[Arny Krueger]

"Audio Empire" wrote in message


While your observation is correct, your conclusion is
flawed. Why would you suspect nonlinearity in a system
made up of discrete components rather than standard
chips? Why would it be more likely that a discrete
circuit would have more nonlinearity than an IC?
Especially when it is easier to correct such nonlinearity
in a discrete circuit (after the fact) than it is in a
chip that is found to be a faulty design? As for
repeatability, as long as precision parts are used, there
should be none.

The challenge with traditional DACs that are made of discrete parts is
linearity. The DAC is composed of a large number of resistors whose
resistance needs to have matched ratios within a tiny fraction - something
like 1/1,000 of a percent. One way to completely avoid this problem is the
well-known Sigma/Delta design, but it appears that DCS has made a big point
of not using this methodology. It is eaiser to match and stabilize the
ratios of resistors on a chip because then all of the parts are tightly
clustered in a microscopic area whose temperature is far more likely to
remain consistent.

That all said, the Stereophile review of the DCS Scarlatti appears to show
sonically blameless technical performance. But so does the DAC in my $30
Sansa Clip +. Yes, the Clip is far less blameless based on technical
measurements, but in actual listening tests, it is blameless.

And that has been my point all along. We have modern equipment with truely
amazing absolute performance and incredible price/performance. But, we have
the same old ears which were designed to hear approaching enemies and
receive verbal messages, not sort audio gear.