VirtualSpeaker 1.2.0

-- Some games using Carbon Sound Manager still fail at launch even when using VirtualSpeaker 1.2.0
This problem affects a variety of commerical games, shareware / freeware games, and game abstraction libraries, which use the Carbon SoundManager to play audio. This problem was caused by an incorrectly set bit (fIsMixable) in one of the virtual audio formats that was used in the emulation of onboard audio hardware. This issue has been corrected in the 1.2.1 release.

I'd like to apologize for the problem, which was due to a bug introduced in the last development build before the release. Games using the Carbon API were tested prior to that build (and would work correctly with VirtualSpeaker as the audio output at that point) but were not sufficiently tested with the final release. The testing phase for the final development build was short, since both of my testers had been using the prior version for quite a while with no problems, and only minor changes (intended for optimization) were made at that point. I use my Cube mainly to capture DV video, and although I was continuously testing VirtualSpeaker for compatibility as the audio output on my primary computer (a G4 tower) during my day to day use, I haven't had much time to play games lately. So, the problem with games using the Carbon SoundManager made it into the release. All work and no play make for bad QA (quality assurance) :)

-- During installation, some permissions may be changed or set incorrectly.
This problem was caused by a setting in the VirtualSpeaker 1.2.0 installation package (which was set to the default) that is ignored by Apple's Installer program. This caused the Installer program to overwrite some permissions that should not have been changed during the installation process. A bug report has been filed with Apple regarding the specific issue in the Installer program which caused the problem, and the installation packages for later releases of VirtualSpeaker are built in a manner which should avoid the problem entirely.


PseudoSpeaker 1.1 and PseudoSpeaker 1.0

-- Internal versions, not released to the general public.
These versions were not intended for use by the general public, and were never officially released. They were distributed to a limited number of people who either knew me directly or happened to post about their problems on a list or forum I read. These versions of the kernel extension do not load properly under MacOS 10.2, and used an external library inherited from the Apple source code examples (libAudioFloatlib.a). In the 1.0 version, the name of the library was unchanged and had a tendency to be overwritten by other code based on Apple's sound driver sample source code that didn't change the name of the library either.