These changes make use of _mixer->getOutputStereo() to select the proper number of audio channels.
This makes it possible to have mono audio produced using the proper DiMUSE internal mixer mono
routines, instead of having the ScummVM mixer folding down a stereo mix.
This code is present in all the original interpreters using Digital iMUSE, and it was implemented
in order to allow slower CD drives to fetch new data without having music skipping during load times.
This works by basically flooding the music buffer with data whenever a big resource, a voice file, or
a SMUSH movie is loaded.
Arguably this is not needed for devices using SSD drives, but since this change had its benefits on a
quite old laptop of mine, and since the memory overhead is very minimal, I have decided to implement it.
Previously we had asserts instructions in place for this; they work under the assumption that the user
always has the sound files available. Unfortunately this assumption breaks in at least two cases:
- The user plays the COMI demo (small version, without sound files);
- The user copies the whole COMI game from the CDs/Steam to the device of choice (e.g. tablet),
and doesn't copy the BUN files over (e.g. for storage issues).
We simply remove the assertions and put a check in its place, and if that check fails we just exit
the function and resume normal execution. This fixes#13845.
This replaces every single strcpy call in the engine with
Common::strlcpy.
Some of these might cases seem a bit pointless, but it is
supposed to discourage future use of strcpy. If people don't
see a single occasion of it they might think twice before they
use it...
The map in IMuseDigiDispatch can contain unaligned blocks if it
contains a TEXT block with a length not divisible by four. This causes
unaligned memory accesses in a few places.
This commit fixes the issue by ensuring only aligned pointers are typed
as `int32 *` while unaligned pointers are typed as `uint8 *` and are
only accessed through helper functions.
Every functionality has been implemented (audio options, text options, saving and loading).
The only thing currently missing from the menu is the thumbnail handling.
The call was unnecessary since the one being fundamental is the call made
to the resource manager mutex in _vm->getResourceAddress(); this fixes
some very rare (but also very real) cases of v7-8 games softlocking.
This fixes instances in which audio chunks with a sample rate equal to a
quarter of the global one were not mixed correctly, causing wave aliasing
during the reconstruction of the analog signal from the digital one.
Without this fix, if we had a list of tracks A->B->C->D and we stopped C, we couldn't have stopped D because we reached the end of the list prematurely.
The missing check can cause an issue in The Dig where the sound slots are being filled with the same sound file (background music) after causing a sequence of these two engine actions:
IMuseDigital::setDigMusicState(): Set music state: stateNoChange
IMuseDigital::setDigMusicState(): Set music state: <any valid state which corresponds to a loadable file>
As usual, thanks athrxx!
Big thanks to athrxx for reporting this issue; the previous calculation for obtaining an optimal number for _maxQueuedStreams did not account for the differences between the sample rate targeted by the audio backend and the one targeted by DiMUSE (which is 22050Hz).
This new formula has been tested for optimal latency and absence of underruns for all supported games and all known combinations of sample rates and output buffer sizes:
[22050, 44100, 48000, 96000], [1024, 2048, 4096, 8192].
AndywinXp
BLooperZ
Le Philousophe
athrxx
Dominik Kreutzer
Donovan Watteau
D G Turner
rsn8887