A great overview of the problems with PulseAudio:
It is 2 years old, but the facts in the article above are still completely true. PulseAudio has made essentially zero progress in the last 2 years, which is why it should be abandoned.
Open up any emulator program on Ubuntu and it will skip like mad. Same with many native games such as Lincity-ng or OpenSonic. This is as most games on Linux depend on sound timing, which the high latency nature of PulseAudio messes up.
I am greatly concerned that the non-functionality of PulseAudio is hampering the beginning of a commercial game industry on Linux. Developers need working APIs to make applications. They will not tolerate game development using a half-working API. I feel that there never be a wide spread game industry on Linux as long as PulseAudio is
in widespread use.
I have nothing against the ideals and theories behind PulseAudio. It is just their implementation does not work and it seems it will never actually work as intended. Libsydney has never come to be. It is time we look at alternatives.
A good possible solution would be switching to OSS4 and writing an audio wrapper for it to make it easier for developers to use. OSS4 is much more simplistic and (arguably) cleaner designed then ALSA, which would likely made this an easier task.
I have already removed PulseAudio completely from my distribution because I have found it greatly interferes with multimedia playback and gaming. I have received no complaints from my users, in fact, many of them have switched over to infinityOS specifically because I do not include PulseAudio.
Let's not waste any more effort on a failure.