Human hearing upper limit for discernable sounds is about half or more than 48kHz standard for high quality audio, but those samples would generally be the peaks and lows of the audio signal. The problem is that interpolating that to a decent sized wave is generally done with large components from a bygone era, analogue-digital audio interfaces, in many cases with some noteworthy latencies, and Transducers alone do not do that.

Any digital to audible sound solution on the market is basically going to be the same sound quality until you get into dollar-store territory.