I have a sketch of a 4x2 mux with PIP input, but I was kinda waiting on a promised newer DSP chip before I pulled the trigger on the design. That might be later this year. It wouldn't be much of a change to do 4x1 and use two of those, so long as the receiving recorder actually has about 88kHz bandwidth at 192kHz.
Very interesting. Possible 4X2 (44.1kHz > 96kHz) option? That would broaden the field of perspective recorders significantly. All the Sony M10s around here would become 4ch capable, the Tascam Dr2ds 8ch capable. Will keep my ears open.
Sure, obviously 4x2 is easier than 4x1. Probably worth checking to see if they actually have 44kHz bandwidth. I have an eval kit for a DSP chip that could be used for proof-of-concept, but its power consumption is rather appalling for field use.
The trick of course is demodulating, you'd need a plugin programmed to do that. I could code the basics of demodulation but the tricky part is writing a PLL that locks onto the carrier. I dunno how to code PLLs, but somebody out there does. I would do it the dummy's method of a manual tuning control. That works so long as the carrier tone doesn't modulate (it shoudn't). But for a lot of applications the typical crystal tolerance is probably good enough. It would certainly be better than just about any cassette tape. Or maybe there is a precoded PLL block that would do it for me (it's been a lot of years since I had time to play with coding VSTs).
This is a fun read, the spacing could be a bit tighter with digital filters, especially since you don't really need dual sidebands:https://en.wikipedia.org/wiki/FM_broadcasting#Quadraphonic_FM