I have two brothers - one is a DJ and the other a sound engineer. Both are also into computer music production.<p>I know (because I've experienced it first hand) that when using software like Reason, Cubase or Logic, the difference between a 2ms and, say, 8ms delay is noticable when using MIDI keyboards. The time between hitting a key on the MIDI keyboard and hearing the note is very noticable.<p>Then again, the issue here is that the time between hitting the key and sound reaching your ear is too high, so theres obviously the added latency of the sound travelling from your soundcard to your ear (since the quoted latencies are those from when sound is generated until its processed by the soundcard). So by minimizing the latency in the OS sound system, you have more time to play with to generate the sound and still hear it before its noticable.<p>As I see it, there are 5 latencies involved:<p><pre><code> keyboard<-1->software<-2->sound system<-3->soundcard<-4->speakers<-5->ears
</code></pre>
There is little that can be done about 1, 4 and 5 (at least, not by the musician or programmer, this is up to the hardware guys - besides perhaps the MIDI driver). We want to maximize how much can be done by the software, so by minimizing 3, we have more time for use in 2 before its noticable.<p>So, yes, for musicians and sound engineers, low latency in the OS sound layer can be important, in my opinion.