Most current DACs measure so well with respect to what humans can hear that aural testing is no longer really relevant. For DACs, the measurements are more a surrogate for engineering care, because it only takes engineering care to make a sonically transparent DAC these days. One that measures poorly indicates sloppy or careless engineering. It could indicate that he designer was purposely coloring the sound to attain some other objective, but none of the DAC manufacturers will own up to that--they all claim transparency.
So, I shop among those that measured well simply because if they were careful about that, they were probably careful about other things, too. But mostly I shop to get the specific connectivity, features and look I need, and use the measurements to confirm that the engineering meets a sensible quality standard. Sometimes, I look at the brand, too, but never as a surrogate for performance, given that good performance data is available.
Those who understand measurements and replace a DAC with a 105 dB SINAD for one with 120 dB SINAD are just exploring a hobby, not expecting any audible difference or improvement.
When I read that one DAC has a bit more open bass than another (a statement I did read recently), and both measure transparently, then there are two possibilities: 1.) we aren't measuring the correct things, and 2.) there is some interaction between the DAC and my system that did not occur when the measurements were made. With DACs, I don't for a moment believe that either of these is relevant. There are other components where the second may be true.
But the first thing is a point of real contention and even religious debate. The measurement folks want to see some evidence that what a reviewer "clearly" hears is an effect that reviewer can reliably hear without prior knowledge of what's producing the sound. This seems a reasonable request, IF the reviewer is making value judgments intended to be instructive to others. The language used by reviewers is often intended to be instructive (recognizing the difference between motive and intent), such as "The Fryburger Electronics Bozostat provide more authority in the bass, and a bit more clarity in the upper treble, but it costs ten times as much, so you would expect that." That statement reveals a bias (biases are just expectations that color our impressions), but it also tells the well-heeled reader that the Bozostat is actually better than the cheaper device, when it might not be. The readers here generally don't expect the more expensive device to sound better, without demonstrating that it does either using measurements, or, in the absence of relevant measurements, by preference testing that controls for those biases.
So, if the argument that we aren't measuring the right things is true, then we should expect that those subjective impressions should be repeatable and unbiased if they are to serve as advice to the reader. If we can't do that, then we shouldn't offer advice at all based on hearing. We can say, "the Dumbostat didn't have the same authority in the bass as did the Bozostat, and we wondered why. Measurements show that as the output of the Dumbostat approaches maximum, we see a trough in the response between 80 and 150 Hz, and this explains the loss of authority. Our usual testing didn't duplicate this particular case." Or "We first thought the Bozostat was more authoritative in the bass, so we checked the levels, this time using a voltmeter, and found that the Bozostat put out 0.3 dB more output than the Dumbostat, and when we adjusted for that, the differences faded away." Even this is acceptable (because it is honest), "Even level-matched, there was something about the Bozostat that impressed me more, but I can't conduct a proper blind test and am probably swayed by its beautiful finish and what I know about Fryburger Electronics."
Of course, this sort of thing is always welcome in reviews, because it tells us stuff the measurements don't: "the Bozostat inspires confidence with its clean design, no-nonsense user interface and beautiful build quality. The Dumbostat performed very well, but the flimsy, poorly fitted case and garish display will make you want to hide it from view so that it doesn't look out of place with your showpieces. The controls worked well enough, though, and it measures and performs so hide it under the cabinet and it can be your little secret."
In other words, write what you personally know and can demonstrate to be true.
Rick "unsubstantiated opinions are not transferrable" Denney
So, I shop among those that measured well simply because if they were careful about that, they were probably careful about other things, too. But mostly I shop to get the specific connectivity, features and look I need, and use the measurements to confirm that the engineering meets a sensible quality standard. Sometimes, I look at the brand, too, but never as a surrogate for performance, given that good performance data is available.
Those who understand measurements and replace a DAC with a 105 dB SINAD for one with 120 dB SINAD are just exploring a hobby, not expecting any audible difference or improvement.
When I read that one DAC has a bit more open bass than another (a statement I did read recently), and both measure transparently, then there are two possibilities: 1.) we aren't measuring the correct things, and 2.) there is some interaction between the DAC and my system that did not occur when the measurements were made. With DACs, I don't for a moment believe that either of these is relevant. There are other components where the second may be true.
But the first thing is a point of real contention and even religious debate. The measurement folks want to see some evidence that what a reviewer "clearly" hears is an effect that reviewer can reliably hear without prior knowledge of what's producing the sound. This seems a reasonable request, IF the reviewer is making value judgments intended to be instructive to others. The language used by reviewers is often intended to be instructive (recognizing the difference between motive and intent), such as "The Fryburger Electronics Bozostat provide more authority in the bass, and a bit more clarity in the upper treble, but it costs ten times as much, so you would expect that." That statement reveals a bias (biases are just expectations that color our impressions), but it also tells the well-heeled reader that the Bozostat is actually better than the cheaper device, when it might not be. The readers here generally don't expect the more expensive device to sound better, without demonstrating that it does either using measurements, or, in the absence of relevant measurements, by preference testing that controls for those biases.
So, if the argument that we aren't measuring the right things is true, then we should expect that those subjective impressions should be repeatable and unbiased if they are to serve as advice to the reader. If we can't do that, then we shouldn't offer advice at all based on hearing. We can say, "the Dumbostat didn't have the same authority in the bass as did the Bozostat, and we wondered why. Measurements show that as the output of the Dumbostat approaches maximum, we see a trough in the response between 80 and 150 Hz, and this explains the loss of authority. Our usual testing didn't duplicate this particular case." Or "We first thought the Bozostat was more authoritative in the bass, so we checked the levels, this time using a voltmeter, and found that the Bozostat put out 0.3 dB more output than the Dumbostat, and when we adjusted for that, the differences faded away." Even this is acceptable (because it is honest), "Even level-matched, there was something about the Bozostat that impressed me more, but I can't conduct a proper blind test and am probably swayed by its beautiful finish and what I know about Fryburger Electronics."
Of course, this sort of thing is always welcome in reviews, because it tells us stuff the measurements don't: "the Bozostat inspires confidence with its clean design, no-nonsense user interface and beautiful build quality. The Dumbostat performed very well, but the flimsy, poorly fitted case and garish display will make you want to hide it from view so that it doesn't look out of place with your showpieces. The controls worked well enough, though, and it measures and performs so hide it under the cabinet and it can be your little secret."
In other words, write what you personally know and can demonstrate to be true.
Rick "unsubstantiated opinions are not transferrable" Denney
Last edited:
