ESS has had the digital volume control presentation part as a
PDF on their website for several years.
I have never figured out why he is using a 16-bit DAC for an example. 24-bit DAC chips have been around for a long time. I've been using only 24-bit and have had at least 24-bit volume control for over 10 years.
He says on slide 5, "As a digital volume control operates on a fixed-width field (ie on that 16 bit number that the DAC receives) it creates noise because the DAC cannot make the fractional part of the number – And this is a large noise!"
So I think he is saying that the noise is due to quantisation errors. If so, then accessing the 32-bit internal data path of the DAC isn't the only solution. You could also access a 64-bit internal data path of the playback software. This will be even more accurate than the 32-bit DAC. In both cases, software or DAC, the final output to DAC chip is 24-bit (slide 14 says the Sabre chip is 24-bit externally). Also, he doesn't mention that you can basically eliminate the noise caused by quantisation errors by using dither.
The only noise a 64-bit digital volume control doesn't push down is the noise floor of the playback system.