an in-memory WAV file. At the moment it's only used in one place, which is
a bit silly, but I hope to use it for the cutscene player to figure out
when to start the lead-out music.
(To do that I'll need to know how long the cutscene is, though. I haven't
looked into how to find that out yet.)
svn-id: r12424
fading-up music to distort and eventually deteriorate into white noise.
This was because I allowed _fade to be set on channels that weren't playing
and because I only checked _fade for equality when deciding when to stop
the fading. It should work much better now, I think.
svn-id: r12392
uninitialized values. Now the only warnings I got were from libmpeg2
itself, and I don't know how serious that is.
I've also added some code - disabled by default - to allow the cutscenes to
run with libmpeg 0.3.1, since that's what I've got on my Linux box. It
appears to work on that one, though I only have the "eye" cutscene on it
yet.
Ogg Vorbis playback is still broken for me under Windows, though. I wonder
if it is because I don't have the very latest Ogg Vorbis libraries on it
(since I didn't manage to compile them under MinGW). But surely the file
format hasn't changed in any important way...?
svn-id: r12374
grows larger, both when fading up or down. This fixes the problem where the
volume would "jump" when changing the fading "direction" of a stream.
Also changed the logic for deciding which music stream to stop if both
streams are playing and a third stream is started. Before it always tried
to pick the one that was fading down. Now it will pick the one with the
lowest volume, assuming that the more faded a stream is the lower its
volume.
Together, this should fix some abrupt music changes at the watchman's hut,
where it would sometimes start two music streams in rapid succession.
svn-id: r12372
interpolation frames. It shouldn't delay anything, but I'm hoping it will
allow the other threads some breathing room, which might help bug #875683.
Or not. We'll see.
svn-id: r12357
minor cleanups. (Mostly spacing and indentation).
Unfortunately the Ogg Vorbis playback causes ScummVM to crash for me, so I
can't say for certain that I didn't break anything.
svn-id: r12356
overlapping.
Enabled some sanity-checking code.
Fixed clearWalkGridList() so that it clears the entire walk grid list. That
was a really stupid bug of mine, and I hope it had at least some bearing on
the rare (for me, at least) walk problems I've seen.
svn-id: r12154
renders the entire screen every frame, but it tries to update (i.e. copy to
the backend) only the parts of the screen that actually changed. At least
approximately so.
svn-id: r12142
