Waveform have nothing to do with it... your element doesn't "fit" inside one cycle...
I do not agree.
Even when you switch one cycle of a sine wave on and off, periodic, the spectrum is that carrier available the full time as a constant carrier of constant amplitude and sidebands of the envelope, in this case of a block wave.
So 700 Hz carrier and sidebands at 350 Hz above and below etc, also folded for a series dots.
AND it is certainly true that you can have a sinewave carrier at 700 Hz with a triangle envelope, just by multiplying the sinewave with the triangle of the same period duration. That yields another smoother spectrum. 6 dB per octave, because a triangle is the integral of a block. It is also possible to switch part of a sinewave on and off. Makes no difference in mathematical handling of the subject.
So that is what I remember from the theory acquired for my ham license. Long time ago, but laws of nature don't change fast, do they?
I think AH8DX is a honest guy, he is very fanatic in exercising, and his approach was obviously specialising on speed not on points.
So he just starts a session in RUFZXP on the highest possible speed, and breaks is off when he did not copy a call above his recorded highest speed.
So a hundred sessions can be broken off, and then bingo. He doesn't claim according to the rules that he copies every run this high speed, once in a time, and due to the large database of about 50000 calls, his records promise to be more then good luck.
So he just hits the enter repeatedly when he copied ONE call, because his only purpose is to record the max speed obtained.
He had a wordpress website, where he published screenshots, showing his progress. A lot of calls with one error. And now he manages to copy at the highest speed sometimes a fully correct call. A short and easy one such as W1OO or a long one because he remembers that one as present in the database when he partially copies it.
Must be physiologically possible when you compare the bandwith with the bandwidth of the ear, and when you compare the complexity of speech with Morse code. The working of the brain is unknown. So when it is done it is possible.
Every guy is as far as I know starting at low speed an trying to increase that, at around 50 wpm meeting some ceiling, possibly because the brain detects high speed Morse as a spectrum and low speed as a time changing signal.
He was exercising the other way around.
Did you ever try to do a Koch course with 800 wpm? It is possible to distinguish two characters, try it and you know it, and then 3 etc.
I expect he did that.
I think it is not fair to deny that somebody managed to do that after years of exercising, on an unusual path.
So I surely congratulate Craigh,