A DAC chip doesn't do anything by itself - it's part of a whole system, requiring externally provided power, data, clocks, filtering, and amplification. The DAC chip manufacturers specify a worst case or "CYA" value in their data sheets. This is a nominal figure which usually corresponds to how the chip performs on their evaluation board using a certain set of assumptions about the quality of the power supply, passive components and such. Then they add a margin for manufacturing variance, which depends on their own yields and tolerances. So it is not a reliable predictor of how the part will perform in a given application - the purpose of the spec is so that the designer can understand the differences across the manufacturer's own product line - but even for that purpose you can't just look at THD+N to decide which is "better".
Ultimately there are many factors besides the DAC chip itself that affect the measured performance of a device - power supply noise, clock noise, EMI, quality of amplification and filtering, etc. No, we're not magically "undistorting" the DAC's output - rather it's that the total performance of the DAC chip is dependent on the environment into which it is designed. It can go the other way too - on one occasion I was unable to achieve the manufacturer's spec in my design, so I ruled out that part. Usually we find that a DAC performs very close to spec in the first board revision that we build. Then we might make some further improvements depending on what we observe in testing. On cost-constrained products there is only so much you can do, but usually it's possible to squeeze out a couple db. However, in the case of Transporter there is a very large improvement over the manufacturer's specs because we are using the chip in a far more advanced design than AKM's evaluation board. For the THD+N figures this is especially due to the jung regs and the selection of certain passives in the amps/filters. The clocks are very important too, but this isn't meaningfully reflected in THD+N tests. -- seanadams ------------------------------------------------------------------------ seanadams's Profile: http://forums.slimdevices.com/member.php?userid=3 View this thread: http://forums.slimdevices.com/showthread.php?t=55083 _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/audiophiles
