understand, it just means that the currently active parts are too important
(i.e. have too high priority) to throw out for this new one.
This makes bugs #1325060 and #1325487 less severe.
svn-id: r19062
As defined in Patch [811623] MT-32 patch for Bug 810564
Added a channel mask to MPU-401 devices so that --native-mt32
may force the device to use only the subset of MIDI channels
actually supported by the MT-32. Also added a best-guess
interpretation of iMuse Part priority in the SysEx 0x00 msg,
since part priorities become more of an issue when the
channel count is cramped.
svn-id: r10409
Implemented a parser for Euphony music. No
FM instrument support yet, as the FM chip used
by FM Towns is not being emulated yet. In the
meantime, a stock FM-emulated GM instrument is
being used instead.
This at least makes the Zak Towns kazoo tune
and the Loom Towns distaff audible. Emulation
of the FM Towns synth chip, or suitable
emulation using the OPL2 synth, is still
under investigation.
svn-id: r10265
a simple number. So that it's still obvious
that the number refers to version, a VER()
macro does nothing to the number but makes
it obvious what the number is for. This
has the exact same effect as the enum that
was going to get more and more lengthy
as the savegame version evolved.
svn-id: r9901
Reverted Hoenicke's change. Why was this change made anyway?
Do not reinstate the "tempo tweak" before talking to me
about why it was needed.
svn-id: r9810
RO streams have built-in timer markers (the
unknown 0xA0 mentioned in FR [742249]) that
override automated methods of updating
VAR_MUSIC_TIMER. Since these timer markers
do not seem to be present in AD resources, the
old mechanism is used if MD_ADLIB is the current
MidiDriver.
svn-id: r9720
in monkeyvga/ega. Tested in monkeyvga, but
not in monkeyega yet. There are still wrinkles
to work out, but the parser for the new format
is in and hooked up.
svn-id: r9684
When scanning (not jumping) to a target location
in a different track, the current track is first
scanned to the end to update the player state
before starting into the new track.
Miscellaneous scan-related cleanup is included.
svn-id: r9589
V6 scripts can specify up to 16 parameters to be passed to
doCommand(). In most cases 8 or fewer parameters were used,
but occasional uses of 9 or 10 parameters resulted in
incorrect iMuse behavior due to lost doCommand() data.
ImTrigger was also upgraded to support 8 parameters instead
of 4, since the incorrect behavior that was observed involved
the use of triggered commands that required more than 4
parameters. Since ImTrigger data is not currently being saved
(which is bad), we do not at this time have to go to another
savegame format to support the additional ImTrigger data
(which is good).
This fix corrects a problem in the Tunnel of Love. When the
"execution scene" music finishes and the hidden door opens,
the Tunnel of Love music is supposed to start up again. In
ScummVM, it never would, do to the loss of parameters for a
trigger-based "Parameter Fade" command.
A bit of miscellaneous cleanup has also been made.
svn-id: r9579
Tightened up OOB track selection handling in
MidiParser. Apparently some SMFs in the Mac
version of MI2 (and other games?) specify
track -1 in their jump commands to indicate
a jump within the same track.
svn-id: r9321
