The idea here is that if two voices are generating the same tone,
the square wave generator should be in sync so that the wave forms
amplify each other, rather than cancelling each other out.
In AGI games, the numeric keypad is used as an alias for cursor keys to
allow stationary and 8 directional input, but while the predictive
dialog is open, this is more sensibly mapped to numeric input to
allow use of the predictive input on desktop ports.
This fixes a leak in PreAGI games (which never deleted their
RandomSource), ensures that PreAGI's RandomSource has a name (and hence
is registered with the event recorder) and even slightly simplifies the
AgiEngine destructor.
This also removes the dependency of engines on the event recorder header
and API, and will make it easier to RandomSources that are not properly
registered.
This unifies the engine names in MetaEngine::getName() and the
credits. In particular drop "Engine" or "engine" from the names when
it was present and use expanded names in credits when the
MetaEngine uses it (e.g. "Beneath a Steel Sky" instead of "BASS").
Right now, a few places in the frontend code still use printf and
consorts. We mark the affected files with a FIXME for now, and
add a dedicated exception for each. To be fixed!
Also tweak FORBIDDEN_SYMBOL_REPLACEMENT to hopefully really always
enforce a compiler error
MKID_BE relied on unspecified behavior of the C++ compiler,
and as such was always a bit unsafe. The new MKTAG macro
is slightly less elegant, but does no longer depend on the
behavior of the compiler.
Inspired by FFmpeg, which has an almost identical macro.
Make the player be centered on a fixed number of "generators" instead of
MIDI channels that arbitrarily allocate generators for notes.
Make the audio stream to be stereo and for sample rate use _sampleRate.
Rewrite the synthesis core:
* Make generators use both oscillators
* Implement swap mode for oscillators
* Fix envelope update frequency
This also should fix some regressions from the previous
commits, related to MidiParser's either being leaked,
or being deleted and then used again (i.e., crashing).
I tested as many games as I had available, but further
testing of all affected engines is called for anyway.
The regression affected AGOS and maybe some others; specifically,
the real MidiDriver would have been deleted twice -- I previously
missed that the Engine instances takes care of freeing the real
MidiDriver, not the MidiPlayer wrapping it.
This commit should clarify the ownership of the real MidiDriver for
most pseudo MidiDrivers.
Many engines follow the advice in audio/midiparser.h and create a
"pseudo-MidiDriver" subclass. But MidiParser really only needs a tiny
subset of the MidiDriver capabilities, namely those found in
MidiDriver_BASE. So we try to subclass from that whenever possible; this
allows us to remove many stub methods, and enables further future
simplifications.
This in turn enables modifying MidiDriver_MPU401::close() to allow
it to be called on a midi driver that has not yet been opened.
The specific issue that triggered me to make these changes was a
crash-upon-quit in HUGO, caused by it instantiating a midi driver,
then encountering an error (missing hugo.dat) *before* having
opened the new midi driver; the general cleanup code then tries
to close the (not yet opened) midi driver -> kaboom
Also fixed some engines which were leaking MidiDriver instances.
Using the timer mechanism for just a simple counter is not just
overkill, its also inaccurate. When using a call frequency of x,
and waiting for y callbacks, the passed time will not be x*y.
The problem amplifies on slower platforms and/or fair thread
schedulers.
Use absolute times instead. Most notably, the walking speed of
the avatar is now smooth on android, but probably on all other
handhelds we support too.
The new code is simpler, avoids a potential buffer overrun
(by avoiding to to use a buffer in the first place), and
hopefully has slightly more sane matching properties.
svn-id: r55135
* names now comply to our naming conventions
* the function computeStreamMD5AsString which computes the MD5
as a hex string now returns it as a Common::String
* add doxygen comments
svn-id: r54121
Forward-ported from branch. Let's do this the easy way for now, even
if I'm not sure why the save dialog shows the autosave slot anyway.
svn-id: r52997
We have to update cursor position, not just terminate the input
buffer. This makes Ctrl-C work as intended, which fixes one half of
bug #3054184 ("SQ1 AGI: keyboard special keys are not functioning
properly"). I'm not sure if the second half actually is a bug.
svn-id: r52899
The initial screen color needs to be 0 and not 0xf for these. Winnie C64/Apple II should be completable now, with some graphical glitches and palette problems.
svn-id: r52881
(After a brief discussion with sev.) AdLib is now the old Sarien sound,
PCjr is the new PCjr emulation. PCjr is still the default; we don't
need to ask getMusicType() how to get there.
svn-id: r52838
Because chanGen() gets called very early, and sometimes it crashes
ScummVM as soon as the game starts. Actually, it's probably enough to
initialise a couple of fields in _tchannel[], but I figured it
couldn't hurt to clear all of _channel[] and _tchannel[].
svn-id: r52741
This is currently done in the engine code. I adapted AGI, AGOS, DRACI,
GROOVIE, LURE, MADE, QUEEN, SAGA, SKY, TINSEL and TOUCHE to send a reset
device on startup. The sound output still works fine (started up a game
from every engine), so this should hopefully not introduce any regressions.
As far as I can tell it seems that SCUMM does send a proper device reset, so
I did not touch it. KYRA only sends a proper reset for MT-32 currently. I am
not sure about SCI though.
This fixes bug #3066826 "SIMON: MIDI notes off when using RTL after SCI".
svn-id: r52736
Formerly the AGI MIDI code did not setup the channel mask properly, in case
"native_mt32" was set. This resulted in one missing channel (i.e. channel 0),
since the MT-32 only responds to data for channels 1-9.
svn-id: r52718
Bug #3034370: "Manhunter: New York keyboard unresponsive". Actually
the keyboard was made responsive after June implementation of set.key,
still the game speed was increasing. Fixed that too.
svn-id: r52369
Bug #3017911: "AGI (Fan): SQ0 - Ego flickering". It is a regression from r49745,
as it appeared, some of the screen updates were too much.
svn-id: r51223
* Remove _isAttached member var and isAttached method
* Engines now always call the onFrame method; whether it does
something is decided by the debugger class resp. its subclasses
* Make detach() protected instead of private, so that subclasses
can invoke it
* Remove _detach_now member var (call detach() instead).
* Rename _frame_countdown to _frameCountdown and properly
document it.
* Add more doxygen comments
* Cleanup
svn-id: r50963
