I'm far from being an expert in digital signal technology, but it seems to me that all of these smaller, specialist DAC makers - Chord, DCS, Meitner, Playback Designs etc - seem to use versions of the same box of tricks. Massive upsampling, mix and match interpolation filters and, frequently, conversion to DSD. I have heard all of these approaches at length, having owned all of the Chord DACs (yes, all of them), and the M Scaler no less than three times, and having spent a week with the Bartok and the Meitner MA3. I also experienced the Playback DSD conversion/upsampling approach with my Nagra Tube DAC, whose digital board is designed by Playback.
My subjective experience suggests all of these approaches are flawed. As always, there are no free lunches in audio or anywhere else, and using technology to produce one "improvement" usually involves a trade off somewhere else. As we know, linear filters measure the best but produce pre- and post-ringing artifacts. Minimum phase filters mitigate this to some extent, but introduce greater post-ringing and other artifacts. Upscaling using the Chord approach - which can be done using Chord's expensive box of tricks, the M Scaler, but can equally be done using any PC - does improve detail but introduces odd tonal artifacts. The same is true of Meitner's and Playback's approach of upscaling and converting to DSD, the latter tending to sound unnaturally smooth and glassy.
Why don't the big chipmakers use these approaches - AKM, ESS, Cirrus, Burr-Brown, Analog Devices etc? They have the resources to do so and I doubt it would be that much more expensive to produce a chip with these capabilities. I suspect they know that there are too many compromises involved and that the result would not sound natural, even though audiophiles are often looking for a distinctive - if unnatural - sound to justify the expense of the DACs.
In my case, having been through so many expensive DACs and remaining unsatisfied, I have "settled" with a Bricasti M3 which, like all of Bricasti's delta-sigma DACs, uses the rather elderly and conventional (and inexpensive) AD1955 chip, but focus their attention on power supplies and implementation. I find this approach sounds more natural than the bespoke FPGA based DACs. I should add that because of the quality of the implementation it also sounds much better than the awful Chinese DAC-of-the month popular on ASR such as the Topping, which I owned briefly.
