I downloaded and tried their updated SDK. It got rid of the major glitching. It looks like there's still some sort of static pop / glitch that occurs in the audio every 4-5 minutes, but I'm not 100% sure about that. It happened at about the same amount of time from the start of the recordings with both SACD players. However it was at a different spot in the track so I know it's not on the disc. I'm guessing it's a slight clock sync issue. The capture card is likely deriving it's own clock from the PC, not synchronizing to clock from the HDMI receiver, but I need to do a little more troubleshooting here. I need to split the audio off to my receiver and listen and see if the static / pop is audio through it. It was audible listening to the live mix on the PC through headphones, so it's not just in the recording.
Then, with the first disc I was messing around with I found the CD layer did not match the SACD mix. I had hoped to use a .cue sheet from the CD layer to cut up the recorded SACD content after lining up the start of each since the SACD recording will just be one big .wav file. I lined up the start of the first track to the ms, but by the end they were out of sync with each other by over 1 second. I need to do more testing since the CD layer was ripped, not recorded over HDMI so I can't conclusively rule out the HDMI interface as the cause of the track length mismatch, but it seems extremely unlikely.
I was trying to see whether the Oppo 24/88.2 or Sony 16/176 would be the better source. I converted down to 20/44.1kHz (I think from memory) using the same command line in SoX for both. Then I discovered the Oppo and the Sony don't output SACD at the same volume over HDMI. It seemed to be about 3dB different. I tried to normalize them to the same level, and then convert to .FLAC but they didn't quite match. The final file sourced from the Sony gave me a larger FLAC file. I don't know if that means it has more detail in the file or if bits are being wasted on artifacts or what. The amount of noise in SACD (DSD) -> PCM conversion above 20kHz is disheartening. You can see the frequency response roll off similar to a CD, and then there's this huge blob of high frequency noise inherent to SACD. Obviously there's a lot more of it from the Sony courtesy of it's 176.4kHz sampling rate vs the Oppo's 88.2kHz, but SoX uses a low pass filter when resampling so all that should get rolled off and averaged out courtesy of Nyquist in either case so I don't think it should cause the file size differences, but I could be wrong.
Anyhow, these various things sort of doused my enthusiasm for the project since it looks like I would have to make two recordings slightly offset from each other (like a 5 second delay) and use one to kill the pops / glitches from the other (assuming they occur on a regular periodic basis) which will be tedious to do. Then, I'd have to manually chop up the files (a pain). Added to that I'm still not sure which player I should use as the source.