This is the exact upgrade that I’m thinking of as well. I’m interested to hear your thoughts on the upgrade from the DX7s.
The DX7s is already so stupendously transparent that I'm not certain whether there are any audible advantages to this upgrade. I'd have to do quite a bit of A-B testing to see whether I can detect any differences, but that wasn't my motivation for upgrading so I haven't bothered to do that. For me, the D90 was attractive mostly for the sake of putting to rest a few minor neuroses:
1. SINAD jump from 108 -> 120 eliminates any doubt of transparency
2. No ESS hump = No OCD triggers
3. My Astell&Kern players use AKM chips, and I love the sound; if there's anything to AKM filters or something, that's another point of appeal
4. The D90 adds Bluetooth support, which I will probably never use, but I
could.
5. Comes with the remote, lol
Fluffy stuff aside, subjective listening has been perfection. How much of that comes from objective transparency and how much is simply a side effect of peace of mind gained from the above list doesn't matter to me. What does is that this setup, paired with the Stellias, is among the cleanest, most detailed, most enjoyable listening experiences I've had anywhere from any equipment.
Since this isn't Audio Pleasure Review, here's a fun experiment that this stack passed with flying colors:
At the very end of the last track on
Dark Side of the Moon ("Eclipse"), a bit of an orchestral version of "Ticket to Ride" by Beatles accidentally made it onto the recording. It's on the Tidal version, just after the quote at about 1:35. The most audible (and recognizable) part occurs about 1:55. The story behind this audio watermark is on Wikipedia:
https://en.wikipedia.org/wiki/Eclipse_(Pink_Floyd_song)
Anyway, that song snippet is extremely faint, and it may not be audible at all on less-resolving equipment. I found it by accident while listening to the Stellias on an A&K DAP, then looked it up and found the Wiki story above. Since then it has become a fun litmus test for various playback chains to get a sense of how much micro-detail comes through. With the D90/789 combo (and the volume turned up a bit), I can clearly hear the entire snippet in detail from just after the spoken bit until the fadeout at the end of the track.