This is a 80min UHD doc with 10 channels of audio.

In the past I remember doing ama file export and getting 100% on all 14 cores.

This time my export was at about 50% cpu ussage.

I did a test just exporting only my video tracks and I am getting fast export speeds again at 100%.

It looks like ama exporting video and audio at once really slows it down.   I wonder if its paralell encoding video and audio rather than serially (one then the other).  Could it be the paralell encoding slowing this down?

I don't want to get into details about my storage, because I don't think its the bottle neck. 

At one point I even deleted all of my audio precomutes and re-rendered them to my nvme drive, and audio still slowed down the export.

The best solution I found was to mixdown the 10 channels of audio first, then export ama with media processor.

So in short, If i only ama export video I get 100% cpu usage.  If I include audio, it drops to 50% cpu usage.

I really want to know whats going on under the hood when media composer is exporting.

Any thoughts on this topic?

thanks