You never will be able to tell it from those numbers being -80dBFS, with a real-life music recordings. Wasting time, wasting efforts. Rather make the thing reliable. The more simple the better. Keep it cool, no local overheated islands.
There is science and experience. If amps are identical, operating linearly, etc. you cannot hear a difference. People hear differences some of which is sighted bias, but what if we assume that there are differences to “dac” sound signatures and amplifier sound “signatures”? After all, the original AES tube vs transistor comparison used soldered RC networks to really make sure each amplifier had its gain and frequency response matched.
A “we can agree” conclusion is that if there is a difference that is smaller than the threshold of lenient audibility but not the threshold of strict audibility, then we have to allow for the possibility than ABX differences can be challenging (because it is beyond the lenient audibility threshold) but might be audible over a period of time because the difference is not fully transparent.
This is the SINAD 90 DAC versus the SINAD 120 dac. If SINAD of 120 is needed across the entire chain for transparency, by definition, the SINAD 90 is not transparent. It’s very possible that the SINAD 90 DAC may be detectable in some cases since it’s not fully transparent. So from the standpoint of engineering, owning the car that goes faster than it is safe to travel on public roads standpoint, it’s not a complete waste.
We do know that the target SINAD is actually lower than 90, which you suggest may be closer to 80, based upon how “good” tube amps can sound.
I have actually sent Amir a SA600 from 1965 which has been restored with a minimum of resto-modding/hot rodding to test. It’s a close as possible to knowing what the SA600 would have sounded like and measured, had the test equipment existed back then. I am going to bet that it’s SINAD is high 60’s and I gotta say, it sounds pretty good. But it sort of makes sense to demand better performance if I am buying something in 2022. Otherwise I should just stick with my vintage gear.
Along those same lines, if your tweak cannot show up in a SINAD 140 APx555 (such as a cable), then that product is worthless. That’s one way we can separate “the possibility” versus the “impossibility”.
—-
I agree 100% that reliability is the first part. For all of the criticism that McIntosh gets, there are few pieces of solid state electronic gear from 1965 that still are useable today and command that kind of pricing.
I just think that reliability has been figured out — it’s just that consumers have fallen into the trap of subscriptions and disposable hardware and software consumption.
We also vote with our wallet. Just as I am more likely to buy a coffee from a friendly coffee shop versus a coffee shop whose owners may be harassing the neighboring cafes even if the actual baristas are great people, and even though the coffee beans at both coffee shops are sourced globally, I am also more likely to buy products from manufacturers with great customer service and community support as long as the products are also reliable and have good ergonomics.