the ak4499ex is a multibit chip iirc, wondering if it'll be similar to the burr brown multibits
The AK4191EX-AK4499EXEQ are delta-sigma, with the distinction of being the first 7-bit chip operation I know about. The ESS9038Pro topped out on 6 bits, but Hyperstream was a pulse-width modulated operation, not delta-sigma.
The 4191-4499 is also not just a DAC chip, but a DAC ecosystem with three times PCB real estate, if you see what I mean. More precise resistors and capacitors aboard as a result can only mean a better single-bit decodes (X7) to alleviate dynamic element matching errors.
The PCM179X series was a hybrid of multibit and delta-sigma, MSBs decoded by the former and LSBs, the latter.
I think it is great that the option exists for a Class A amplification stage, but that may be more of a novelty or pride of ownership thing more than audio quality.
I am huge for Class A (and hence discrete operation) as it really means a more linear open-loop amplification without relying on negative feedback to close the loop (and the excess negative feedback necessary).
I pay far less attention to how something is biased between Class A or AB, and just focus on being blessed either way. I have a discrete circuit in my hands.
Which has so many audible improvements, and such better performance into low-impedance loads. I could detail why in detail if required - but Class A, Class AB bias audible effects are a bit of a misnomer. A red herring making people lose focus of the real point: a discrete circuit.
As an aside, my belief why we may not be hearing the difference switching between Class A and Class AB is:
1. IEM loads or power levels are so benign to begin with that Class AB operation is never falling out of Class A, and staying that way as such. Even in Class AB, bias levels are high enough it's staying Class A.
2. Failing to differentiate between Class A/Class AB appears much more in balanced use. The common-mode rejection could be hiding the crossover distortion enough, working with whatever feedback is then applied. I've heard the switch between A/AB all too evidently on a Cayin N6ii/E01, which alas was single-ended and did not benefit from common-mode rejection.
3. Some transducers just may not be transparent enough to show up a different switching between bias types.