[Darryl Ramm]

Yes. Each sample rate and buffer size you want to work at with hardware inserts. *and* use of any stupid settings like ignore errors or "minimize additional latency" changes stuff, but that should not be unexpected. And if you run with that on you may have more problems than HW insert latency.

Hopefully that is not be too many measurements. Which for me would be three buffer sizes, all at 96 kHz. YMMV. You can get fast at doing this. Small test session and zoom in on the timeline of the driving clicks and use loop record to "punch in" to record the delayed signal so you don't even need to go scrolling around and zooming out/in to find click pulses. Make a table print out the values, stick it on your computer and/or interface. Done.

Like I said it's a UI/UX embarrassment for Avid, and if anybody at Avid actually cared they could improve it greatly. But it's not rocket science to use. And I personally would not go to clunky old Digilink solely for this, if I was already on Digilink sure I'd like how it works.

I'm not sure why folks have so many problems getting this going. Some may be not appreciating dependency on IO buffer, sample rate and other settings. Some may be confused with plugin delay compensation and how that is all automatic. Or maybe there is a bug I've not run into? But one way to make this harder than it should be is to fark around on a working session with plugins and latency comp on trying to change insert latency numbers. Create a small test session, no plugins. Latency comp off (disables insert latency correction as well as plugin correction). Measure the insert latency as described in the reference guide (As BScout has pointed out here as well recently - measure in samples convert to ms (down to 1/100 ms accuracy) enter value in the insert delay box. Turn on latency comp (to enable the insert latency you just entered) and see if you need to tweak the number more. Enter the same latency for other equivalent inserts.