My first encounter with K2 Mastering

I have purchased 3 K2HD CDs in the last several months; Miles Davis-Kind of Blue, Eagles-Hell Freezes Over, Cat Stevens-Greatest Hits.

I'll admit they are on the pricey side but they're the best CD versions of Kind of Blue and Hell Freezes Over I've ever heard.

Although the Cat Stevens is excellent I don't think it sounds as consistently fine as the other two.

I wonder if magic of K2 HD lies in better mastering-downsampling or it is whole recording-mastering-downsampling process. Such recordings like Kind of Blue cannot be re-recoded.

Kijanki,
I meant to answer your question but it slipped my mind! Anyway, better late than never....
In the original K2 & in the newer K2HD processes the key does lie in the A-D conversion from the master tapes/original digital recording and then in the subsequent down-conversion & truncation to 16 bits. Here is some info on the K2HD mastering process:
http://hificable.dk/k2-hd-mastering/

Kijanki,
here is a bit more technical info on JVC's K2 process:
http://en.wikipedia.org/wiki/Extended_Resolution_Compact_Disc

it seems that besides the A/D & down-conversion, the aspects of making a physical disk are also precision controlled.
This reminds me of several (yester year) home-use CD burners (such as the Yamaha) where we could burn longer pits. These burned CDs sounded much better than the commercial CDs but the penalty was that all the tracks not always fitted onto 1 CD.

Bombaywalla, Thank you for the links - very interesting. Many people assume that 16/44.1 has to be perfect because Nyquist says so. Nyquist criteria requires no signal above half of the sampling frequency or signal would fold (alias) to 0Hz and up. That is impossible to do perfectly and any close to brick wall filter introduces non-linear phase (uneven group delays) screwing up summing of harmonics. Also Nyquist criteria applies to continuous waves only. Reconstruction of signal becomes difficult with short duration time and high frequency (cymbals), since SINC functions (that suppose to correct limited frequency response caused by limited time/duration) sum very limited number of terms introducing interpolation error. That's where people hear more natural sound from Vinyl or SACD. On the other hand Master tapes have less of a problem being recorded at 192kHz, more than 9.6x upper limit of 20kHz compare to 2.2x in redbook CD. All this suggest that downsampling process is extremely important to CD sound quality. I already enjoy standard redbook CD the way it is (perhaps have less than golden ears) but quality of this K2 mastering, while still in 16/44.1 format, make me ecstatic.