The basic idea with DSD is:
To (in the recording studio) store the signal like it comes from the 1-bit AD-Converter stage inside the chip, before it‘s recalculated (noise shaped and filtered) and packed into PCM format.
Then feed this “raw” signal (at the consumer’s place) back into the DA-Converter’s final stage only.
The idea sounds convincing, as it bypasses two recalculation inter-steps:
- 1-bit to PCM + Antialiasing Filter,
- and PCM to 1-bit + Reconstruction Filter.
What renders the format unusable is the fact that you cannot process this bit-stream in any way:
No editing, no multitrack mixing, no EQ, nothing, not even change the loudness, during production (or on the consumer side like inside ADI-2).
For the recording engineer this simply would mean working with both hands tied to the back.
It sacrifices all that can be done during production, even artistically, for using a theoretically “better” recording format.
This in the light of the fact that the recording format is by far the least significant factor in producing a good recording!
Remark:
DoP = DSD over PCM is just a transport format of DSD.
DSD data is packed into a PCM stream, to make it possible to handle DSD inside normal computer (PCM) architecture.