I'd chop it in half.
Keep all of the HDMIs + LAN on the stand-alone right unit.
Keep all of the digital audio I/Os + LAN + 2x USB3.1s, on the left stand-alone unit.
When the HDMI4.2 comes out, you can retire it but still have the digital audio (left) side functional.
AIO/AVP at this price range, ends up requiring expansive updates, in the (not too distant) future.
It sounds a bit extreme, but as it happens I completely agree.
When I first read Kal Rubinson's review of the highly publicised
Storm EVO, I thought what a waste of money. Kal used the EVO with the
Merging Hapi, and it started to make sense after reading that review. There are lots of good multi channel converters that perform like stereo DACs (rather than like AVPs) so if you separate decoding and conversion, you can improve the performance of the system.
After that I wondered about the other functions performed by an immersive audio system, and they started to stack up : Video Switching, Audio Extraction, Dolby Decoding, Atmos Rendering, Amplitude Equalisation, Phase Equalisation, Reconstruction Filter, Digital to Analogue Conversion, Preamp, Volume Control, Analogue Buffer, Power Amplifier.
There are individual HW & SW elements or layers that can perform each of those functions very efficiently. You can also adapt your system to include different architectures like networked audio distribution, or active speakers, etc. It
is convenient to do them all in an AVR, but then every layer is compromised. In practice you can integrate or separate any of those functions, but only if you have a processor with digital outputs.
I tried to pull that together into a layered audio architecture above, with the audio signals progressing downwards from the sources at the top.
Each layer corresponds to high-performing, high-value, off-the-shelf products that can be freely integrated and interchanged using industry-standard interfaces.
Companies like Storm, Trinnov, Meridian, Genelec and Neumann will try to do that for you, and give great results, but they're expensive.
I wanted an open, flexible, scalable, non-proprietary architecture, rather than something that trapped you to a particular manufacturer.
Nerds like me have obsessed with separating audio and video for a long time. Meridian did just that with the HD621 and UHD722, and any processor with digital outputs allows the multi-channel DA converter to be physically separated from video signals. If eARC is the only input, there are no digital video signals at all on that connection. That also provides a layer of video/audio separation, as it only sends digital audio signals into the audio processor / decoder.
Now here's the point.
A layered architecture lets you hold on to things that don't change much, like power amps, and stick with those for a long time.
For the faster-moving technologies like immersive audio, you can upgrade without having to dig deep every time.
This also applies to HDMI standards, which seem to make AVRs and AVPs obsolete very quickly.
This came up in
this thread and
this and
this - the miniDSP TIDE16 sticks to HDMI v2.0, but many people want HDMI 2.1 for video gaming.
One solution is to use the TV for source selection, and use a single eARC connection to the audio processor.
Alternatively, use an inexpensive HDMI switch with two, four or even eight inputs, and many of them also support eARC inputs and outputs.
Some of them are very sophisticated, with matrix switching, web GUI control from any device, and per-channel custom EDIDs.
Source selection and audio extraction sit at the top of the architecture, but they don't have to be integrated into an AVP.
Instead you could use an ARVUS, Apantac, Elytone or Hyperion processor just for decoding, and handle the HDMI 2.1 switching with one of these:
Finally, yes, when video standards move on again, as they always do, chuck out the cheap HDMI switch, and buy a newer and up to date one.
