What do you think about my typical experience shared in
my post #362 on my project thread?
This experience illustrate something that is very well known from the beginning : the presence of the noise-shaped quantisation noise of DSD signals in the tens of kHz region when signal is analysed in the frequency domain in a relatively narrow frequency spectrum. It is well known that this noise inherent to DSD (or sigma-delta signals) extends much above tens of kilohertz if one can observe it with a wider bandwidth spectrum analyser.
As you wrote yourself, you "
suspect" that that HF noise "
could be harmful" for tweeter or super-tweeters or for pets "
if" that noise would be reproduced as sound waves by super-tweeters, ie it's "
speculation", from your own words.
As DSJR just wrote, he is concerned by "
possible" inter-modulation products in the audible bandwidth.
That is exactly the point I have stressed : there is much speculation spread about the suspected effects of DSD HF quantisation noise. We all know this noise exists. Measuring this noise
ad nauseam will do nothing more to improve the knowledge of the existence of this noise, which is very well known and is a byproduct inherent to the way DSD (or sigma-delta modulated signal) works.
What is needed is proper experimentation to assess these
speculations. Do the remainder of this noise (after the mandatory low-pass filter to attenuate it, of course) disturb the downstream line level electronics and power amplifiers as is claimed ? Do the remainder of this noise have an effect, thermal or mechanical, on the speakers, especially tweeters? If this noise spectrum were to be reproduced by adequate speakers, would it be perceivable?
Although no specific experiment to make those assessments are known to me, we know the outcome of some tests designed to establish the audibility or lack thereof of different kinds of signal, PCM and DSD. The result of this tests* so far is that most likely it is impossible to distinguish by ear a PCM analogue to digital to analogue conversion chain from a DSD (1 bit delta-sigma) analogue to digital to analogue conversion chain. If telling apart PCM from DSD is not possible, to me it also means that DSD HF quantisation noise introduces no audible impairment in downstream electronics or speakers within the test set-up.
Thus, so far, the scarce clues we have about the topic of the effect of DSD HF quantisation noise tend to lead to the conclusion that this noise have no effect, at least no
audible effect.
I am absolutely bewildered that in 2024, after so much talks and speculation spread about the feared effect of DSD HF quantisation noise, no one in the world having access to DSD test signals and test equipments, notably in the Hi-Fi press, eg Mr Atkinson at
Stereophile or Mr Miller at
Hi-Fi News and Record Review or Mr Montanucci at
Audioreview, had the simple idea to measure a preamp or power amp by subjecting it with test signals generated from DSD sources (eg from Philips's Super Audio CD DAC Test Disc) and compare the results with those obtained with test signals generated by their usual means in order to see if the feared inter-modulation products in the audio band are there or not when replaying DSD test signals! Absolutely bewildered!
* The infamous
Meyer and Moran paper, the famous blind-test with Tonmeisters at Detmold from
Blech and Yang, the blind-test Mr Howard subjected himself that I cited
at the end of this post, dualazmak's family own experimental test results stated
here.
