you can't measure tones in the neural network.

Actually you can measure neural networks (just special ways, can read nice books on training and evaluation of them), and to your surprise no one still proven and showed that brain closely resemble any type of modern neural network (in really it is very far from even most complex ones).

But I talked not about this, but about acoustical and mechanical parts of ear.

The null box measures the wire only, unless I misunderstand.

Null box measure all it should measure. Measuring your perception is not the task of it. Same as it does not measure waves height at Honolulu.

BTW, no one is stopping ppl from using cheap gear

Well, capitalism is stopping you buying products with less margin in them, as banks do not provide loans to such manufacturers, you do not see them so much at expos or at any other media, they just do not have enough money to fund all this. And eventually most of them close or will be bough by firms who sell you more crappy and expensive stuff.

People make some serious money on YouTube who say that a Stradivarius violin is the same as a cheap violin. These are very popular. And, in fact, in double blind tests ppl could not tell them apart. They have been tested many times, and the results are consistent. I would rather have the Strad even if in the test people prefer the clone. Is it bad science or is experience worth something?

It is capitalism :-) Bad illness of humanity.

@caveport It has already been done. Mic emulators can pass a double blind test. Also, I did not say that all differences could be eliminated, I said they could get down to the level of difference as the null box, or to the point where you can't tell them apart.
Re: bias. I recently saw a post of a photo some guy took with his cell phone and he asked the group why the photo was better than his fancy camera. Instead of telling him it was because it was outdoors in really good light where the smaller sensor gave him a bit larger and sharper DOF, they all tried to convince him that the fancy camera had the better photo and that he was deluded (he wasn't, it didn't, the cell phone photo was tack sharp).
So, are we going to just use our cell phones? No. Well, maybe. Sometimes. (But eventually, probably.) Would we use the cell phone all the time if some box told us it was the same? Nope.
What about if we thought we could see a difference? If we were sure we could see a difference. But the box said uh-uh. Would we switch? Nope.
Remember, the video clearly says that no two wires test the same (best quote ever, IMHO).
T-Shirt material. Words to buy gear by.

@caveport Can't say I blame you.
Anyone else: you can see the difference in some typical cables here:

https://www.soundonsound.com/forum/viewtopic.php?f=15&t=58458

Q.

Apart from the differences in construction which are well documented on this forum, what are the advantages of choosing one over the other? Does one reject EMI better than RFI [...]?

R.

In simple terms, the screen is intended to minimise RFI, while starquad is intended to minimise EMI. Both can be used together (and often are).

In a standard balanced cable there are two signal cores which are twisted together, and this twisting is the critical element in rejecting interference from magnetic fields. By twisting the two signal wires together, both wires take turns being closer to and farther from the interference source, and so on average both wires will receive the same overall magnetic field strength and have the same induced currents which can then be cancelled out precisely at the differential receiver.

And in general, this arrangement works well... except when the interference source is very close indeed. In this special case, the twisting 'period' or 'lay-length' is usually too long to maintain the required average exposure to the same strength of interference. And so we end up with an imbalance in induced interference currents.

The solution presented in starquad cables is to add a second pair of signal wires twisted (very tightly) in with the first pair. So you have four signal wires (hence the 'quad' part of the name), arranged in a cross shape (the 'star'). The key to making this work is to wire opposite pairs together.

In this way, the local interference source will be strongest in the nearest wire, but weakest in that wire's sibling, while the two wires in the second pair will both be equidistant and both receive a field strength which is mid-way between the strong and weak levels of the other two wires. Consequently, the two pairs both end up with the same exposure and so the induced interference signals will be cancelled out.

Starquad cables are very good at rejecting very close sources of EM interference -- such as when laying mic cables in a cable duct with lighting cables -- but are essentially no better than standard balanced cables when the interference source is more than a few feet away. And the extra pair of signal wires increases the capacitance quite significantly, which can become an issue in very long cable runs or if you are trying to force AES3 digital signals down the cable.

Broadcasters tend to use star-quad cable by default, but in my own location recording rig I have only one cable drum filled with starquad cables, and three with standard balanced cable with double Reussen screens! And I use the latter far more often!

It's not bullshit. It is how shielding works in balanced cables. This stuff goes back to the earliest days of 600 ohm balanced audio. The only other comment is that people seem to be confusing shielding quality with signal wire quality and transfer. The original video is not about shielding, only signal quality. The premise is that two wires that measure the same will sound the same. Nothing else. He was also not talking about microphone cable but interconnect cables.

Sorry for the comment when I promised I would not! :-)

Shielding is different than quad. Many cables, even fairly pricey ones, have gaps or holes in the shielding. These have been tested, and Gotham and Canare test very high. You can see the test results on the Canare website. Gotham has the best shielding, which is why a lot of pros use it. Canare has had the quad cable from early on with a solid, flexible lightweight build which is why a lot of pros also use Canare.
You only have to make one film where you pick up hum, a walkie talkie, a radio channel etc before you ditch all your cable and get quad.