Could this be one reason tubes (and perhaps records) sound better?


This is not a new finding, rather it keeps cropping up in the hearing literature...
"White noise improves hearing":

https://www.sciencedaily.com/releases/2019/11/191112142926.htm
128x128lesdomes
I'm not sure tubes sound better - it is just a preference.  Noise is often used in imaging to improve resolution, but in sound it is likely to improve dynamics.  Distorted guitar sounds more dynamic than clean Jazz guitar at the same level.  Added noise is a form of distortion.
It might be that in striving to create noise-free systems, solid-state and digital audio went below subjectively optimal levels. That suggests experimenting with adding slight amounts of white noise to virtually noise-free solid-state systems to see how that affects aesthetic evaluations. This could be done using a continuously adjustable noise source to dial-in a most-pleasing level, which will of course vary with the listener as well as the source and other features of the system. 
Not only is this not a new finding in hearing research, its also not a new finding at all, in fact it has long been known that human perception is improved by the addition of noise.

Do a search for dither. Dither is noise added to a video signal that results in greater perceived resolution and clarity. Dither is standard practice in video mastering. Been that way a very long time.

More recently, and even less well known, Ted Denney III introduced Synergistic Research ECT, HFT, and PHT, a whole range of transducers that operate on the principle of dither or added noise. The noise introduced by these tiny little things is inaudible, yet somehow improves clarity, depth, extension and detail to an amazing extent.

The fact (and with this much research over this long a time frame, its a fact) that white noise improves perception is however nothing to do with tubes and records sounding better than transistors. Its nowhere near as simple as that. Rather what this does instead is shed light on the complex nature of what we so blithely dismiss as "noise".

I’ll give an example to illustrate my point. One night back when I was still playing CDs about as often as records we got to the end of one side when my wife said she was amazed how quiet it was. The record. Which I thought she meant, because it was one of my more silent ones. But she said no, she meant its quiet compared to the CD. Which certainly cannot be the case, for sure not by the normal meaning of noise being hiss and ticks and pops and all the other obvious noises. These are NOT what she meant!

I’m constantly fascinated with the reactions of normal, as in non-audiophile, folks, people who really just want to enjoy a really good musical experience. So I ask questions, and always carefully worded questions, very carefully worded so as to influence their answers as little as possible.

As long as that takes to explain its shorter than the exact questions and answers, which nobody can remember verbatim years later anyway. The gist of it is that she perceives the records as having less noise because to her the noise she hears when playing a CD is not hiss, the noise is the entire signal. CD, digital, solid state- these things measure great but sound bad because the whole damn signal they put out is noise! Or to put it slightly more politically correct yet technically accurate, the noise is so inextricably woven into the signal as to make it near impossible to differentiate it out. The signal is the noise!

Now, they are not all the same, these SS and digital devices. Some are better. Those we compliment by saying they are analog-like. Close to analog. This is after all the highest compliment you can pay a digital component. The highest compliment you can pay a solid state amp is it sounds tubey. Even better if it has tubey magic.

Nobody ever in the history of audio referred to transistor magic. I just did a search for "CD magic" came up with some magic tricks. Search "tubey magic" pages of hits, all in reference to the sound. Precisely nobody ever said solid state is magic. Measures well. That’s about it.

Last thing, the white noise thing, strikes me as a very close analog to bias. Tubes, tape deck heads, lots of stuff use bias current. Bias is used in all kinds of cases where the initial response from zero is poor, but the response from an elevated state is much better. So a very small bias current helps keep the part in an excited and more responsive state. Seems pretty obvious to me that is part of what is going on here. But then if you’ve read this far it should be obvious that what is obvious to me is actually kinda hard for most to get their minds around.

I’m telling ya, they just aren’t paying me anywhere near enough for this quality writing.
@millercarbon  I believe that we got used to particular system sound, different from reality.  We know that in concert the sound is different but at home we like what we hear.  When I replaced CDP with Benchmark DAC I had impression that sound is strange.  It sounded too clean - like something was missing.  One person stated that he can hear all instruments separately, but prefer to hear them as one (sound blob).  We need to learn to listen.