All of that covered, dude. And I'm comparing these results to my i7/4790 as clearly mentioned.
Amcrest T2599 also included this time around. Same crappy stream in BI even at H.264, but not quite as pronounced. This is where I'm most disappointed since you also may recall that H.265 utilization is hardly the objective here. Quite the opposite for any number of reasons in my particular situation. The other cameras didn't provide an H.264 option, hence the frustration factor and inconclusive result there.
However...,
Both cameras work fine using H.265 in VLC, my Dahua NV4108, and look almost as smooth as BI in freaking tinyCam on an Android TV box.
After weeks of messing around with this and two new servers later,
Blue Iris is the only contributing constant left in the mix. Kinda hard to rationalize it being anything else at this point -- and I'm not the only one who has or is wrestling with this issue as you are well aware. I just don't happen to buy the 'gotta be something else' crap any more because I've tried everything else. Cables, switches, monitors, servers, HDDs, SSDs, memory..., even multiple versions of Blue Iris with two fresh Eval boxes running side by side right now into the same switch with the same cameras. Every configuration permutation each camera has to offer including the latest Amcrest firmware and multiple beta variations directly from Reolink senior level support as they continue to try improving their own protocol implementations, particularly for their older 4K hardware platforms.
The real bitch is that re-encoded MP4 exports from both cameras look beautiful even from the shitty raw h.265 .bvr clips. That's fine and dandy for those who are only concerned with that functionality. My use is primarily live surveillance via desktop mode with the occasional need or desire for a full screen solo'd display that doesn't jerk and jump at every I-frame or just freeze for several frames when a school bus enters the scene and just disappears by the time the stream returns.
With all the above considered, I still can't help wondering if there aren't
some type of lingering performance handicaps inherent in the Evaluation Versions.
An i5/9500 should handle what I'm throwing at it in its sleep. Hell, so should an i5/6500.
That's why I asked.