In March, b67c2d72d6 moved some MIDI
initialization from the main thread to the MIDI thread. This caused
MidiPlayer_Midi::sysEx() to run on the MIDI thread for the first time.
This is a problem because it calls SciEngine:sleep(), which polls
events, and that causes MacOS to throw an exception for calling
SDL_PollEvent() on a non-main thread.
While investigating, it also turns out that MidiPlayer_Midi::sysEx()
and MidiPlayer_Fb01::sysEx() were calling OSystem::updateScreen(),
and that also shouldn't be happening on a non-main thread.
Now SciEngine::sleep() is only called on the main thread, and
OSystem::delayMillis() is called on the MIDI timer thread.
Continuing to call sleep() on the main thread keeps the UI responsive
when loading patches, which can take several seconds.
The OSystem::updateScreen() calls had no effect and have been removed.
Fixes bug #12947
This reintroduces two original bugs that I found too nasty to implement when I made the driver. These are present in all SCI0 CMS drivers. They do actually influence the sound in an positive way and make it more faithful when comparing to DOSBox. I have added an #ifdef option so it can be easily enabled and disabled for testing.
There are many situations that GM sounds worse than MT, even for games
that were written for GM.
I think this is related to this.
And, there is a note in midi.cpp:
// NOTE: SSCI uses channels 1 through 8 for General MIDI as well, in the drivers I checked
Also, issue #6686 reports that using the "higher" channels causes
problem on a real SC-55.
This change solves at least: #6686, #9735 and #10297
This adds support for the Roland GS drumkits used by SQ5 and QfG3.
The original Sierra GM driver does not pass the drumkit select MIDI
messages to the GM device, but they do exist in the MIDI data.
Another issue is that the drumkit numbers used are incorrect. This
does work on the original Roland SC-55 devices because they correct
the drumkit numbers. Later devices do not do this. Code has been
added to correct these wrong drumkit numbers.
My commit 0e734722 causes lockups in SCUMM (sometimes) and SCI (very often). I didn't like the way I had fixed this before, but in the end I now had to do it in a similar way.
- The mutex was added to avoid the triggering of the assert in backends/midi/windows.cpp, line 95. Meanwhile, this issue has been addressed differently.
- SCI does not per se require a mutex for the sound drivers. The engine is mostly thread-safe by avoiding driver calls through the main thread.
- Backport some code of waltervn's fork at his recommendation (taken from 97604200 and 4c3bfee5). That code has caught some bitrot and I were too lazy to manually resolve the merge conflicts. All changes have been re-checked with disasm.
- Add initTrack() implementation for SCI0 EARLY/LATE
- Minor cleanup
- Please note that this commit does not fix any shortcomings of the SCI sound engine
After implementing such a dialog into the fb01 driver it did make sense to me to also have this as a feature for all other aftermarket drivers/patches.
So now the sound drivers can report missing files after the failed open() call which will then be displayed in a dialog. Which will at least be more helpful than our usual error messages...
Fix SCI0 (LATE/EARLY) variants of the driver which were broken (didn't play at all). This might be of my own doing, since I introduced the initTrack() method into the driver class and the fb01 driver didn't have one so far.
SCI1 didn't seem to require much fixing. I modified some things according to my findings in the driver disasms. QFG2 and JONES seem to be fine. I am not too happy with KQ5. It has volume issues, but they might be present in the original, too.
I also added an isOpen() check and a mutex to avoid threading issues. When aborting SCI (either quitting ScummVM or returning ot the launcher) while using the fb01 driver I frequently (more often than not) got the assert from backends/midi/windows.cpp, line 95. This fixes that.
I've done plenty of checks and experiments with the sound bank initialization. But I found no bugs or possible improvements there. Hard to tell whether the sound is right. That device seems to have a mind of its own...
Pitchwheel and channel volume settings need to be updated when the channels get reassigned.
This bug caused faulty music at least in KQ5 when walking from the dwarves scene into the harp playing tree scene.
(This triggered an assert in PQ2 at the airport. I've added a warning instead, since the original driver has no handling for that sort of thing. Invalid programs will simply point into invalid memory blocks)
I haven't found an elegant and non-intrusive way to squeeze SCI0 support into LordHoto's existing code. The drivers are too different. So I made some rearrangements. The basic mechanisms of LordHoto's SCI1 code should remain the same as before, though. I only introduced some more classes, moved some code into these classes and renamed some things (mainly for myself, so as not to get confused).
I fixed two voice mapping bugs in the existing driver code. The first bug in bindVocies() effectively hindered the driver from playing anything at all when the CMS_DISABLE_VOICE_MAPPING #define wasn't set (_voice[i].channel == 0xFF instead of _voice[i].channel != 0xFF). The second bug in unbindVoices() was not a complete show stopper, but the function simply did not "unbind the voice". The line which does the actual removal of the channel assignment was missing.
The SCI0 driver portions have been tested with: PQ2, KQ4, LSL3, QFG1, ICE and COC.
SCI_0_EARLY versions apparently don't support the CMS. At least I haven't seen a driver file so far. And there seems to be no no instrument patch resource. Although the latter issue needn't necessarily be one, since the patch data array in the driver is actually preset with data (which gets overwritten as soon as a patch file is loaded). Maybe this would work for SCI_0_EARLY. However, I haven't tested this, since I really would have have a look at a driver file first if one actually exists. For now, I have limited the driver to SCI_0_LATE.
SCI1 has been tested with KQ5 and LSL5 (not extensively, just to see whether anything got broken and whether my voice mapping fixes work).
Fix up SCI0_LATE variant of the driver so that it actually plays anything (and even correctly). SCI1 hasn't seen any testing from me so far. I don't know whether that version works. Same for SCI0_EARLY...
I haven't found an elegant and non-intrusive way to squeeze SCI0 support into LordHoto's existing code. The drivers are too different. So I made some rearrangements. The basic mechanisms of LordHoto's SCI1 code should remain the same as before, though. I only introduced some more classes, moved some code into these classes and renamed some things (mainly for myself, so as not to get confused).
I fixed two voice mapping bugs in the existing driver code. The first bug in bindVocies() effectively hindered the driver from playing anything at all when the CMS_DISABLE_VOICE_MAPPING #define wasn't set (_voice[i].channel == 0xFF instead of _voice[i].channel != 0xFF). The second bug in unbindVoices() was not a complete show stopper, but the function simply did not "unbind the voice". The line which does the actual removal of the channel assignment was missing.
The SCI0 driver portions have been tested with: PQ2, KQ4, LSL3, QFG1, ICE and COC.
SCI_0_EARLY versions apparently don't support the CMS. At least I haven't seen a driver file so far. And there seems to be no no instrument patch resource. Although the latter issue needn't necessarily be one, since the patch data array in the driver is actually preset with data (which gets overwritten as soon as a patch file is loaded). Maybe this would work for SCI_0_EARLY. However, I haven't tested this, since I really would have have a look at a driver file first if one actually exists. For now, I have limited the driver to SCI_0_LATE.
SCI1 has been tested with KQ5 and LSL5 (not extensively, just to see whether anything got broken and whether my voice mapping fixes work).
I haven't found an elegant and non-intrusive way to squeeze SCI0 support into LordHoto's existing code. The drivers are too different. So I made some rearrangements. The basic mechanisms of LordHoto's SCI1 code should remain the same as before, though. I only introduced some more classes, moved some code into these classes and renamed some things (mainly for myself, so as not to get confused).
I fixed two voice mapping bugs in the existing driver code. The first bug in bindVocies() effectively hindered the driver from playing anything at all when the CMS_DISABLE_VOICE_MAPPING #define wasn't set (_voice[i].channel == 0xFF instead of _voice[i].channel != 0xFF). The second bug in unbindVoices() was not a complete show stopper, but the function simply did not "unbind the voice". The line which does the actual removal of the channel assignment was missing.
The SCI0 driver portions have been tested with: PQ2, KQ4, LSL3, QFG1, ICE and COC.
SCI_0_EARLY versions apparently don't support the CMS. At least I haven't seen a driver file so far. And there seems to be no no instrument patch resource. Although the latter issue needn't necessarily be one, since the patch data array in the driver is actually preset with data (which gets overwritten as soon as a patch file is loaded). Maybe this would work for SCI_0_EARLY. However, I haven't tested this, since I really would have have a look at a driver file first if one actually exists. For now, I have limited the driver to SCI_0_LATE.
SCI1 has been tested with KQ5 and LSL5 (not extensively, just to see whether anything got broken and whether my voice mapping fixes work).
athrxx
Cameron Cawley
Donovan Watteau
Eugene Sandulenko
Orgad Shaneh
sluicebox
Paul Gilbert
Filippos Karapetis
NMIError
Walter van Niftrik
Thierry Crozat
Zvika Haramaty
Bastien Bouclet
D G Turner
Walter van Niftrik