I assumed it was something of that general nature, Merc. It's good to know the details.
More broadly, in my market, sound is a complete dead issue. Want to know how many soundcards I've sold this year? One. It was a Creative Sound Blaster Audigy, bottom-end model, to someone who came in and ordered just the sound card itself, nothing else. (He already had a nice system. Needed an Audigy — any Audigy — for some particular special purpose application the details of which have slipped my memory me.)
System buyers sometimes ask about the standard of the on-board sound. One in 10, maybe. No more than that. Some say "my old board has on-board sound and it's not very good, do I need a sound card?" I tell them that on-board sound has improved a lot, that they can expect a fair random current mainboard to produce sound of equal or better quality to the old Sound Blaster Vibra 128. (Nearly everyone has owned a Vibra 128 at one time or another.)
They say "Oh, that will be fine. Forget the sound card. Now, about that LCD monitor ...."
Fair dinkum, I could ship computers without any sound at all and half of them wouldn't even notice. As a selling point, sound is a complete waste of time. No-one is interested.
--------------------
Doug. Yes, you are quite right. For a first-ever chipset, the Nforce 1 was a damn fine effort. It's just a pity that they don't seem to be able to improve on it. I think their problem is that they aren't putting in the engineer time on driver improvements. Have you noticed that they practically never release updated Nforce drivers? Trust me, it ain't because the v1.0 was perfect in every detail.
A good simple, practical measure of a chipset's quality is the number of what we call "floaty-round motherboards" we have in the workshop at any given time. These are boards that, for one reason or another, got unpacked, usually bolted into a system, and then came out again. Usually, it's because they have some kind of a question mark over them. Maybe they are faulty, maybe they are OK. Sooner or later, we have to go through them one by one and determine if the fault is (a) real, and (b) reproducable enough to RMA them.
Good boards don't keep "floaty round" status for long. They get pressed into service for one odd job or another quite quickly. Out-and-out faulty boards don't hang round for long either: they soon get marked with the fault and go out the back onto the RMA pile.
It's the "works most of the time" boards that stay of floaty-round status practically forever. You can't pin an honest-to-goodness fault on them that's reproducable enough to RMA the board (not without spending a ridiculous amount of time on each one) and yet every time you need a board to do an actual job you have to choose another one because you can't actually trust the floaty-round board not to come back again.
Right now we have about a half-dozen floaty-round motherboards. Five of them have Nvidia Nforce II chipsets. (The other one is an Asrock KM-600 that refuses to power up for several hours about twice a month, then works perfectly again.)
Nope, it's not a brand thing. Of our five problem children, three are (from memory) Albatrons and two Gigabytes. One of the five is an on-board video job, the others are high-end boards with goodies like Firewire. Two different manufacturers, three different models. And don't even think about mentioning those pox-ridden ASUS disasters — they float straight over to the garbage bin: they were dreadful: greater than 50% return rate. Considering that even at the height of their popularity, Nforce II boards only accounted for maybe 20% of our motherboard sales mix, that 5:1 ration is a telling indictment of the chipset. And it's not unusual. We RMAed an unusually high proportion of them when we were handling more of them.
We also has heaps of weird sound problems with the Nforce II. Like about 30% of systems shipped. You can't even plug in a stand-alone sound card and/or swap the motherboard: the noise through the speakers just comes back anyway after a reinstall and a week or two. The only cure is to substitute a VIA board. I'm virtually certain it's a driver problem.
But guess what? Nvidia don't do driver revisions.