
tt		tuning program number (0 - 127)
F7		EOX
----------------------------------------------------------------------

(5) Key-Based Tuning Dump (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 04 bb tt <tuning name> [xx yy zz] ... chksum F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI tuning standard"
04		sub-ID#2 = "key-based tuning dump (Non Real-Time)"
bb		tuning bank number (0 - 127)
		(described as 1-128 in MIDI Tuning Specification)
tt		tuning program number (0 - 127)
<tuning name>	16 ASCII characters
[xx yy zz]	frequency data for one note (repeated 128 times)
chksum		checksum (XOR of all bytes excluding F0, F7, and chksum)
F7		EOX
----------------------------------------------------------------------

(6) Scale/Octave Tuning Dump 1-Byte Form (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 05 bb tt <tuning name> [xx] ... chksum F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI tuning standard"
05		sub-ID#2 = "scale/octave tuning dump 1-byte form
				(Non Real-Time)"
bb		tuning bank number (0 - 127)
		(described as 1-128 in MIDI Tuning Specification)
tt		tuning program number (0 - 127)
<tuning name>	16 ASCII characters
[xx]		frequency data for C,C#,... B (12 bytes total)
			00H means -64 Cent
			40H means +/- 0 Cent
			7FH means +63 Cent
chksum		checksum (XOR of all bytes excluding F0, F7, and chksum)
F7		EOX
----------------------------------------------------------------------

(7) Scale/Octave Tuning Dump 2-Byte Form (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 06 bb tt <tuning name> [xx yy] ... chksum F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI tuning standard"
06		sub-ID#2 = "scale/octave tuning dump 2-byte form
				(Non Real-Time)"
bb		tuning bank number (0 - 127)
		(described as 1-128 in MIDI Tuning Specification)
tt		tuning program number (0 - 127)
<tuning name>	16 ASCII characters
[xx yy]		frequency data for C,C#,... B (24 bytes total)
			00H 00H means -100 cents (8,192 steps of .012207 cents)
			40H 00H means 0 cents (equal temperament)
			7FH 7FH means +100 cents (8,191 steps of .012207 cents)
chksum		checksum (XOR of all bytes excluding F0, F7, and chksum)
F7		EOX
----------------------------------------------------------------------

(8) Single Note Tuning Change (Bank) (Real-Time)
----------------------------------------------------------------------
F0 7F <device ID> 08 07 bb tt ll [kk xx yy zz] ... F7

F0 7F		Universal Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI tuning standard"
07		sub-ID#2 = "single note tuning change (Bank) (Real-Time)"
bb		tuning bank number (0 - 127)
		(described as 1-128 in MIDI Tuning Specification)
tt		tuning program number (0 - 127)
ll		number of changes (1 change = 1 set of [kk xx yy zz])
[kk		MIDI key number
 xx yy zz]	frequency data for that key (repeated 'll' number of times)
F7		EOX
----------------------------------------------------------------------

(9) Single Note Tuning Change (Bank) (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 07 bb tt ll [kk xx yy zz] ... F7

F0 7E		Universal Non-Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI tuning standard"
07		sub-ID#2 = "single note tuning change (Bank) (Non Real-Time)"
bb		tuning bank number (0 - 127)
		(described as 1-128 in MIDI Tuning Specification)
tt		tuning program number (0 - 127)
ll		number of changes (1 change = 1 set of [kk xx yy zz])
[kk		MIDI key number
 xx yy zz]	frequency data for that key (repeated 'll'number of times)
F7		EOX
----------------------------------------------------------------------

(10) Scale/Octave Tuning 1-Byte Form (Real-Time)
----------------------------------------------------------------------
F0 7F <device ID> 08 08 ff gg hh [ss] ... F7

F0 7F		Universal Real-Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
08		sub-ID#2 = "scale/octave tuning 1-byte form (Real-Time)"
ff		channel/options byte 1
			bits 0 to 1 = channel 15 to 16
			bits 2 to 6 = reserved for future expansion
gg		channel byte 2 - bits 0 to 6 = channel 8 to 14
hh		channel byte 3 - bits 0 to 6 = channel 1 to 7
[ss]		12 byte tuning offset of 12 semitones from C to B
			00H means -64 cents
			40H means 0 cents (equal temperament)
			7FH means +63 cents
F7		EOX
----------------------------------------------------------------------

(11) Scale/Octave Tuning 1-Byte Form (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 08 ff gg hh [ss] ... F7

F0 7E		Universal Non Real-Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
08		sub-ID#2 = "scale/octave tuning 1-byte form (Non Real-Time)"
ff		channel/options byte 1
			bits 0 to 1 = channel 15 to 16
			bits 2 to 6 = reserved for future expansion
gg		channel byte 2 - bits 0 to 6 = channel 8 to 14
hh		channel byte 3 - bits 0 to 6 = channel 1 to 7
[ss]		12 byte tuning offset of 12 semitones from C to B
			00H means -64 cents
			40H means 0 cents (equal temperament)
			7FH means +63 cents
F7		EOX
----------------------------------------------------------------------

(12) Scale/Octave Tuning 2-Byte Form (Real-Time)
----------------------------------------------------------------------
F0 7F <device ID> 08 09 ff gg hh [ss tt] ... F7

F0 7F		Universal Real-Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
09		sub-ID#2 = "scale/octave tuning 2-byte form (Real-Time)"
ff		channel/options byte 1
			bits 0 to 1 = channel 15 to 16
			bits 2 to 6 = reserved for future expansion
gg		channel byte 2 - bits 0 to 6 = channel 8 to 14
hh		channel byte 3 - bits 0 to 6 = channel 1 to 7
[ss tt]		24 byte tuning offset of 2 bytes per semitone from C to B
			00H 00H means -100 cents (8,192 steps of .012207 cents)
			40H 00H means 0 cents (equal temperament)
			7FH 7FH means +100 cents (8,191 steps of .012207 cents)
F7		EOX
----------------------------------------------------------------------

(13) Scale/Octave Tuning 2-Byte Form (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 09 ff gg hh [ss tt] ... F7

F0 7E		Universal Non Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
09		sub-ID#2 = "scale/octave tuning 2-byte form (Non Real-Time)"
ff		channel/options byte 1
			bits 0 to 1 = channel 15 to 16
			bits 2 to 6 = reserved for future expansion
gg		channel byte 2 - bits 0 to 6 = channel 8 to 14
hh		channel byte 3 - bits 0 to 6 = channel 1 to 7
[ss tt]		24 byte tuning offset of 2 bytes per semitone from C to B
			00H 00H means -100 cents (8,192 steps of .012207 cents)
			40H 00H means 0 cents (equal temperament)
			7FH 7FH means +100 cents (8,191 steps of .012207 cents)
F7		EOX
----------------------------------------------------------------------

(14) Temperament Tonality Control Tuning (Real-Time)
----------------------------------------------------------------------
F0 7F <device ID> 08 0A sf mi F7

F0 7F		Universal Real-Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
0A		sub-ID#2 = "temperament tonality control tuning
				(Real-Time)"
sf		number of sharp/flat (1 byte)
			39H means 7 flats
			3FH means 1 flat
			40H means key of C
			41H means 1 sharp
			47H means 7 sharps
mi		major/minor (1 byte)
			00H means major key
			01H means minor key
			02H means passing major key
			03H means passing minor key
F7		EOX
----------------------------------------------------------------------

(15) Temperament Tonality Control Tuning (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 0A sf mi F7

F0 7E		Universal Non Real-Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
0A		sub-ID#2 = "temperament tonality control tuning
				(Non Real-Time)"
sf		number of sharp/flat (1 byte)
			39H means 7 flats
			3FH means 1 flat
			40H means key of C
			41H means 1 sharp
			47H means 7 sharps
mi		major/minor (1 byte)
			00H means major key
			01H means minor key
			02H means passing major key
			03H means passing minor key
F7		EOX
----------------------------------------------------------------------

(16) Temperament Type Control Tuning (Real-Time)
----------------------------------------------------------------------
F0 7F <device ID> 08 0B ff gg hh tt F7

F0 7F		Universal Real-Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
0B		sub-ID#2 = "temperament type control tuning (Real-Time)"
ff		channel/options byte 1
			bits 0 to 1 = channel 15 to 16
			bit 2 = port A/B
			bits 3 to 6 = reserved for future expansion
gg		channel byte 2 - bits 0 to 6 = channel 8 to 14
hh		channel byte 3 - bits 0 to 6 = channel 1 to 7
tt		temperament type (1 byte)
			00H means equal temperament
			01H means Pythagoras tuning
			02H means mean-tone tuning
			03H means pure intonation
			40H means user-defined temperament #0
			41H means user-defined temperament #1
			42H means user-defined temperament #2
			43H means user-defined temperament #3
F7		EOX
----------------------------------------------------------------------

(17) Temperament Type Control Tuning (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 0B ff gg hh tt F7

F0 7E		Universal Non Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
0B		sub-ID#2 = "temperament type control tuning (Non Real-Time)"
ff		channel/options byte 1
			bits 0 to 1 = channel 15 to 16
			bit 2 = port A/B
			bits 3 to 6 = reserved for future expansion
gg		channel byte 2 - bits 0 to 6 = channel 8 to 14
hh		channel byte 3 - bits 0 to 6 = channel 1 to 7
tt		temperament type (1 byte)
			00H means equal temperament
			01H means Pythagoras tuning
			02H means mean-tone tuning
			03H means pure intonation
			40H means user-defined temperament #0
			41H means user-defined temperament #1
			42H means user-defined temperament #2
			43H means user-defined temperament #3
F7		EOX
----------------------------------------------------------------------

(18) User-defined Temperament Entry (Non Real-Time)
----------------------------------------------------------------------
F0 7E <device ID> 08 0C tt <temper name>
		ll [fh fl bh bl aa bb cc dd ee ff] ... F7

F0 7E		Universal Non Real Time SysEx header
<device ID>	ID of target device (7F = all devices)
08		sub-ID#1 = "MIDI Tuning Standard"
0C		sub-ID#2 = "user-defined temperament entry (Non Real-Time)"
tt		temperament program number (0 - 63)
<temper name>	16 ASCII characters
ll		number of formula (1 formula =
				1 set of [fh fl bh bl aa bb cc dd ee ff])
[fh		applying pitch bit mask byte 1
			bits 0 to 3 = circle of fifth forward 8 to 11
			bits 4 to 5 = reserved for future expansion
			bit 6 = major flag (reversal)
 fl		applying pitch bit mask byte 2
			bits 0 to 6 = circle of fifth forward 1 to 7
 bh		applying pitch bit mask byte 3
			bits 0 to 3 = circle of fifth backward 8 to 11
			bits 4 to 5 = reserved for future expansion
			bit 6 = minor flag (reversal)
 bl		applying pitch bit mask byte 4
			bits 0 to 6 = circle of fifth backward 1 to 7
 aa bb		fraction (aa/bb)
 cc dd ee ff]	power ((cc/dd)^(ee/ff)) (repeated 'll' number of times)
F7		EOX
----------------------------------------------------------------------

======================================================================
The major/minor in the temperament tonality
======================================================================

The basic chords used in general music in C major are not only C, G, F
but also Am, Em, Dm which appear frequently.  There may also be Cm,
Gm, Fm, A, E, D, and so on.  Since these chords are not supported only
in pure intonation (C major), players need to change temperaments
according to progress of music.

To solve the issue, TiMidity++ prepares

(1) pure intonation (C major)
	based on the pitch of C in Pythagoras tuning (C major)
(2) pure intonation (A minor)
	based on the pitch of A in Pythagoras tuning (A minor)
(3) pure intonation (passing C major)
	based on the pitch of A in Pythagoras tuning (C major)
(4) pure intonation (passing A minor)
	based on the pitch of C in Pythagoras tuning (A minor)

I will explain more precisely.  The following table gives the lattice
(Cartesian model) of the scale system:

-----------------------------------------------------------------------------
D--   A--   E--   B--   F#--  C#--  G#--  D#--  A#--  E#--  B#--  F##-- C##--
Bb-   F-    C-    G-    D-    A-    E-    B-    F#-   C#-   G#-   D#-   A#-
Gb    Db    Ab    Eb    Bb    F     C     G     D     A     E     B     F#
Ebb+  Bbb+  Fb+   Cb+   Gb+   Db+   Ab+   Eb+   Bb+   F+    C+    G+    D+
Cbb++ Gbb++ Dbb++ Abb++ Ebb++ Bbb++ Fb++  Cb++  Gb++  Db++  Ab++  Eb++  Bb++
-----------------------------------------------------------------------------

The notation "ABCDEFG" is according to Pythagoras tuning.  The
notation "+", "-", "++" and "--" mean 1sc higher, 1sc lower, 2sc
higher and 2sc lower respectively.

A certain pure intonation is given as 12 sounds arranged by the
rectangle of 4x3 from the lattice.  For example, C tuning, A tuning,
A- tuning and C+ tuning are given as following tables respectively:

[C tuning (C major)]
----------------------
A-    E-    B-    F#-
F     C     G     D
Db+   Ab+   Eb+   Bb+
----------------------

[A tuning (A minor)]
----------------------
F#-   C#-   G#-   D#-
D     A     E     B
Bb+   F+    C+    G+
----------------------

[C+ tuning (passing C major)]
----------------------
A     E     B     F#
F+    C+    G+    D+
Db++  Ab++  Eb++  Bb++
----------------------

[A- tuning (passing A minor)]
----------------------
F#--  C#--  G#--  D#--
D-    A-    E-    B-
Bb    F     C     G
----------------------

I think it is nice to select the tuning combination whose pitch of
parallel key is slightly lower for major music, and slightly higher
for minor music.

======================================================================
Preset temperament of Temperament Type Control Tuning
======================================================================

First, Pythagoras tuning (major) chromatic scale is expressed by the
following recurrence relations.  Here, the index [] is a offset of the
tonic.  The operation results are surely settled between 1 and 2, so
they will be made into half or double if necessary.

[Pythagoras tuning (major) chromatic scale]
pytha_maj[ 0] = 1                       # C  1
pytha_maj[ 7] = pytha_maj[ 0] * 3/2     # G  3/2
pytha_maj[ 2] = pytha_maj[ 7] * 3/2     # D  9/8
pytha_maj[ 9] = pytha_maj[ 2] * 3/2     # A  27/16
pytha_maj[ 4] = pytha_maj[ 9] * 3/2     # E  81/64
pytha_maj[11] = pytha_maj[ 4] * 3/2     # B  243/128
pytha_maj[ 6] = pytha_maj[11] * 3/2     # F# 729/512
--
pytha_maj[ 5] = pytha_maj[ 0] * 2/3     # F  4/3
pytha_maj[10] = pytha_maj[ 5] * 2/3     # Bb 16/9
pytha_maj[ 3] = pytha_maj[10] * 2/3     # Eb 32/27
pytha_maj[ 8] = pytha_maj[ 3] * 2/3     # Ab 128/81
pytha_maj[ 1] = pytha_maj[ 8] * 2/3     # Db 256/243

On the other hand, pure intonation (major) chromatic scale can be
expressed by the following recurrence relations.  Here, sc means a
syntonic comma (81/80).

[pure intonation (major) chromatic scale]
pure_maj[ 0] = 1                        # C  1
pure_maj[ 7] = pure_maj[ 0] * 3/2       # G  3/2
pure_maj[ 2] = pure_maj[ 7] * 3/2       # D  9/8
pure_maj[ 9] = pure_maj[ 2] * 3/2 / sc  # A  5/3
pure_maj[ 4] = pure_maj[ 9] * 3/2       # E  5/4
pure_maj[11] = pure_maj[ 4] * 3/2       # B  15/8
pure_maj[ 6] = pure_maj[11] * 3/2       # F# 45/32
--
pure_maj[ 5] = pure_maj[ 0] * 2/3       # F  4/3
pure_maj[10] = pure_maj[ 5] * 2/3 * sc  # Bb 9/5
pure_maj[ 3] = pure_maj[10] * 2/3       # Eb 6/5
pure_maj[ 8] = pure_maj[ 3] * 2/3       # Ab 8/5
pure_maj[ 1] = pure_maj[ 8] * 2/3       # Db 16/15

It can be understood that pure intonation is similar to Pythagoras
tuning fundamentally except descending with 1sc at A and rising with
1sc at B flat while going up and down respectively from the tonic in
the circle of fifths.

Similarly, expressed Pythagoras tuning and pure intonation (minor)
chromatic scale by the following recurrence relations.  Although the
fractions written to right-hand side is terrible values, the
recurrence relations themselves are very simple.

[Pythagoras tuning (minor) chromatic scale]
pytha_min[ 0] = 1                       # C  1
pytha_min[ 7] = pytha_min[ 0] * 3/2     # G  3/2
pytha_min[ 2] = pytha_min[ 7] * 3/2     # D  9/8
pytha_min[ 9] = pytha_min[ 2] * 3/2     # A  27/16
pytha_min[ 4] = pytha_min[ 9] * 3/2     # E  81/64
pytha_min[11] = pytha_min[ 4] * 3/2     # B  243/128
pytha_min[ 6] = pytha_min[11] * 3/2     # F# 729/512
pytha_min[ 1] = pytha_min[ 6] * 3/2     # C# 2187/2048
pytha_min[ 8] = pytha_min[ 1] * 3/2     # G# 6561/4096
pytha_min[ 3] = pytha_min[ 8] * 3/2     # D# 19683/16384
--
pytha_min[ 5] = pytha_min[ 0] * 2/3     # F  4/3
pytha_min[10] = pytha_min[ 5] * 2/3     # Bb 16/9

[pure intonation (minor) chromatic scale]
pure_min[ 0] = 1                  * sc  # C  1     * sc
pure_min[ 7] = pure_min[ 0] * 3/2       # G  3/2   * sc
pure_min[ 2] = pure_min[ 7] * 3/2 / sc  # D  10/9  * sc
pure_min[ 9] = pure_min[ 2] * 3/2       # A  5/3   * sc
pure_min[ 4] = pure_min[ 9] * 3/2       # E  5/4   * sc
pure_min[11] = pure_min[ 4] * 3/2       # B  15/8  * sc
pure_min[ 6] = pure_min[11] * 3/2 / sc  # F# 25/18 * sc
pure_min[ 1] = pure_min[ 6] * 3/2       # C# 25/24 * sc
pure_min[ 8] = pure_min[ 1] * 3/2       # G# 25/16 * sc
pure_min[ 3] = pure_min[ 8] * 3/2       # D# 75/64 * sc
--
pure_min[ 5] = pure_min[ 0] * 2/3       # F  4/3   * sc
pure_min[10] = pure_min[ 5] * 2/3       # Bb 16/9  * sc

The differences from the major tuning are that the boundary of
Pythagoras tuning goes up three positions, that the positions of
descending with syntonic comma are changed, and that pure intonation
is adjusted 1sc higher so that melodic parts' tonic (Pythagoras
tuning) and harmonic parts' tonic (pure intonation) are overlapped.

By the way, mean-tone tuning is also prepared besides Pythagoras
tuning and pure intonation as preset temperament of TiMidity++.  While
mean-tone tuning (major) is based on the general one whose major
thirds are pure, mean-tone tuning (minor) is based on Salinas tuning
whose minor thirds are pure.  Both mean-tone tuning (major) chromatic
scale and mean-tone tuning (minor) chromatic scale can be expressed by
the following recurrence relations.

[mean-tone tuning (major) chromatic scale]
mt_maj[ 0] = 1                          # C  1
mt_maj[ 7] = mt_maj[ 0] * 5^(1/4)       # G  5^(1/4)
mt_maj[ 2] = mt_maj[ 7] * 5^(1/4)       # D  5^(1/2) / 2
mt_maj[ 9] = mt_maj[ 2] * 5^(1/4)       # A  5^(3/4) / 2
mt_maj[ 4] = mt_maj[ 9] * 5^(1/4)       # E  5/4
mt_maj[11] = mt_maj[ 4] * 5^(1/4)       # B  5^(5/4) / 4
mt_maj[ 6] = mt_maj[11] * 5^(1/4)       # F# 5^(3/2) / 8
--
mt_maj[ 5] = mt_maj[ 0] / 5^(1/4)       # F  2 / 5^(1/4)
mt_maj[10] = mt_maj[ 5] / 5^(1/4)       # Bb 4 / 5^(1/2)
mt_maj[ 3] = mt_maj[10] / 5^(1/4)       # Eb 4 / 5^(3/4)
mt_maj[ 8] = mt_maj[ 3] / 5^(1/4)       # Ab 8/5
mt_maj[ 1] = mt_maj[ 8] / 5^(1/4)       # Db 8 / 5^(5/4)

[mean-tone tuning (minor) chromatic scale]
mt_min[ 0] = 1          * sc            # C  1                 * sc
mt_min[ 7] = mt_min[ 0] * (10/3)^(1/3)  # G  (10/3)^(1/3)      * sc
mt_min[ 2] = mt_min[ 7] * (10/3)^(1/3)  # D  (10/3)^(2/3) /  2 * sc
mt_min[ 9] = mt_min[ 2] * (10/3)^(1/3)  # A  5/3               * sc
mt_min[ 4] = mt_min[ 9] * (10/3)^(1/3)  # E  (10/3)^(4/3) /  4 * sc
mt_min[11] = mt_min[ 4] * (10/3)^(1/3)  # B  (10/3)^(5/3) /  4 * sc
mt_min[ 6] = mt_min[11] * (10/3)^(1/3)  # F# 25/18             * sc
mt_min[ 1] = mt_min[ 6] * (10/3)^(1/3)  # C# (10/3)^(7/3) / 16 * sc
mt_min[ 8] = mt_min[ 1] * (10/3)^(1/3)  # G# (10/3)^(8/3) / 16 * sc
mt_min[ 3] = mt_min[ 8] * (10/3)^(1/3)  # D# 125/108           * sc
--
mt_min[ 5] = mt_min[ 0] / (10/3)^(1/3)  # F  2 / (10/3)^(1/3)  * sc
mt_min[10] = mt_min[ 5] / (10/3)^(1/3)  # Bb 4 / (10/3)^(2/3)  * sc

The point that the boundary of mean-tone tuning goes up three
positions, and that mean-tone tuning is adjusted 1sc higher, are the
same situation as Pythagoras tuning and pure intonation.

Now, I think that mean-tone tuning could use for a harmony-melody
because of the characteristic that is more harmony-like than
Pythagoras tuning, and a scale is not uneven like pure intonation.

======================================================================
User-defined temperament entry
======================================================================

The function of user-defined temperament entry (experimental) is
implemented in TiMidity++.  This corresponds to (18) of MIDI Tuning
Standard summary (see the top of this document).  For example, it can
generate various temperaments by the following SysEx's.

[equal temperament]
f0 7e 00 08 0c 00                                   ; temper prog number
65 71 75 61 6c 00 00 00 00 00 00 00 00 00 00 00     ; "equal"
01                                                  ; number of formula
0f 7f 00 00 01 01 02 01 07 0c                       ; (both) 2^(7/12)
f7

[Pythagoras tuning]
f0 7e 00 08 0c 01                                   ; temper prog number
50 79 74 68 61 67 6f 72 61 73 00 00 00 00 00 00     ; "Pythagoras"
02                                                  ; number of formula
00 3f 40 1f 03 02 01 01 00 01                       ; (maj) 3/2
43 7f 00 03 03 02 01 01 00 01                       ; (min) 3/2
f7

[mean-tone tuning]
f0 7e 00 08 0c 02                                   ; temper prog number
6d 65 61 6e 2d 74 6f 6e 65 00 00 00 00 00 00 00     ; "mean-tone"
02                                                  ; number of formula
00 3f 40 1f 01 01 05 01 01 04                       ; (maj) 5^(1/4)
43 7f 00 03 01 01 0a 03 01 03                       ; (min) (10/3)^(1/3)
f7

[pure intonation]
f0 7e 00 08 0c 03                                   ; temper prog number
70 75 72 65 20 69 6e 74 6f 6e 61 74 69 6f 6e 00     ; "pure intonation"
04                                                  ; number of formula
00 3f 40 1f 03 02 01 01 00 01                       ; (maj) 3/2
00 04 40 02 05 01 02 03 04 01                       ; (maj) 5*(2/3)^4
43 7f 00 03 03 02 01 01 00 01                       ; (min) 3/2
40 22 00 00 05 01 02 03 04 01                       ; (min) 5*(2/3)^4
f7

[Kirnberger-3]
f0 7e 00 08 0c 00                                   ; temper prog number
4b 69 72 6e 62 65 72 67 65 72 2d 33 00 00 00 00     ; "Kirnberger-3"
02                                                  ; number of formula
00 0f 00 00 01 01 05 01 01 04                       ; (both) 5^(1/4)
00 30 00 1f 03 02 01 01 00 01                       ; (both) 3/2
f7

[Hirashima temperament]
f0 7e 00 08 0c 01                                   ; temper prog number
48 69 72 61 73 68 69 6d 61 00 00 00 00 00 00 00     ; "Hirashima"
02                                                  ; number of formula
00 1f 00 03 01 01 05 01 01 04                       ; (both) 5^(1/4)
00 00 00 3c 03 02 01 01 00 01                       ; (both) 3/2
f7

[Werckmeister-3]
f0 7e 00 08 0c 02                                   ; temper prog number
57 65 72 63 6b 6d 65 69 73 74 65 72 2d 33 00 00     ; "Werckmeister-3"
02                                                  ; number of formula
00 07 00 00 01 09 02 01 0f 04                       ; (both) 2^(15/4)/9
00 18 00 3f 03 02 01 01 00 01                       ; (both) 3/2
f7

[well-temperament]
f0 7e 00 08 0c 03                                   ; temper prog number
77 65 6c 6c 2d 74 65 6d 70 65 72 00 00 00 00 00     ; "well-temper"
02                                                  ; number of formula
00 07 00 00 01 09 02 01 0f 04                       ; (both) 2^(15/4)/9
00 00 01 7f 03 02 01 01 00 01                       ; (both) 3/2
f7

----
TAMUKI Shoichi <tamuki@linet.gr.jp>

================================================================
	** Timidity SoundFont Extension **

	    written by Takashi Iwai
		<iwai@dragon.mm.t.u-tokyo.ac.jp>
		<http://bahamut.mm.t.u-tokyo.ac.jp/~iwai/>

	patch level 1: April 2, 1997
================================================================

* WHAT'S THIS?

This is an extension to use samples in SoundFont files with
timidity-0.2i.  You can employ both SoundFont file together with
ordinary GUS patch files.  Both SBK and SF2 formats are supported.


* USAGE

Two commands are newly added in configuration.

To specify the SoundFont file to be used, just add a line in config
file like:

	 soundfont sffile [order=number]

The first parameter is the file name to be loaded.  The file itself
is stored once after reading all configurations, then converted to
the internal records except wave sample data.

The optional argument specifies the order of searching.
'order=0' means to search the SoundFont file at first, then search
the GUS patches if the appropriate sample is not found.
'order=1' means to search the SoundFont file after GUS patches.

Another command 'font' is supplied to control the behavior of sample
selection.  If you don't want to use some samples in the SoundFont
file, specify the sample via 'exclude' sub-command.

	font exclude bank [preset [keynote]]

The first parameter is MIDI bank number of the sample to be removed.
The optional second parameter is MIDI program number of the sample.
For drum samples, specify 128 as bank, and drumset number as preset,
and keynote number for the drum sample.

You can change the order of individual sample (or bank) via "order"
sub-command.

	font order number bank [preset [keynote]]

The first parameter is the order number (zero or one) to be changed,
and the sequent parameters are as well as in exclude command above.


* BUGS & TODO'S

- noises on some bass drum samples
- support of modulation envelope
- support of cut off / resonance
- support of chorus / reverb


* CHANGES

- pl.1
	+ fix volume envelope calcuation
	+ add font command
	+ fix font-exclude control

---------------------------------------------------------------------
This is the README file of TkMidity Ver.1.5, Tcl/Tk interface for
TiMidity MIDI Converter/Player.

TkMidity realizes the Tk interface panel communicating with true
TiMidity program. By using Tk, you can enjoy a beautiful Motif-like
window without Motif libraries.


* WHAT'S NEW in 1.5

- Trace window using timer callback
- Forward/backward buttons
- A couple of bug fixes..


* CONTENTS

This archive contains the following files:

README.tk	- this file
tk_c.c		- tk-interface control source file
tkmidity.ptcl	- main tcl/tk source to be preprocessed
tkpanel.tcl	- main control panel tcl/tk script
browser.tcl	- file browser
misc.tcl	- miscellaneous subroutines
tkbitmaps/*.xbm	- bitmap files for TkMidity


* USAGE

There are four modes newly featured to TiMidity; repeat, shuffle,
auto-start and auto-exit modes. Repeat mode plays musics after all
files are finished repeatedly. Shuffle mode means the random pick-up
playing. When Auto-start is on, the TkMidity begins playing music as
soon as program starts. Auto-exit means to quit TkMidity automatically
after all songs are over. Each setting can be saved by Save Config
menu.

You can change the display configuration in Displays menu. This
configuration also can be saved on the start-up file by Save Config
menu.

From ver.1.3, File Open/Close menues and direct keyboard
controls are supported. You can append arbitrary files from file
browser. The shortcut key actions are as follows:

	[Enter]		: Start Playing
	[Space]		: Pause / Start Again
	[c]		: Stop Playing
	[q]		: Quit TkMidity
	[p] or [Left]	: Previous File
	[n] or [Right]	: Next File
	[v] or [Down]	: Volume Down (5%)
	[V] or [Up]	: Volume Up (5%)
	[F10]		: Menu Mode
	[Alt]+[Any]	: Select Each Menu

From this version (1.4), trace mode window is realized. You can see a
funny movements of volume and panning of each channel if you specify
the option flag in command line (see manual).


* PROGRAM NOTES

This version requires Tcl7.5/Tk4.1 libraries.  Unlike the older tk
interface, timidity links the tcl/tk libraries on its binary, not
using wish program.  Also, shared memory access must be permitted.


* TROUBLE SHOOTING

+The present script verifies the existence of the file before pass to
TiMidity, but occasionally this could happen...

		Takashi Iwai	<iwai@dragon.mm.t.u-tokyo.ac.jp>
				<http://bahamut.mm.t.u-tokyo.ac.jp/~iwai/>

How to compile TiMidity++ for Windows

This file contains fllowing instructions.

1.Mingw in Msys environment
2.Turbo C++ in Msys environment
3.OpenWatcom C++ in Msys environment
4.Visual C++ in Msys environment
5.Digital Mars in Msys environment
6.Pelles C in Msys environment

1.Mingw in Msys environment

(1)Setup Mingw and MSYS
    1)Setup Mingw(http://sourceforge.net/projects/mingw/)
      and MSYS (See MingwWiki(http://mingw.sourceforge.net/MinGWiki/))
(2)Setup needed libraries
        2-0)get pexports from Mingw
            "pexports" is included in mingw-utils-0.3.tar.gz, get it and extract it.
        2-1)pdcurses
            Get pdcurses-2.6.0-2003.07.21-1.exe from Mingw. and extract them.
            Move curses.h to include path and libcurses.a libarary path.
        2-2)oggvorbis(http://www.vorbis.com/)
            get OggVorbis-win32sdk-1.0.1.zip and extract it.
            make export libraries
                pexports xxxx.dll >xxxx.def
                dlltool --dllname xxxx.dll --input-def xxxx.def --output-lib libxxxx.a
            Change include\ogg\os_type.h line 29 as following
                (os_types.h)
                29 #  if !defined(__GNUC__) || defined(__MINGW32__)
            set environment variables in batch file like this.
                REM OggVorbis
                set PATH=\usr\local\oggvorbis-win32sdk-1.0.1\bin;\usr\local\oggvorbis-win32sdk-1.0.1\lib;%PATH%
                set C_INCLUDE_PATH=/usr/local/oggvorbis-win32sdk-1.0.1/include:%C_INCLUDE_PATH
                set LD_LIBRARY_PATH=/usr/local/oggvorbis-win32sdk-1.0.1/lib:%LD_LIBRARY_PATH%

        2-3)gogo no coder(http://www.marinecat.net/mct_top.htm)
            get gogo.dll from Gogo no Coder(http://www.marinecat.net/cgi/lcount/count.cgi?page=3112&name=wing311.exe&downex=wing311a.exe)
            get gogo.h files from Gogo noCoder source(http://www.marinecat.net/cgi/lcount/count.cgi?page=3111&name=petit311-src.lzh&downex=petit311.lzh)
            move gogo.h gogo\include\gogo
            move gogo.dll gogo\lib
            make export libraries
                pexports gogo.dll >gogo.def
                dlltool --dllname gogo.dll --input-def gogo.def --output-lib libgogo.a
            set environment variables in batch file like this.
                REM GOGO
                set PATH=\usr\local\gogo\bin;\usr\local\gogo\lib;%PATH%
                set C_INCLUDE_PATH=/usr/local/gogo/include:%C_INCLUDE_PATH%
                set LD_LIBRARY_PATH=/usr/local/gogo/lib:%LD_LIBRARY_PATH%
        2-4)flac(http://flac.sourceforge.net/)
            Get "http://downloads.sourceforge.net/flac/flac-1.2.1-devel-win.zip"  and extract it.
            Change include\*\export.h line 58 as following
              (export.h)
              58 #if defined(FLAC__NO_DLL) || !defined(_MSC_VER) \
              || !defined(__BORLANDC__) || !defined(__CYGWIN32__) || !defined(__MINGW32__)
            set environment variables in batch file like this.
              REM FLAC
              set PATH=\usr\local\flac-1.2.1-devel-win\lib;;%PATH%
              set C_INCLUDE_PATH=/usr/local/flac-1.2.1-devel-win/include:%C_INCLUDE_PATH%
              set LD_LIBRARY_PATH=/usr/local/flac-1.2.1-devel-win/bin:%LD_LIBRARY_PATH%

        2-5)portaudio(http://www.portaudio.com/)
            Download portaudio v1.19. and extarct it.
            Move portaudio.h to include path.
            Get portaudio.dll from Csound5.08.2-gnu-win32-f.exe(http://sourceforge.net/project/showfiles.php?group_id=81968).
			Rename portaudio.dll to libportaudio.dll and move it to library path.
            Only portaudo.h is needed for compiling TiMidity++.

(3)Make TiMidity++ binaries
        3-1)timw32g.exe
            (configure)
             CFLAGS="-O2" ./configure --enable-network --enable-w32gui --enable-spline=gauss \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             strip timidity.exe
             mv timidity.exe timw32g.exe
        3-2)twsyng.exe
            (configure)
             CFLAGS="-O2" ./configure --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             strip timidity.exe
             mv timidity.exe twsyng.exe

        3-3)twsynsrv.exe
            (configure)
              CFLAGS="-O2" ./configure --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
             add config.h following line
                #define TWSYNSRV 1
           (make)
             make
             strip timidity.exe
             mv timidity.exe twsynsrv.exe

        3-4)timidity.exe
            (configure)
             CFLAGS="-O2" ./configure --enable-interface=ncurses,vt100,winsyn --enable-network --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             strip timidity.exe
        3-5)timiditydrv.dll
            (configure)
             CFLAGS="-O2" ./configure --enable-winsyn --enable-windrv --enable-spline=linear \
             --enable-audio=w32,portaudio
            (make)
             make
             cd windrv
             strip timiditydrv.dll


2.Turbo C++ in Msys environment

(1)Setup BorlandC and MSYS
    1)Setup BorlandC
        Downlod Turbo C++ Explorer(http://www.codegear.com/jp/downloads/free/turbo)
        Write bcc32.cfg and ilink 32.cfg.
        (bcc32.cfg)
          -IC:\Borland\BDS\4.0\include
          -LC:\Borland\BDS\4.0\lib
          -LC:\Borland\BDS\4.0\lib\psdk
          -DWINVER=0x0400
          -D_WIN32_WINNT=0x0400
        (ilink32.cfg )
            -LC:\Borland\BDS\4.0\lib;C:\Borland\BDS\4.0\lib\psdk
        ** Caution !! **
          Folder name which include '-'s is not acceptable.
          ilink32 can't understand such folder name.
    2)Setup  MSYS (See MingwWiki(http://mingw.sourceforge.net/MinGWiki/))
        Remove mingw path from /etc/fstab.
        Add fllowing line at the top of Msys.bat and create bcc_env.bat.
              set PATH=C:\Borland\BDS\4.0\bin;%PATH%

(2)Setup needed libraries
        2-0-1)implib.exe(to make import library from dll)
           implib -a -c xxx.lib xxx.dll
        2-0-2)coff2omf.exe(VCC library to BCC library)
           coff2omf  xxxx.lib xxx_bcpp.lib.

        2-1)pdcurses
           pdcurses-2.6.0-src.tar.bz2 from Mingw(http://sourceforge.net/project/showfiles.php?group_id=2435)
           and make them.
           rename pdcurses.lib libpdcurses.lib
           Add include path and library path bcc32.cfg.
           Add library path in ilink32.cfg
        2-2)oggvorbis(http://www.vorbis.com/)
           get OggVorbis-win32sdk-1.0.1.zip and extract it.
           Add include path and library path bcc32.cfg.
           Add library path in ilink32.cfg
           Only hederfiles are need for compiling TiMidity++
        2-3)gogo no coder(http://www.marinecat.net/mct_top.htm)
           get gogo.dll from Gogo no Coder(http://www.marinecat.net/cgi/lcount/count.cgi?page=3112&name=wing311.exe&downex=wing311a.exe)
           get gogo.h files from Gogo no Coder source(http://www.marinecat.net/cgi/lcount/count.cgi?page=3111&name=petit311-src.lzh&downex=petit311.lzh)
           move gogo.h gogo\include\gogo
           (for timidity gogo.lib is not necessary)
           Add include path and library path bcc32.cfg.
           Add library path in ilink32.cfg
           Only hederfiles are need for TiMidity++
        2-5)flac(http://flac.sourceforge.net/)
           get "http://downloads.sourceforge.net/flac/flac-1.2.1-devel-win.zip" and extract it.
           Change include\*\export.h line 58 as following
                (export.h)
                58 #if defined(FLAC__NO_DLL) || !defined(_MSC_VER) \
                    || !defined(__BORLANDC__) || !defined(__CYGWIN32__) || !defined(__MINGW32__)
           Add include path and library path bcc32.cfg.
           Add library path in ilink32.cfg
        2-6)portaudio(http://www.portaudio.com/)
             I don't know how to comple portaudio with bcc commandline tools.
             Only portaudio.h is needed for compiling TiMidity++.
             You can get portaudio.dll from Csound5.08.2-gnu-win32-f.exe(http://sourceforge.net/project/showfiles.php?group_id=81968).

(3)Make TiMidity++ binaries
        3-0) perl -pe 's/CC\s-o\s\S\S*\s/CC /g' configure >configure_bc
                ( -o xxxx options are not work correctly with BCC)

        3-1)timw32g.exe
            (configure)
             CC="bcc32" CPP="cpp32" CFLAGS="" ./configure_bc  --enable-w32gui --enable-spline=gauss \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             mv timidity.exe timw32g.exe
        3-2)twsyng.exe
            (configure)
            CC="bcc32" CPP="cpp32" CFLAGS="" \
            ./configure_bc --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             mv timidity.exe twsyng.exe
        3-3)twsynsrv.exe
            (configure)
             CC="bcc32" CPP="cpp32" CFLAGS=""\
             ./configure_bc --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
             add config.h following line
                #define TWSYNSRV 1
             (make)
             make
             mv timidity.exe twsynsrv.exe
       3-4)timidity.exe
            (configure)
            CC="bcc32" CPP="cpp32" CFLAGS="" \
             ./configure_bc --enable-interface=vt100,winsyn,ncurses --enable-network --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
       3-5)timiditydrv.dll
            (configure)
            CC="bcc32" CPP="cpp32" CFLAGS="" \
             ./configure_bc  --enable-winsyn --enable-windrv --enable-spline=linear \
             --enable-audio=w32,portaudio
            (make)
             make


3.OpenWatcom C++ in Msys environment

(1)OpenWatcom and MSYS
    1)Set OpenWatcom(http://www.openwatcom.org)
       1-1)Download OpenWatcom, install and configure them
       1-2)Download Microsoft Platform SDK install and configure them.
            Replace rc.exe of OpenWatom with the one of Platform SDK's one.
            or get wrc.exe from wine-prgs-0.9.5.zip(http://www.winehq.com/), and rename rc.exe
            (OpenWatcom's rc.exe can't use).
   2)Setup Mngw and MSYS (See MingwWiki(http://mingw.sourceforge.net/MinGWiki/))
        Remove mingw path from /etc/fstab.
        Add fllowing line at the top of Msys.bat and create wcc_env.bat.
              call wcc_env.bat
        wcc_env.bat is like follow ing
            @echo off
            set LIB=
            set INCLUDE=
            call i:\watcom\setvars.bat
                 ----
               ( setteing of various env. val. s )
                  ----
(2)Setup needed libraries(The wcc386 option -5s is recomended !! Because of TiMidity++ compatibility.!!)
        2-0) make import Libray from dll
           mlib -n -b out.lib in.dll
        2-1)pdcurses
           Get  pdcurses-2.5.0  from GnuWin32(http://sourceforge.net/projects/gnuwin32/)
           and make import Libraries from dlls.
           "libpdcurses.lib"  is needed (not pdcurses.lib).
           Move curses.h to include path and libpdcurses.lib libarary path.
        2-2)oggvorbis(http://www.vorbis.com/)
           Get OggVorbis-win32sdk-1.0.1.zip and extract it.
           Edit include\ogg\os_types.h.
              (os_types.h)
              29 #  if defined(__WATCOMC__)
              30 /* MSVC/Borland */
              31 typedef __int64 ogg_int64_t;
              32 typedef int ogg_int32_t;
              33 typedef unsigned int ogg_uint32_t;
              34 typedef short ogg_int16_t;
              35 typedef unsigned short ogg_uint16_t;
              36 #  else
                   -----
              52 #  endif
           Make inport Libraries from dll
           Set environment variables in batch file like this.
                REM OggVorbis
                set PATH=\usr\local\oggvorbis-win32sdk-1.0.1_wcc\bin;\usr\local\oggvorbis-win32sdk-1.0.1_wcc\lib;%PATH%
                set C_INCLUDE_PATH=/usr/local/oggvorbis-win32sdk-1.0.1_wcc/include:%C_INCLUDE_PATH
                set LIB=\usr\local\oggvorbis-win32sdk-1.0.1_wcc\lib:%LIB%
           Only *.h files are needed for TiMidity++.
        2-3)gogo no coder(http://www.marinecat.net/mct_top.htm)
           Get gogo.dll from Gogo no Coder(http://www.marinecat.net/cgi/lcount/count.cgi?page=3112&name=wing311.exe&downex=wing311a.exe)
            Get gogo.h files from Gogo noCoder source(http://www.marinecat.net/cgi/lcount/count.cgi?page=3111&name=petit311-src.lzh&downex=petit311.lzh)
            Move gogo.h gogo\include\gogo
            (for timidity gogo.lib not necessary)
           Set environment variables in batch file like this.
                REM GOGO
                set PATH=\usr\local\gogo_wcc\bin;\usr\local\gogo_wcc\lib;%PATH%
                set C_INCLUDE_PATH=/usr/local/gogo_wcc/include:%C_INCLUDE_PATH%
                set LIB=\usr\local\gogo_wcc\lib:%LIB%
           Only *.h files are needed for compiling TiMidity++.
        2-5)flac(http://flac.sourceforge.net/)
            get "http://downloads.sourceforge.net/flac/flac-1.2.1-devel-win.zip" and extract it.
            make inport Libraries from dll
            set environment variables in batch file like this.
                REM FLAC
                set PATH=\usr\local\flac-1.2.1-win_wcc\bin;%PATH%
                set C_INCLUDE_PATH=/usr/local/flac-1.2.1-win_wcc/include:%C_INCLUDE_PATH%
                set LIB=\usr\local\flac-1.2.1-win_wcc\bin:%LIB%
           Only *.h files are needed for compiling TiMidity++.
        2-6)portaudio(http://www.portaudio.com/)
            I don't know how to comple portaudio with OpenWatcom commandline tools.
            Only port audio.h is needed for compling TiMidity++.
            You can get portaudio.dll from Csound5.08.2-gnu-win32-f.exe(http://sourceforge.net/project/showfiles.php?group_id=81968).

(3)Make TiMidity++ binaries
        3-0-1)wcc386_w.sh & wpp386_w.sh( They are in scripts/directory )
            Wcc386 is not familier to GNU autoconf tools,
            Use my wrapper wcc386_w.sh, instead of wcc386.exe
        3-1)timw32g.exe
            (configure)
            CC="wcc386_w.sh" CPP="wcc386_w.sh -p"  CFLAGS="-d0 -obll+riemcht" \
           ./configure --enable-network --enable-w32gui --enable-spline=gauss \
            --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio \
            --disable-oggtest --disable-vorbistest --disable-libFLACtest --disable-libOggFLACtest
            (make)
             make
             mv timidity.exe timw32g.exe
        3-2)twsyng.exe
            (configure)
            CC="wcc386_w.sh" CPP="wcc386_w.sh -p"  CFLAGS="-d0 -obll+riemcht" \
            ./configure --enable-network --enable-winsyng --enable-spline=linear \
            --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio \
            --disable-oggtest --disable-vorbistest --disable-libFLACtest --disable-libOggFLACtest
             (make)
             make
             mv timidity.exe twsyng.exe
        3-3)twsynsrv.exe
            (configure)
            CC="wcc386_w.sh" CPP="wcc386_w.sh -p"  CFLAGS="-d0 -obll+riemcht" \
            ./configure --enable-network --enable-winsyng --enable-spline=linear \
            --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio \
            --disable-oggtest --disable-vorbistest --disable-libFLACtest --disable-libOggFLACtest
            add config.h following line
                #define TWSYNSRV 1
             (make)
             make
             mv timidity.exe twsynsrv.exe
        3-4)timidity.exe
            (configure)
            CC="wcc386_w.sh" CPP="wcc386_w.sh -p"  CFLAGS="-d0 -obll+riemcht" \
            ./configure --enable-interface=ncurses,vt100,winsyn --enable-network --enable-spline=linear \
            --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio \
            --disable-oggtest --disable-vorbistest --disable-libFLACtest --disable-libOggFLACtest
            (make)
            make
       3-5)timiditydrv.dll
            (configure)
            CC="wcc386_w.sh" CPP="wcc386_w.sh -p"  CFLAGS="-d0 -obll+riemcht" \
             ./configure  --enable-winsyn --enable-windrv --enable-spline=linear \
             --enable-audio=w32,portaudio
            (make)
             make


4.Visual C++ in Msys environment

(1)Setup VisualC++ and MSYS
    1)Setup Visual C++
        Download
            Visual C++ 2008 Express Edition(http://www.microsoft.com/japan/msdn/vstudio/express/default.aspx)
            Microsoft Platform SDK
        install and configure them.

   2)Setup Mngw and MSYS (See MingwWiki(http://mingw.sourceforge.net/MinGWiki/))
        Remove mingw path from /etc/fstab.
        Add fllowing line at the top of Msys.bat and create bcc_env.bat.
              call vc_msys_env.bat
        vc_msys_env.bat is like follow ing
            @echo off
            call c:\"Program Files\Microsoft Platform SDK"\SetEnv.Cmd /2000 /RETAIL
            call c:"\Program Files\Microsoft Visual Studio 9.0"\Common7\Tools\vsvars32.bat

            Set INCLUDE=c:\DXSDK\include;%INCLUDE%
            Set LIB=c:\DXSDK\lib;%LIB%

                  ----
               ( setteing of various env. val. s )
                  ----
(2)Setup needed libraries
        2-1)pdcurses
           pdcurses-2.6.0-src.tar.bz2 from Mingw(http://sourceforge.net/project/showfiles.php?group_id=2435)
           and make them.
           rename pdcurses.lib libpdcurses.lib
           move curses.h to include path and libpdcurses.lib libarary path.
        2-2)oggvorbis(http://www.vorbis.com/)
           get OggVorbis-win32sdk-1.0.1.zip and extract it.
           set environment variables in batch file like this.
                REM OggVorbis
                set PATH=\usr\local\oggvorbis-win32sdk-1.0.1\bin;\usr\local\oggvorbis-win32sdk-1.0.1\lib;%PATH%
                set C_INCLUDE_PATH=/usr/local/oggvorbis-win32sdk-1.0.1/include:%C_INCLUDE_PATH
                set LIB=\usr\local\oggvorbis-win32sdk-1.0.1\lib:%LIB%

        2-3)gogo no coder(http://www.marinecat.net/mct_top.htm)
           Get gogo.dll from Gogo no Coder(http://www.marinecat.net/cgi/lcount/count.cgi?page=3112&name=wing311.exe&downex=wing311a.exe)
           Get gogo.h files from Gogo noCoder source(http://www.marinecat.net/cgi/lcount/count.cgi?page=3111&name=petit311-src.lzh&downex=petit311.lzh)
           Move gogo.h gogo\include\gogo
            (for timidity gogo.lib not necessary)
           Set environment variables in batch file like this.
                REM GOGO
                set PATH=\usr\local\gogo\bin;\usr\local\gogo\lib;%PATH%
                set C_INCLUDE_PATH=/usr/local/gogo/include:%C_INCLUDE_PATH%
                set LIB=\usr\local\gogo\lib:%LIB%
        2-5)flac(http://flac.sourceforge.net/)
            Get "http://downloads.sourceforge.net/flac/flac-1.2.1-devel-win.zip"  and extract it.
            Set environment variables in batch file like this.
                REM FLAC
                set PATH=\usr\local\flac-1.2.1-win\bin;;%PATH%
                set C_INCLUDE_PATH=/usr/local/flac-1.2.1-win/include:%C_INCLUDE_PATH%
                set LIB=\usr\local\flac-1.2.1-win\bin:%LIB%
        2-6)portaudio(http://www.portaudio.com/).
             I don't know how to comple portaudio with vc commandline tools.
            But for compling TiMidity++ only portaudio.h is needed.
            You can get portaudio.dll from Csound5.08.2-gnu-win32-f.exe(http://sourceforge.net/project/showfiles.php?group_id=81968).

(3)Make TiMidity++ binaries
        3-1)timw32g.exe
            (configure)
             CC="cl" CPP="cl.exe -EP"  CFLAGS="-O2" \
            ./configure --enable-network --enable-w32gui --enable-spline=gauss \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
            make
             mv timidity.exe timw32g.exe
        3-2)twsyng.exe
            (configure)
            CC="cl" CPP="cl.exe -EP"  CFLAGS="-O2" \
            ./configure --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             mv timidity.exe twsyng.exe
        3-3)twsynsrv.exe
            (configure)
             CC="cl" CPP="cl.exe -EP"  CFLAGS="-O2"\
             ./configure --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
             add config.h following line
                #define TWSYNSRV 1
             (make)
             make
             mv timidity.exe twsynsrv.exe
        3-4)timidity.exe
            (configure)
            CC="cl" CPP="cl.exe -EP"  CFLAGS="-O2" \
             ./configure --enable-interface=ncurses,vt100,winsyn --enable-network --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
        3-5)timiditydrv.dll
            (configure)
            CC="cl" CXX="cl" CPP="cl.exe -EP"  CFLAGS="-O2" \
              ./configure --enable-winsyn --enable-windrv --enable-spline=linear \
               --enable-audio=w32,portaudio
            (make)
             make


5.Digital Mars in Msys environment

(1) Digtal Mars & MSYS
	1) Installing up Digtal Mars(http://www.digitalmars.com/)
		1-1) Download Digtal Mars.
		2-2) replace rc.exe by that from Microsoft Platform SDK.
		�i rc.exe included  Digital Mars package is not suitable for compiling TiMidity.�j
	2)Setting up MSYS(http://sourceforge.net/projects/mingw/�@http://mingw.sourceforge.net/MinGWiki/))
		Remove mingw concening path from "/etc/fstab".
		Make dm_env.bat(like following file) is called msys.bat.
		(Add top of msys.bat�j
			call dm_env.bat
		(dm_env.bat example�j
			set LIB=
			set INCLUDE=

			Set PATH=i:\dm\bin;%PATH%
			Set INCLUDE=i:\dm\include;i:\dm\include\win32;%INCLUDE%
			Set LIB=i:\dm\lib;%LIB%

			Set PATH=i:\usr\local\gogo\bin;%PATH%
			Set INCLUDE=i:\usr\local\gogo\include;%INCLUDE%
			Set LIB=i:\usr\local\gogo\lib;%LIB%
				�i�����j
				 ----
(2) Setup Various Libraries
	2-0) How to make import libraries from dlls
		implib out.lib in.dll
	2-1) pcurses
		"Get pdcurses-2.5.0" from GnuWin32(http://sourceforge.net/projects/gnuwin32/).
		Change curses.h as following
		281 #if defined( _MSC_VER )|| defined(__DMC__)       /* defined by compiler */
		977 #if !defined(PDC_STATIC_BUILD) && (defined(_MSC_VER) || defined(__DMC__))&& defined(WIN32) && !defined(CURSES_LIBRARY)
		988 # if !defined(PDC_STATIC_BUILD) && (defined(_MSC_VER) || defined(__DMC__)) && defined(WIN32)
		Make import library with system option as follows.
			$implib /system libpdcurses.lib pdcurses.dll
		Change name of pccurses.lil to libpdcuses.lib/

    2-2) oggvorbis(http://www.vorbis.com/)
    	Use "OggVorbis-win32sdk-1.0.1.zip".
           Edit include\ogg\os_types.h as follows.
              (os_types.h)
              36 #  elif defined(__MINGW32__) || defined(__DMC__)
		Make import libraries from dlls

    2-3) Gogo no Coder (mp3 encoding dll) (http://www.marinecat.net/mct_top.htm)
           Get gogo.dll from Gogo no Coder(http://www.marinecat.net/cgi/lcount/count.cgi?page=3112&name=wing311.exe&downex=wing311a.exe)
           Get gogo.h files from Gogo noCoder source(http://www.marinecat.net/cgi/lcount/count.cgi?page=3111&name=petit311-src.lzh&downex=petit311.lzh)
		   Make gogo.lib from gogo.dll
    	   Move gogo.h gogo\include\gogo
    	   Move gogo.dll libgogo.a gogo\lib

	2-4) flac(http://flac.sourceforge.net/)
		Use "http://downloads.sourceforge.net/flac/flac-1.2.1-devel-win.zip"
		Edit line 58 of include\*\export.h s.
			(export.h)
			58 #if defined(FLAC__NO_DLL) || !defined(_MSC_VER) \
				|| !defined(__BORLANDC__) || !defined(__CYGWIN32__) || !defined(__MINGW32__) \
				|| !defined(__WATCOMC__) || !defined(__DMC__)
		Make import libraries from dlls.

	2-5) portaudio(http://www.portaudio.com/)
		Only portaudio.h is required for compiling of TiMidity++
        You can get portaudio.dll from Csound5.08.2-gnu-win32-f.exe(http://sourceforge.net/project/showfiles.php?group_id=81968).


(3) Compiling TiMIdity++
        3-0-1) "LF -> CRLF" conversion
           $sh script/unix2dos.sh
        3-0-1) perl -pe 's/CC\s-o\s\S\S*\s/CC /g' configure |perl -pe 's/CXX\s-o\s\S\S*\s/CXX /g' - >configure_dm
                (because "-o xxxx" is not suitbable dmc.exe )
        3-0-2) Get cpp.exe from Free Pascal Compiler(http://www.freepascal.org/),
             and rename it as fpcpp.exe.
            �ipreprocessor of dmc.exe is not compatible "gnu autotools"�j

        3-1)timw32g.exe
            (configure)
            CC="dmc -Jm -w -mn -5 -o" CPP="fpcpp -D__NT__ -I/i/dm/include" \
             ./configure_dm --enable-network --enable-w32gui --enable-spline=gauss \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             mv timidity.exe timw32g.exe
        3-2)twsyng.exe
            (configure)
            CC="dmc -Jm -w -mn -5 -o" CPP="fpcpp -D__NT__ -I/i/dm/include" \
             ./configure_dm --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
             (make)
             make
             mv timidity.exe twsyng.exe
        3-3)twsynsrv.exe
            (configure)
            CC="dmc -Jm -w -mn -5 -o" CPP="fpcpp -D__NT__ -I/i/dm/include" \
             ./configure_dm --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            add config.h �̍Ō�Ɉȉ��̍s��lj��B
                #define TWSYNSRV 1
             (make)
             make
             mv timidity.exe twsynsrv.exe
        3-4)timidity.exe
            (configure)
            CC="dmc -Jm -w -mn -5 -o" CPP="fpcpp -D__NT__ -I/i/dm/include" \
             ./configure_dm --enable-interface=ncurses,vt100,winsyn --enable-network --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
            make
       3-5)timiditydrv.dll
            (configure)
            CC="dmc -Jm -w -mn -5 -o" CPP="fpcpp -D__NT__ -I/i/dm/include" \
             ./configure_dm --enable-interface=windrv,winsyn --enable-network --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
            make
              commentout timiditydrv.h 23:
              23 //#error this stub requires an updated version of <rpcndr.h>
           	make



6.Pelles C in Msys environment

(1)Setting up  Pelles C and MSYS
	1) Pelles C
		Get Pelles C  from �ihttp://www.smorgasbordet.com/pellesc/).
	�Q�jMSYS(http://sourceforge.net/projects/mingw/  http://mingw.sourceforge.net/MinGWiki/))
		Remove pathes concerning with MINGW from "/etc/fstab".
		Call new file "pocc_env.bat" from msys.bat.
		(Top of msys.bat�j
			call pocc_env.bat
		("vcc_env.bat" example�j
			@echo off
			call call c:\PellesC\bin\povars32.bat

			Set INCLUDE=c:\DXSDK\include;%INCLUDE%
			Set LIB=c:\DXSDK\lib;%LIB%
				... continue ...
				 ----
(2) Setting up various libraries
	2-1) pcurses
		Get "pdcurses-2.6.0-src.tar.bz2" from "Mingw' web site", and compile it.
		Rname pccurses.lib to libpdcuses.lib.

    2-2) oggvorbis(http://www.vorbis.com/)
    	Get "OggVorbis-win32sdk-1.0.1.zip�h(http://www.vorbis.com/files/1.0.1/windows/OggVorbis-win32sdk-1.0.1.zip)

    2-3) Gogo no Coder (mp3 encoding dll) (http://www.marinecat.net/mct_top.htm)
           Get gogo.dll from Gogo no Coder(http://www.marinecat.net/cgi/lcount/count.cgi?page=3112&name=wing311.exe&downex=wing311a.exe)
           Get gogo.h files from Gogo noCoder source(http://www.marinecat.net/cgi/lcount/count.cgi?page=3111&name=petit311-src.lzh&downex=petit311.lzh)
		   Make gogo.lib from gogo.dll
    	   Move gogo.h gogo\include\gogo
    	   Move gogo.dll libgogo.a gogo\lib

	2-4) flac(http://flac.sourceforge.net/)
		Use "http://downloads.sourceforge.net/flac/flac-1.2.1-devel-win.zip".
		Edit "line 58"s of include\*\export.h as follows.
			(export.h)
			58  #if defined(FLAC__NO_DLL) || !defined(_MSC_VER) \
				|| !defined(__BORLANDC__) || !defined(__CYGWIN32__) || !defined(__MINGW32__) \
				|| !defined(__WATCOMC__) || !defined(__DMC__)

	2-5) portaudio(http://www.portaudio.com/)
		Only "portaudio.h" is requierd for compiling.
        You can get portaudio.dll from Csound5.08.2-gnu-win32-f.exe(http://sourceforge.net/project/showfiles.php?group_id=81968).

(3)Compiling TiMIdity++
		3-0) perl -pe 's/CC\s-o\s\S\S*\s/CC /g' configure >configure_pocc
                (" -o xxxx " in configure is not acceptable by pocc)
        3-1)timw32g.exe
            (configure)
            CC="cc" CPP="pocc.exe -E"  CFLAGS="-MT" ./configure_pocc  \
            --enable-network --enable-w32gui --enable-spline=gauss \
            --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
            make
             mv timidity.exe timw32g.exe
        3-2)twsyng.exe
            (configure)
            CC="cc" CPP="pocc.exe -E"  CFLAGS="-MT" ./configure_pocc  \
            --enable-network --enable-winsyng --enable-spline=linear \
            --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
             mv timidity.exe twsyng.exe
        3-3)twsynsrv.exe
            (configure)
             CC="cc" CPP="pocc.exe -E"  CFLAGS="-MT" ./configure_pocc  \
             --enable-network --enable-winsyng --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
             add config.h following line
                #define TWSYNSRV 1
             (make)
             make
             mv timidity.exe twsynsrv.exe
        3-4)timidity.exe
            (configure)
            CC="cc" CPP="pocc.exe -E"  CFLAGS="-MT" ./configure_pocc  \
            --enable-interface=ncurses,vt100,winsyn --enable-network --enable-spline=linear \
             --enable-audio=w32,vorbis,gogo,ogg,flac,portaudio
            (make)
             make
        3-5)timiditydrv.dll
            (configure)
            CC="cc" CPP="pocc.exe -E"  CFLAGS="-MT" ./configure_pocc  \
            --enable-winsyn --enable-windrv --enable-spline=linear \
              --enable-audio=w32,portaudio
            (make)
             make


2008.4.10 Keishi Suenaga(skeishi@mutt.freemail.ne.jp)

---------------------------------------------------------------------
	HOW TO USE Xaw interface of TiMidity++

Xaw interface of TiMidity prepares a GUI only with Athena Widget sets
and Xt libraries that are standard toolkits of X Window System.


* WHAT'S NEW in 2.14.0

- Support for loading and saving playlists.
- Vastly improved file dialog (Separate file/directory panes, filtering, etc.).
- Scroll wheel support.
- Supports Drag and Drop via the XDND protocol.
- Tooltips for buttons are implemented.
- Can mute individual channels in trace mode.
- Fixed time thumb. Now it can also be used to change the position in the midi.
- Support displaying more than 16 channels.
- Many many bugs fixed and other small improvments.

* INSTALLATION

The xaw interface is compiled by passing either "--enable-interface=xaw"
or "--enable-dynamic=xaw" as parameters for the configure script.
TiMidity++ ordianary install will than contain the xaw interface.

The are several ways to configure the xaw interface before compilation:

A) Some parameters can be passed to GNU configure to modify the xaw
interface. Namely:

--with-xawlib=<comma delimetered list of toolkit names>
--with-xaw-resource-prefix=<prefix to resource dir>
--enable-xdnd, --disable-xdnd
--enable-offix, --disable-offix

--with-xawlib:

TiMidity++ supports multiple Xaw toolkits. By default it searches for them
in this order: xaw3d, xaw. If you wish to force compilation
with a specific toolkit, provide the --with-xawlib=<name> parameter to
configure, e.g.:

  ./configure --enable-interface=xaw --with-xawlib=xaw

or even:

  ./configure --enable-interface=xaw --with-xawlib=neXtaw,xaw3d,XawPlus,xaw

  to force a different search order.

--enable-xdnd, --enable-offix, --disable-xdnd, --disable-offix

Enable/Disable support for XDND/Offix drag and drop protocols. By default
both are disabled.

--with-xaw-resource-prefix=<prefix>

Optional parameter. If passed, timidity will try to install the resource files
to <prefix>/app-defaults/ and <prefix>/ja_JP.EUCJP/app-defaults.

B) The xaw.h and x_trace.h files can be edited as well to change some options:

If you want to use Label Widget for lyrics as in version 1.1,
remove /* */ so that this line is visible:

  #define WIDGET_IS_LABEL_WIDGET

Although $(HOME)/.xtimidity will be regarded as an initial personal
configuration file, you can rename it by changing the line;

  #define INITIAL_CONFIG ".xtimidity"

If you'd like to see more/less than 16 channels in one screen when in
trace mode, change

  #define D_VISIBLE_CHANNELS 16

in x_trace.h.

(There's a way to change that after compilation: changing the traceHeight
resource will change the number of visible channels in one screen.)

!!! NOTE !!!
Don't forget to make clean before build if these files are edited.

* USAGE

Xaw interface shows the following file menu:

  LOAD
    Load a new MIDI file.

  SAVE
    Save the currently shown MIDI file.

  LOAD PLAYLIST
    Loads a playlist.

  SAVE PLAYLIST
    Saves a playlist.

  SAVE CONFIG
    Save the current settings and modes e.g. shuffle,
    repeat, auto-start and auto-exit flags to ~/.xtimidity .

  HIDE/UNHIDE MESSAGES
    Toggles displaying text widget showing messages and lyrics.

  HIDE/UNHIDE TRACE
    Toggles trace canvas.

  SHUFFLE
    Toggles shuffle flag.

  REPEAT
    Toggle repeat flag.

  AUTO-START
    Toggle auto-start flag.

  AUTO-EXIT
    Toggle auto-exit flag to exit or not after all songs are over.

  FILE LIST
    Popup the playing file list window.

  EXTEND MODES
    Popup the extend mode control window.

  ABOUT
    Information about Xaw interface of TiMidity++.

  QUIT
    terminate timidity.

From ver.1.3, File Load and File List menues that enabled to add and/or
delete playing list.  You can append or delete arbitrary files from file
browser.  To play a file, you can select it by either of pull-down menu
on the title bar or File List window.
The shortcut key actions on each window are as follows:

<Main Window>
	[Enter] or [r]	: Start Playing
	[Space]		: Pause / Start Again
	[a]		: Show about window
	[l]		: Show file list
	[s]		: Stop Playing
	[q]		: Quit TiMidity++
	[p] or [Left]	: Previous File
	[n] or [Right]	: Next File
	[v] or [Down]	: Volume Down (10%)
	[V] or [Up]	: Volume Up (10%)
	[<]		: Slower tempo
	[>]		: Faster tempo
	[-]		: Lower pitch
	[+]		: Raise pitch
	[Alt + F], [z]	: Show main menu
	[Alt + N]	: Load Files
	[Ctrl + V]	: Save File
	[Alt + L]	: Load Playlist
	[Alt + P]	: Save Playlist
	[Alt + S]	: Save settings to ~/.xtimidity
	[Ctrl + M]	: Hide/Show Messages
	[Ctrl + T]	: Hide/Show Trace canvas
	[Ctrl + S]	: Toggles Shuffle state
	[Ctrl + R]	: Toggles Repeat state
	[Ctrl + D]	: Toggles Tooltips
	[Ctrl + O]	: Show options window
	[g]		: Turns on/off spectrogram (if timidity was compiled
			  with --enable-spectrogram)

<File List Window>
	[p] or [Up]	: Move the cursor to the previous file
	[n] or [Down]	: Move the cursor to the next file
	[Right]		: Move the cursor 5 lines backward
	[Left]		: Move the cursor 5 lines forward
	[v]		: Volume Down (10%)
	[V]		: Volume Up (10%)
	[Enter] or [R]	: Start Playing the selected file
	[d]		: Delete the selected file entry
	[A]		: Delete all entries
	[Space]		: Pause / Start Again
	[s]		: Stop Playing
	[q]		: Quit TiMidity++
	[c]		: Close the window
	[ESC]		: Close the window

<Extend Mode Window>
	[c]		: Close the window
	[ESC]		: Close the window
	[q]		: Quit TiMidity++

<Dialog in Load File Window>
	[Tab]			: File or directory completion
	[Enter]			: Select the directory entries
	[Alt-Enter]		: Selects all visible files
	[Escape] or [CTRL]+[g]	: Cancel and close the window
	[Ctrl-a]		: Move the cursor to the top
	[Ctrl-e]		: Move the cursor to the line-end
	[Ctrl-k]		: Delete strings after the cursor
	[Ctrl-`]		: Turn on filter
By entering a string containing a wildcard in the dialog, the filter will
be turned on, only displaying the matching files (case-insensitive). e.g.
*.mid shows all filenames ending in '.mid', '.MID', etc.
[AB]* shows all filenames beginning with 'a','A','b' or 'B'.
[!c]* shows all filenames not starting with 'c'.

* TRACE MODE
Specify the interface option flag 't' in command line like as
'timidity -iat' when you execute TiMidity (see timidity man page).
You can see a funny movements of volume, expression, panning,
pitchbending, reverb, chorus and voices et cetera of each channel on the
trace canvas.  Click left mouse button on the trace screen and you can
toggle trace screen between 'pitchbend and instrument' or 'tonebank,
reverb and chorus'. Click right mouse button and you can (un)mute a channel.
Click middle mouse button, and you can play a channel solo.
Wheel or 'j' key can scroll the trace to see more channels than the usual
16 channels. If the trace is too tall for your resolution, simply set
the traceHeight resource to a lower value.

* RESOURCES

Application class name is 'TiMidity', so specity resources in
.Xdefaults or .Xresources in your home directory as follows for example;

  TiMidity*textbackground: LemonChiffon
  TiMidity*background: PeachPuff1
  TiMidity*menubutton: goldenrod
  TiMidity*buttonbackground: gold
  TiMidity*labelfont: -adobe-helvetica-bold-o-*-*-14-*-75-75-*-*-*-*

Application specific resources supported by TiMidity are listed below:

Name                Class                   Default
----------------------------------------------------------
arrangeTitle        ArrangeTitle            False
  Show MIDI titles on the title bar of the main window.
background          Background              gray65
  Background color of the main window.
menubutton          MenuButtonBackground    #CCFF33
  Menu button color.
textbackground      TextBackground          gray85
  Background color of Text Widget.
toggleforeground    ToggleForeground        MediumBlue
  Foreground color of toggle buttons.
buttonforeground    ButtonForeground        blue
  Foreground color of command buttons and other transient windows.
buttonbackground    ButtonBackground        gray76
  Background color of command buttons and other transient windows.
text2background     Text2Background         gray80
  Background color of labels in file dialogs.

labelfont           FontSet      -adobe-helvetica-bold-r-*-*-14-*-75-75-*-*-*-*
  Label widget font.
volumefont          FontSet      -adobe-helvetica-bold-r-*-*-12-*-75-75-*-*-*-*
  Volume Label font.
textfontset         FontSet      -*-*-medium-r-normal--14-*-*-*-*-*-*-*
  Text font for message box and so on.
ttitlefont          FontSet      -*-fixed-medium-r-normal--14-*-*-*-*-*-*-*
  MIDI title font at the bottom of trace window.
textHeight          TextHeight   120
  Height of the lyric widget (is 30 if WIDGET_IS_LABEL_WIDGET is defined).
menuWidth           MenuWidth    200
  Width of popup menu.
moreString          String       More...
  Label of submenus on the pop-up menu.
fileLabel           String       file...
  Title of main menu.
noPlaying           String       [ No Playing File ]
  String appended to title when no file is being played.
popup_confirm_title String       Dialog
  Title of confirm boxes.

These resources are used only in trace mode:

Name                Class                   Default
----------------------------------------------------------
tracefont           FontSet      -*-*-medium-r-normal--14-*-*-*-*-*-*-*
  Trace window font.
traceWidth          TraceWidth     627
  Width of trace widget.
traceHeight         TraceHeight    332
  Height of trace widget.
tracebackground     TraceBackground         gray90
  Background color of trace canvas.
velforeground       VelForeground           orange
  Color of bar in 'vel' (velocity) column for 'normal' channels.
veldrumforeground   VelDrumForeground       red
  Color of bar in 'vel' column, when the channel is a drum channel.
volforeground       VolForeground           LightPink
  Color of bar in 'vol' column.
expforeground       ExpForeground           aquamarine
  Color of bar in 'expr' column.
panforeground       PanForeground           blue
  Color of triangle in pan column.
rimcolor            RimColor                gray20
  Color of the border between channel bars.
boxcolor            BoxColor                gray76
  Color of the channel bars.
captioncolor        CaptionColor            DarkSlateGrey
  Color of text in the trace widget.
whitekeycolor       WhiteKeyColor           white
  Color of white keys in keyboard.
blackkeycolor       BlackKeyColor           black
  Color of black keys in keyboard.
playingkeycolor     PlayingKeyColor         maroon1
  Color of played key in keyboard.
sustainedkeycolor   SustainedKeyColor       red4
  Color of sustained key in keyboard.
reverbcolor         ReverbColor             PaleGoldenrod
  Color of bar in reverb column.
choruscolor         ChorusColor             yellow
  Color of bar in chorus column.
untitled            String       <No title>
  String used in foot when no file is being played.

Note that all font resource are of type FontSet.
Also traceWidth setting does not work yet.

!!! NOTE !!!
Some resource files were attached to TiMidity++1.3.6 or later.
If you create a resource file of another locale except English or
Japanese, please mail it to the timidity mailing list
<timidity-talk@lists.sourceforge.net>.
After building TiMidity++, put TiMidity.ad into the resources directory
such as;
 $(X11R6)/lib/X11/app-defaults
And put TiMidity-uj.ad, which is the Japanese-EUC resource file into;
 $(X11R6)/lib/X11/$(LANG)/app-defaults
For example, those are as follows if the Japanese-EUC locale of your
system is ja_JP.EUC as FreeBSD;
 # cp TiMidity.ad /usr/X11R6/lib/X11/app-defaults/TiMidity
 # cp TiMidity-uj.ad /usr/X11R6/lib/X11/app-defaults/ja_JP.EUC/TiMidity

* SETTINGS FILE
 ~/.xtimidity contains settings which are used by the xaw interface.
 When "SAVE CONFIG" is selected, timidity will save the current configuration
 to the file.

 The file has a simple format. Each line begins with "set", than an option name,
 followed by the option's value.
 The options are:

Name                Allowed Values          Default
----------------------------------------------------------
RepeatPlay           0/1                    0
  Repeat playing the playlist after the last file was played.
ShufflePlay          0/1                    0
  Play the playlist in a random order.
ExtOptions           0-63                   Depends on compliation options.
  Enable various sequencer extensional modes (see options window for list).
ChorusOption         0/1                    Depends on compliation options.
  Enable chorus.
CurVol               0-800                  70
  Initial volume.
Showdotfiles         0/1                    0
  Show files and directories beginning with '.' in the file dialog.
DefaultDir           a valid dir name.      The user's home directory.
  The directory first displayed in the file dialog.
Disp:trace           0/1                    0
  Always start in trace mode.
Disp:text            0/1                    1
  Display the lyrics/messages widget at startup.
Tooltips             0/1                    1 [1]
  Display tooltips for buttons.
Autostart            0/1                    1
  Start playing immediately when the interface is started.
AutoExit             0/1                    0
  Exit after playing has stopped.
ConfirmExit          0/1                    0
  Confirm exit from interface if user tries to close it.
SaveList             0/1                    1
  Save playlist to ~/.xtimidity when saving configuration. If Autostart is 1,
  it will begin playing immediately when timidity is next started.
SaveConfig           0/1                    1
  Save config at exit.
File                 a valid file name.     None
  This file will be loaded at next start. If SaveList is 1, the previous
  playlist will be saved using this option. This option may occur multiple
  times in ~/.xtimidity.

[1] When timidity is compiled with xaw3d, tooltips are by default off.
    Tooltips can not be enabled when the neXtaw toolkit is used.

* REPORTS

Please send problems concerning Xaw interface, feature requests,
bug reports or comments to the TiMidity mailing list
<timidity-talk@lists.sourceforge.net>.

  Regards,

	Yoshishige Arai	<ryo2@on.rim.or.jp>

------------------------------------------------------------------
This document describes about TiMidity++ skin interface.

 Skin interface is the front-end of TiMidity++ to allow users to
control TiMidity++ with funny GUI. The face of GUI is changable
by using "skin data" of WinAmp / x11amp.


* SETTING UP THE X-SKIN INTERFACE *

 First, choose a skin data to use and place it on some directory.
And set this path to the environment variable "timidity_skin"
 e.g, If your favourite skin data (*.bmp) is placed on
 ~/.x11amp/Skins/timidity/ :

% setenv timidity_skin ~/.x11amp/Skins/timidity/                 (csh)
$ timidity_skin=~/.x11amp/Skins/timidity ; export timidity_skin  (sh)

 If your skin data archive file is on /dos/programs/winamp/skins/winamp.zip:

% setenv timidity_skin /dos/programs/winamp/skins/winamp.zip
$ timidity_skin=/dos/programs/winamp/skins/winamp.zip ; export timidity_skin

 Only one skin can be specified. There are no selector of skin data.
 Since default skin is not prepared, if the environment variable isn't
specified or the data specified is broken, TiMidity++ won't boot.

 This skin interface requires 10 BMP files for skin data such that:

 main, titlebar, playpaus, cbuttons, monoster,
 posbar, shufrep, text, volume, number

 If your data losts some of these files, please borrow the corresponding
data from other skin data.
 In case the data has viscolor.txt, the color of spectrum analyzer is
determined by this file. If not, the default colors are used.


* BOOTING TiMidity++ *

 To use skin interface, boot TiMidity++ with command-line option -ii(tv).
After initializing of internal of TiMidity++, a window looks like WinAmp :)
will raise.


* USAGE OF INTERFACE *

 Features of this interface now available are:

  Prev/Play/Pause/Stop/Next/ Shuffle/Repeat/ Volume/ Exit

with every corresponding buttons.
 Clicking the area of displaying elapsed-time changes the contents of
this area as elapsed-time / remained-time.
 Clicking the area below the time displaying area changes the contents of
this area as spectrum analyzer. If you want to display the spectrum analyzer,
you should compile TiMidity++ with configure option --enable-spectrogram.

 You can move interface's window by grabbing any place of this window.

 To exit this interface, click the little button on right corner.


* BUG? *

 Text displaying area can display only English, digits and some graphical
characters. Any of multi-byte character won't displayed.
 Cannot set panpot, Equalizer and PlayList of TiMidity++.
 There no effect by clicking the eject button.
 Cannot iconize, smalling.


 YOU CAN USE THIS INTERFACE AT YOUR OWN RISK.

                              Daisuke Nagano <breeze_geo@geocities.co.jp>
                      http://www.geocities.co.jp/Playtown/4370/index.html