Since my name got mentioned - there is no magic in Mojo, just thirty years of hard work researching DAC design using carefully controlled listening tests as a tool and making no assumptions. I create IP, and design silicon audio chips too; I have billion dollar semiconductor companies buying my services and patented IP; these companies do not sign multi million $ contracts on magic either.
What is often forgotten is that real science is about what we do not know, not what we do know; and in terms of understanding how the brain processes data from the ears to separate instruments out into separate entities science has very little understanding. My work has been centred upon looking at aberrations that interfere with the brains processing, and I have found some very interesting things - some errors are audible even when they are well below the threshold of audibility of the ears. I often think that the situation is akin to the ears having 16 bit resolution - but put a properly dithered 16 bit signal into a FFT and you can resolve signals well below the 16 bit limit. Thus, small signals that are below the resolution of the ear have important subjective consequences for the brains processing of the ear data.
One major aspect of what I do is centred upon the timing of transients. Mojo has some 500 times more processing power in the interpolation filter than conventional high performance DAC's. This is done because timing is an important perceptual cue - it is something I had studied and realised 30 years ago. You can read more about it here:
http://phys.org/news/2013-02-human-fourier-uncertainty-principle.html
This work is based on a paper published in Physical Review Letters. That physics journal does not believe in magic either.
Now the interpolation filter theory is very straightforward and proven mathematics. You can see more here:
https://en.wikipedia.org/wiki/Whittaker%E2%80%93Shannon_interpolation_formula
Now the maths is clear; if you use an infinite tap length infinitely oversampled filter with a sinc response you will perfectly recover the original bandwidth limited analogue signal. My work in this was the realisation thirty years ago that using conventional limited tap length filters would have important timing consequences in that the timing of transients would have an uncertainty; and these timing errors would have serious subjective consequences. That's why for the past 20 years I have been designing and improving my own algorithm to reduce these timing errors and designed extremely long tap length WTA filters. It is not magic just hard science and rigorous listening tests.
But perhaps the academic stuff is not your thing. So take a look at these measurements of Mojo:
This is FFT of Mojo at 2.5v RMS into a 300 ohm load. This is measured using my APX555. What is truly remarkable about this is the absence of any noise floor modulation - you can see this in the red trace with no signal. The noise floor at -175dB is the same whether the output is at 2.5v or nothing. No other DAC at any price (excepting other Chord DAC's) has this complete lack of measured noise floor modulation and its one important attribute as to why Mojo sounds so smooth and refined.
Getting this level of performance is not about magic; there are numerous ways for a DAC to create noise floor modulation, and I had to go to ridiculous lengths to achieve this level of performance.
To conclude; 500 pages in 8 weeks and dozens of awards and 5 star reviews is not magic - just thirty years of very advanced engineering.
Rob