I'm not entirely convinced PCI latency is the sole caused in every case. I was able to reproduce the "squeal of death" a few years back, but couldn't find enough people who had this problem to do any comprehensive testing with my "solution".
It was with system running a 850 Duron and Epox 8KHA+ (VIA kt266a). The sqeal would come during gaming sessions or whenever listing to music while surfing. The sound card was an SBLive 5.1 running under WinXP (pre sp1). After trying literally everything, I finally replaced the card with a TB Santa Cruz.
Initially, the problem disappeared, but then one night the squeal came back during a session of UT and then another time while running a different game. This convinced me that the problem was still there, but just not as prevalant as before.
It was around about this time that I noticed what I thought was a completely unrelated issue. The CPU voltage tended to fluctuate slightly every now and then even though I wasn't overclocking. Since voltage fluctuations are fairly common, I didn't think any of it, but thought that it would be a good idea to stabilize it a bit by upping the current a bit - in a sense, over-volting the CPU. Don't remember how much I raised it, but it was enough to cover the "low" range fluctuation.
The squeal never came back after that. Now since the problem was never as common with the TBSC as it was with the SBLive 5.1, I didn't actually think the problem had gone away. After while, I needed the TBSC for another box so I stuck the SBLive 5.1 back in. To my surprise, the squeal never came back even with the SBLive 5.1 running in the system. I kept this box running with the SBLive 5.1 for the next 6-7 months or so and not once did it exhibit the problem. I even tested it briefly with an Audigy OEM during this time and it seemed to work fine as well.
Just to add, I've never experienced this problem with a non-VIA based motherboard.