It would be really cool if there was a way to put the extra DSP in the PCI card and make it backwards compatible with the original Multiface/Digiface boxes.  Can you say "upgrade path"? :-) (Can RME say "Cha-Ching"?)

Why would current Digiface owners upgrade a Digiface box - it doesn't have a "sound" of its own - so making an upgrade HDSP card with the Totalmix FX DSP would seem logical (if technically possible).  I guess Steadyclock might be a nice addition with a new Digiface revision - but most AD/DA's would be externally clocked in a Digiface system, anyway (no appreciable benefit with Steadyclock AFAIK)...

If RME does revamp the MF/DF - I hope the Digiface gets 4 banks of ADAT's plus simultaneous AES and SPDIF like the RayDat fryingpan .  A new Multiface should have no less than 2 banks of ADAT I/O - also with simultaneous SPDIF and AES.  High Power Headphone Outs are a must (maybe 2 pairs?).

Long live PCI!  I guess PCIe is alright, too rotfl (but currently too expensive).  I've yet to use a Firewire or USB interface on ANY of my DAW's - and proud of it!

Of course - but that's no reason to stop current development of PCI/PCIe audio interfaces.  Design for the hardware of today

I'm only suggesting that IF the Multiface/Digiface will be updated - it would be nice if the PCI/PCIe cards were backwards-compatible with current MF/DF boxes - and possibly with the addition of TotalmixFX DSP.  I believe the FPGA currently lives inside the MF/DF box, but maybe there is a way to suppliment with additional DSP on the PCI/PCIe card HeadScratch

Lightpeak seems functional for some applications - but I think an electrical connection is beneficial if for nothing more than its convenient ability to provide power and data in a single cable.  Lightpeak could get ratified to include an additional copper power feed when it leaps to 100GB/s - I'd be OK with that :-)  Otherwise I think USB3 has long legs for personal peripherals for the next decade and beyond - and is 100% backwards compatible.  USB3 might not be ideal for Audio - but I bet RME and other hacking geniuses will work it out (like RME did with USB2.0 only some 10 years after it was released - and it was thought to be completely useless for tons of low-latency I/O until the FFUC proved everyone wrong ).

Cheap MADI would be nice, too - but MADI is only like 15MB/s (125mpbs) max!  Fine for general Audio (I love it!) but too expensive.  I agree that we might as well future-proof out the ying-yang and use Lightpeak as a future MADI replacement, and get the benefit of mass-production pricing.  I'd bet the per-unit licensing on Lightpeak would end up cheaper than ADAT - and the components will be cheap.

MADI's jitter spec is pretty awful as-is, so I doubt Lightpeak would be much worse rotfl

Lightpeak seems to be "payload agnostic" - so you can basically use the Tx/Rx system for whatever data protocol you wish (MADI 2.0!).  Seems flexible for high-bandwidth requirements - those applications where USB3 won't suffice (USB3 will remain king on those personal gadgets that don't need more than 500MB/s or so - it should last a long long time for that task!)

Rock on