I'm not making excuses, but in all fairness, the audio bug is not something that was even really discovered in the wild for about 6 months and the severity was limited in impact. There was also a simple work-around... DDT in the world of QA (Don't do that). I know this wasn't palatable to many, but it had to be triaged against the myriad of other issues that are all demanding resources on any given day.
In general, it's just not economical to test every possible use case. In addition, there's often insufficient business case to justify a fix for some defects (eg. suitable work-arounds exist or the negative consequences are insignificant compared to the cost of fixing). All software (and therefore most hardware) ships with bugs, it's just a question of the number, severity, and the use cases which trigger them that determines the perceived quality. Most QA departments establish benchmarks which must be met (eg. no sev 1 defects, less than X sev 2 defects, etc.). There's likely hundreds of known bugs in OSX 10.6.2 and likely just as many unknown bugs, but the majority are of insignificant severity or take unusual circumstances to invoke.
Given Apple's fairly liberal warranty (and the costs associated with it) and discerning user base (aka picky), I'd say they are probably very sensitive to the risks of poor QA and probably take more measures than most to minimize their after sales support and warranty costs and therefore invest more than average in QA.
