For anyone curious,
The only reliable option i found is to externalize the audio analysis and bring the data into Unreal via your favorite communication protocol… I like OSC… This way the systems stay reactive on frame the entire time. I am keyframing a custom event that sends out a message to start the audio track after my particles are in my favorite rest state. We can then hit play and let the magic happen. The screen capture works as expected, using shadowplay, im able to capture 2k - 60 fps no problem on the same machine. for 4k capture, we are using another machine with a cpature card.
This is not an uncommon method in RT video content, I was only hoping that there was a more internal solution. But until it is known, for those looking to stay away from render times, this is a solid approach.