But really, your claim is disingenuous on its face. All my Artifact Audibility test attempts to show is at what level below music a nasty sounding artifact can be heard. It has nothing to do with digital always adding disturbing artifacts, or whatever it is you believe. If that were even true, those artifacts would show up when recording and playing back a pure sine wave, or some other known source. So already your claim is easily proven false using basic audio test equipment. I should have posted this link earlier:
What your Artifact test does is demonstrate the masking principle. It does not get at the issue of how artifacts introduce coloration by adding brightness. It does very effectively show how a sound that's 40 db down (or even less) might not be audible when louder sounds are present.
I think I have a Nakamichi machine still laying around somewhere. I would need refurbishing though. Most of the rubber bits have perished.
Here is the problem: When you are demonstrating an artifact like the effect of aliasing (which should be inherent in your digital gear) you can't use the same gear to show 'with' and 'without'; that won't work since it would always be 'with'.
Distortion (including aliasing or inharmonic distortion) is a bit insidious as there is this assumption that the distortion artifacts are masked or buried in the noise. Sometimes this is true (certain harmonics can mask the presence of other harmonics) but the ear has a way of allowing us to pick out very low distortion artifacts not by detecting them as themselves, but by converting them to tonality. In a way they ride on top of the signal rather than at the bottom (metaphorically speaking). So what we would be looking for in a test is to see which is brighter (without actual FR errors being introduced).
This, Ethan, is why I've been wondering if you don't 'get it'; your responses to me have all been about masking, which is something other, and something I don't contest (masking being an important rule of human hearing). Perhaps it might help to know that in the last ten years, Dr. Herbert Melcher has shown that the human ear/brain system has tipping points. An example of that is if the playback has insufficient speed (risetime) the brain has a tipping point where the music processing is transferred from the limbic centers to the cerebral cortex.
Another example of a tipping point is where the brain will favor distortion as tonality over actual FR errors! Now both of these things were not known back in the 1970s and are examples of 'now that we know that, we can do proper engineering to take them into account.'
So in a case where we want to demonstrate aliasing, or how the presence of trace amounts of higher ordered harmonic distortion can be detected without hearing the actual tones, its likely that the test would be constructed on a platform that first is immune to the problem itself, so that the problem can be demonstrated. Then once that is done what we are looking for is tonal differences between the two examples 'with' and 'without'.
You can see right away the difficulty! To demonstrate aliasing, it can't be done on a digital system since aliasing is inherent; to demonstrate higher ordered harmonics you can't have them inherent in the system which probably means no transistors in the signal path.
Atmasphere help me understand the artifacts you talk about. Is this related to the fact that human hearing is not flat but more sensitive at certain frequencies as shown here or something else?
It does matter that our ears are tuned to bird song frequencies (IOW Fletcher-Munson); this is the result of evolution as birds are the early warning system of the presence of a predator in the environment. You can regard this extra sensitivity as a complication!
But its more than that- the ear/brain system converts distortion to tonality. This is why one amp can sound bright while another does not, even though while on the bench both amps measure flat. The difference is that the former has more higher-ordered harmonic distortion and so sounds bright despite being perfectly flat in bandwidth.
Seventh harmonic causing a metallic quality: I've heard that several places, but in this case the easiest to remember is John Curl (one of the top solid state designers alive today).
https://www.youtube.com/watch?v=fZwS-oyqc3w
goto about 36:35