In my experience, "ad hoc" comparisons of disparate digital audio products such as the MPS-5, DAD AX24 and Berkeley Alpha DAC (and others) without consideration and optimization of the signal chain for each product are misleading and will lead to erroneous conclusions, not even to factor in the wild card variable of personal taste.
At a high level the digital audio chain consists of a source (physical media or electronic file), playback software, source-to-DAC interface (firewire, USB, AES/EBU, coax, toslink, i2S), DAC and pre-amp (not to mention cables, power supply, vibration isolation, etc.).
The MPS-5 is an "all in one" source, playback software, source-to-DAC interface and DAC. All of these variables have been optimzied inside the box by the designer. The DAD AX24 is a DAC only (for these discussions; not sure if it has a pre-amp function) and the Berkeley Alpha DAC is a DAC/Pre-Amp.
I don't know about the DAD AX24, but I can say from experience to my ears, the Berkeley Alpha DAC only sounds its best on CDRB using its filter 1.24 (1.16 is default), its own pre-amp section, and a high quality AES/EBU interface. For example, in my system if I feed the Berkeley an AES input directly from a Lynx AES16 without inserting a noise isolation and reclocker device like the Antelope DA the sound quality is significantly degraded (presumably due to the effects of EMI/RFI noise and jitter). And of course, comparing any of these products without listening to both CDRB and hi-rez seems incomplete.
I'm sure the MPS-5 and DAD AX24 are excellent products, as well as the MH ULN-8 which gets raves, but if one is to conduct a valid comparison then the entire signal chain must be optimized for each piece IMHO.