Hey
Let's talk about the ringing myth. On impulse response usually we see that the ringing level is pretty high like 10÷ of the impulse level and sometimes higher depending on the filter.
Also the rising time of the impulse vary between filters nos beeing the better.
1. Why don't we call the ringing of filter noise?
(And measures it it would be pretty high)
2. Does ringing of the filter mask some content in fast transient music as it's pretty high?
3. Can the fastness of nos raising time of impulse response explain natural tone perception? In case of fast transient complex music.
4. Does ringing occurs only in impulse response, or does it occurs also on real music on other conditions?
I keep thinking this guy is just playing us and happy to waste our time but I'll take a shot...
It is not a myth, it is a real output of a real filter hit with a real (albeit unrealistic) signal. Like many, many things it has been morphed into a distinguishing factor among products to differentiate among them even if the effect is irrelevant. Sort of like the bandwidth wars of years ago; "My amp is better than your amp because mine has 1 MHz of bandwidth and you only have 100 kHz!" The extra bandwidth usually did more harm than good, in both cases...
1. Because it is not noise. It is correlated to the filter response and not a random thing like noise. It could be considered distortion though is not nonlinear distortion people typically consider distortion. Ringing is usually considered separately from noise and classical distortion. And BTW, NOS DACs have anti-image filters at their outputs that exhibit the same sort of ringing (albeit different mechanisms cause it), and in fact the analog filter ringing is usually worse because without oversampling the transition band (filter bandwidth from passing signals to rejecting signals) is much narrower due to the lower sampling frequency.
2. People think of things like drum strikes, piano hammers, plucked strings and the like as "fast transients" even though they tend to be fairly low in bandwidth. Really fast transients require high frequency content, typically above what we can hear, and thus ringing in filters at frequencies beyond (above) audibility will not in any way affect "fast transient music". As has been said, the recording chain will eliminate those frequencies anyway. And of course oversampled converters have wider bandwidth to begin with (see below), sort of blasting the NOS being "faster" argument. Non Over Sampled = lower bandwidth.
3. An oversampled data converter has greater bandwidth and thus faster transient response than a NOS design. Not sure where you are going with this... Handling "transient complex music" requires reasonable bandwidth with high linearity (low distortion) and high resolution (wide dynamic range). Delta-sigma converters have repeatedly exhibited greater dynamic range and better linearity and thus should be more suited for complex musical passages. The issues they had with tones and limit cycles was solved twenty, thirty years ago and NOS DACs have plenty of their own issues. I tend to doubt (being from Missouri) you could tell them apart in a blind test unless one or the other had some sort of measurable issue (show me).
4. Frequency response and impulse response are frequency-domain and time-domain representations of the same transfer function. One can be converted to provide the other. If the "real music" has significant energy up in the filter's transition band, be it a NOS or delta-sigma output filter, then it will excite the filter and lead to ringing. In the real world such energy is small (very low in amplitude/volume) and filtered by the recording and production (mixing, mastering) process so I believe it is a non-issue. But I am not a mastering engineer nor speaker nor audio component designer blessed with ears to hear such things (and that of course is the usual audiophile's recourse; "you just aren't good enough to hear or measure it" -- balderdash.)