The point is that a 16 bit 44.1 stream gets you an FR from the lowest range of human hearing to the highest. So there is no point in achieving an FR higher than 20 KHz. And actually most people can not even hear that high at any point in their lives. However, setting 20 KHz is a reasonable place to set the upper limit of human hearing, so going above that is definitely pointless.
The next issue is dynamic range. In recording systems this amounts to the difference between the loudest part of the program without distortion and the softest parts not descending into the noise floor. When it comes to noise floor you have to consider not only the signal to noise ratio of recording and reproducing chain, but also the background of the listening environment.
In digital systems unlike analog, distortion becomes the bigger problem in the quiet signals and not the loud. Having said that if you do set the recording level too high, then you risk running out of bits, in which case distortion suddenly becomes catastrophic. So when making digital recordings, you have to keep a close eye on the bitmeter which will be part of the recording screen, like I have in my WaveLab screen.
Now the problem becomes with quiet signals in that eventually you get to a place where the system has to choose between 1 and 0, so the error rate becomes 100%.
This problem is overcome by adding white noise to the program. This process is known as dithering. It is vital to get this correct. The lower the bit rate, the more dither is required. So this is why the bit rate does determine the achievable dynamic range. However a 16 bit recording will give you a 96 db dynamic range at least. This is more than enough for most program, especially taking into account the noise level of even the quietest domestic environments. So that is why the CD was well chosen. It gives you an FR out to 20 KHz reliably and actually with a bit of reserve, and you get a 96 db dynamic range 44.1 16 bit. That gives you a better FR and dynamic range than analog recording or reproducing systems.
Now, if you had understood what you referred to as those boring videos, you would have seen why recording engineers do actually record at higher rate than CD spec. This is because they want reserve, and want to eliminate as far as possible running out of bits in a live session. However in the remastering for public consumption there is no loss in issuing the recording for public consumption at 16 bit 44.1 KHz or certainly no higher than 48 KHz. Doing so is truly a waste of resources and as I said wantonly clogs the Internet. The Berlin Philharmonic Digital Concert Hall stream in 16 bit 44.1 lossless FLAC streams. Lossless algorithms use streams that save bandwidth but can be fully restored on the receiving end. FLAC and AFLAC are the most ubiquitous. I think the BPO are the world leaders in the streaming of digital AV program. Their audio quality is astounding.
This is all explained in what you referred to as those boring videos. These were produced by some of the greatest experts on digital audio.
Now I have been into digital audio from the early years. I started making digital broadcasts on location for the local public radio station in 1984.
I was one of the first, if not the first to broadcast live broadcasts on location over the Internet. So I have experience with this for 40 years now.
Below is a picture of a first generation CD player that I repaired playing a 20 KHz signal. You can see the right and left channels on my scope. Both are perfect. This was my first CD player a Revox. You can see that the 20 KHz waves are perfect. It had a problem that I had to correct, and that is why it was on the bench. This player was 14 bit 44.1 like quite a few first generation players were. No analog recording system can match that.
This Hi-Res craze is on the same spectrum as speaker wire and power cables that cost the same as exotic jewelry. It is all absolute bunk the lot of it.
