Agree. The OPs first move here should be to get a processor with digital inputs. I know the ARC preamp is a well built piece and all, but there is nothing magical about it. And it does nothing that a good modern AVP can't do as well or better. Time to move on.If you have a modern AVP that has HDMI, I wouldn't suggest continuing to use the analog outputs on the Oppo. I know the converters in the Oppo are very good, but I doubt they'd be any better than what the AVP has. Worse a lot of AVPs do everything (a lot of the time even volume control) in the digital domain, so it's very likely you're just doing an extra unnecessary conversion by taking the analog input, converting it to digital for manipulation and then back to analog. Much better to feed the AVP any digital signal and let it handle it.
Use the Oppo to play discs, not as the sole DAC for your system.
Agree. The DSOTM Atmos disc is wonky. Pick another Atmos reference disc. And until the 5.1 Animals mix, James Guthrie's mixes were very meh.Don't use DSOTM to judge how Atmos played as 5.1 sounds on your system. There is something very wrong with the DSOTM disc encoding, as discussed elsewhere on QQ. I have a 5.0 system with an Atmos capable AVR. If I play either the Atmos test tones or the album in straight 5.0 it is as you describe, very "floaty" and no real localisation of anything. The test tones each play in 4 out of my 5 speakers. But if I engage Atmos in my Arcam AVR31 and have it render that to 5.0 (an entirely valid option) then it all sounds much better, the test tones appear where they should and the localisation in the album comes back.
So while this is a valid exercise for you, DSOTM is the wrong disc to use for testing. One solution you could consider is leave your speaker setup alone and get an Atmos capable decoder or AVR and then like me you can have an Atmos render down to your speaker layout.