Why use an obviously defective implemention of a dac to prove your point?
My point was that DACs could be incorrectly implemented, even in a device that is otherwise well designed and regarded.
Has anyone heard a difference between the correct implementation of the ESS dac and one with the ESS hump? Again, separate vs integrated and audibility.
I don't recall anyone on ASR claiming that the ESS hump was audible, I certainly wouldn't. So maybe not the best example to choose for this discussion. However, I chose this example because it is a well documented issue (on ASR) and has been investigated and solved by several manufacturers and ASR members and the issue is known to be related to the components connected to the DAC chip i.e. 'the implementation'.
Personally, I don't see the need for a stand-alone DAC if the DACs in your other equipment are performant. I was happy to replace my Topping x50 stack (D50s/A50s/P50) on my desk with a Topping EX5 (integrated DAC/HP & Pre-amp) because the EX5 solved a ground loop issue (balanced output), is neater both on and under my desk (single unit, no interconnects, single integrated PSU) and easier to use (bigger display and volume control).
From a measurement perspective, the EX5 is a 'downgrade' from the x50 stack (in the HP amp, rather than the DAC), but the benefits I mention above far outweigh any concerns I might have about this, to me, inaudible difference,