If it turns out that everything that had previously been fixed by
this manager is broken by this change, everything that had been
fixed probably could have been fixed by just increasing the audio
buffer size in SdlMixerManager. :\
The previous default buffer size of 4096 samples for 44kHz mixer
would add up to 93ms of audio latency, which is fine for early
adventure games, but this is significantly more latency than is
acceptable for games with full motion video. For these games,
the latency needs to be kept within roughly +15ms and -45ms of
video frame presentation to avoid lip sync problems. With this
change, the default audio buffer size is calculated to be 1024
samples at 44kHz (which happens to match what DOSBox uses).
There is a possibility that the reduced latency may cause issues
that did not previously exist with things like the MT-32 emulator,
where a larger buffer size allowed for a larger window where
high-complexity synthesis that could not be generated in realtime
could be balanced out by low-complexity synthesis that could be
generated faster than realtime. In this case, rather than
increasing the system mixer buffer size again, please move the
MT-32 emulator into its own thread and give it its own larger ring
buffer into which it can generate more sample data in advance,
independently from the rest of the audio system.
For other systems where this buffer size reduction might cause a
problem with audio drop-outs, a new audio_buffer_size
configuration option has been added to allow users to tweak the
audio buffer size to match their machine's ability to generate
audio samples.
Fixes Trac#10033. Also improves playback of samples in SCI that
were programmed to restart across several consecutive frames,
relying on lower audio latency in the original engine for this to
not sound bad, like the hopping sound at the start of chapter 1
of KQ7, and the sound of turning on the power in the digger train
in the Lighthouse volcano.
Removing this GUI control was suggested as far back as 2011 at
<http://lists.scummvm.org/pipermail/scummvm-devel/2011-November/010416.html>.
There were no objections, but it was never removed. When working
on audio latency bugs, I independently rediscovered that the GUI
option was broken: the per-game options would *never* work, and the
option would not take effect until ScummVM was restarted because
there is no API for interacting with the backend audio mixer. So
now, it is finally gone.
Primarily for the sake of future troubleshooting, configurability
of the audio sample frequency within SdlMixerManager is maintained
for the moment, but now users will need to edit their ScummVM
configuration file manually to change it.
OpenGL backends don't always support the pixel format that is
returned by the Indeo 3 decoder when playing the GK2A.AVI from the
GK2 demo. If this happens, use the backend's preferred pixel format
and convert in software.
If a backend doesn't support any 16-bit or 32-bit format, the
playback code will error out. This is probably fine, since there
are not really any of those any more.
Fixes Trac#9994.
_hwscreen is always initialized to 16bpp so the supported 32bpp
pixel formats would never be put into the list of supported pixel
formats, making it useless for engines to query for usable 32bpp
pixel formats.
This patch changes things so that the native desktop pixel format
is at the top of the supported formats list, and all pixel formats
<= the default desktop pixel format will now show up in the list
of supported formats. ("Supported" is somewhat of a misnomer here
since there is no hardware querying beyond checking the default
desktop pixel format. SDL generally accepts a wide variety of pixel
formats and tries to convert them to whatever the hardware
supports.)
This is an attempt to fix a compilation error on some platforms.
The error message seems to indicate that log2 might be a define
on thos platforms.
Note that the log2 implementation in ModuleModXmS3m is not the binary
logarithm, and we cannot use Common::intLog2.
In commit f0581bab4a the index value
for the array was changed from a constant to a variable, but uint8
is too small to contain 293, so this value was ending up as 37.
When -Wundeclared-selector is enabled (recommended by Apple), the
calls to the setBadgeLabel selector in MacOSXTaskbarManager are
warned on because NSDockTile declarations are not included because
they do not exist in macOS 10.4 and earlier. While I don't know
that we are even supporting such old macOS versions these days, it
is simple enough to fix this problem when compiling to modern
macOS versions by conditionally including the necessary header.
c19c10d548 fixed the bad parsing of
module.mk which meant the Xcode generator broke. For the moment,
just move the hack for browser_osx.mm into the generator, until
there is a more elegant solution to this problem (if ever).
FluidSynth does build and run successfully, but it is necessary
to get the correct static libraries. It can be manually built
against dylib, if desired, by manually adding the fluidsynth.dylib
to the Frameworks. Since this will result in a runtime dependency,
this is not set up by default.
Libraries were being included unconditionally regardless of the
arguments passed to create_project, and in the case of macOS were
being added using manual linker flags and other such hacks instead
of using the Frameworks list.
Previously, _topNibble was not reset at the beginning of a new
audio block, and the alignment byte at the end of odd blocks was
being read as audio data, which caused audible clicks and
out-of-bounds sample generation. There may have also been read
errors related to the use of continue/break keywords inside of a
macro wrapped with do-while(0).
The introduction of partial block reads in this code when it was
converted from ffmpeg to a ReadStream interface was also confusing
and somewhat inefficient (calling SeekableReadStream::pos
frequently), so this code has been refactored for clarity and to
improve efficiency by reducing the number of virtual calls. Error
detection has also been improved somewhat by ensuring that there
are enough bytes to read a block header, and that the step indexes
in the header are within the valid range.
