OK, here is the re-test of the Emotiva RMC-1 with the version 1.9 firmware.
Let's start with our dashboard:
Wow, that is a huge improvement. Those spurious tones that went on forever were adding to the noise floor, dragging SINAD way down (see inset for the original version 1.8 test). The new SINAD places the RMC-1 in a new performance bracket:
It also raises the status of RMC-1 to the top of the class among AVRs and Processors tested:
Note that I cleaned up the above graph, taking out the 2.7 volt output of the HTP-1 as that was not a fair comparison against likes of RMC-1 which produce 4 volt output.
You can see the improvement easier when we compare THD+N against output voltage:
IMD also improved:
There is still an issue with Reference stereo having higher noise level. I wonder if there is some conversion/dither pipeline it is going through. Its distortion though is lower than Direct mode however as reflected in the red graph.
Most remarkable difference came in Jitter test:
What looked like a hardware problem was a software stage that was modulating the input akin to what jitter does.
Conclusions
The big turn off for me in the original review was not just one thing, but so many things not working right. The new version cuts out the number of problematic tests way down, resulting in much improved performance. I can't yet recommend the processor due to lack of Dirac support as of this writing. If that were to become available, I would give it that.
FYI, I spoke to Emotiva folks about an hour. Much better people than I gave them credit for prior to that. They have a tough job to do, integrating a number of black boxes without much leverage on their many technology suppliers. I have been in their shoes (and the other side as technology provider), and feel the pain. I offered my help on their future development should they need it.