Hi @stavrose
Sorry for my delayed reply, I have spent a fascinating day at the exhibition ‘The Music is Black: A British Story’ in the newly opened V&A East in London. I was there three hours - but needed much more time to take it all in!
The reason for my question to you was that I wanted to understand the level of supposed ‘intelligence’ in the various AI tools and their capability in responding to questions about audio reproduction in the home environment.
For example; depending on the AI tool used (and also whether one pays for ‘better’ answers), a query text submitted by one user may produce a different result to an identical query text submitted by another user. My understanding is that this variation relates to the history of previous queries of the particular user.
The conclusion I have drawn from this one sample example (of asking for the results of your AI query) is that AI tools are plain dumb. The AI tools are most likely aggregating much of the potential nonsense that gets repeated on social media and hence provide no meaningful or really informative (c.f. technically accurate) coverage for the question posed.
PS: I appreciate that perhaps you and I take different perspectives on this?
I am after all a ‘Mad Measurement Monk’ (reference).
Nevertheless. I feel that if one wishes to explore the more likely reasons for perceived differences in sound emerging over several days for a (physically) unchanged arrangement of electrical equipment, the areas to look at are:
-
The specifications of equipment involved (and any variations with temperature of the equipment and sound transmission medium - i.e. air).
-
The relative humidity within the listening environment (i.e. the air).
There is previous thread exploring some of this (Physics and HiFi).
FYI ‘firmware’ (as an explanation) would not be at the top of my list.
EDIT NOTE: The following paragraph has been amended to more accurately describe the situation pertaining to the graph and comment below. It is perhaps unfortunate that the technical reviewer for HiFi News did not report the specific software/firmware version number when testing and reporting.
Nevertheless, it is the case that sometimes (when in early development or perhaps due to small errors in software coding) digital audio equipment may not handle some aspects of digital input reliably and consistently separate to the above considerations. The Naim NSS 333 is one such example and at its launch in 2023 the HiFi News magazine tested the product including the jitter performance of digital inputs of USB and S/PDIF.
The published results and technical comment in HiFiNews were as follows:-
"in this instance there was a marked difference in low-rate jitter between USB (10psec) and S/PDIF (3345psec) inputs [see Graph 2]. Naim will be issuing a firmware update to remedy this."
High res. zoom jitter spectra with 48kHz/24-bit data (S/PDIF, red with mkrs; USB-A, black)
Copyright: HiFiNews (December 2023).
My apologies for this lengthy reply - it may stimulate further thoughts and discussion - but as I don’t own such equpment I have no current means of veryfying the HiFiNews Lab Report from 2023?
.
