The post-merger (PM) phase of binary neutron star (BNS) mergers carries information about the merger remnant and dense-matter equation of state, but its observational prospects remain uncertain. We compare two ways of accessing this emission: the stochastic gravitational-wave background (SGWB) from the BNS population and direct PM-band detections from individually resolved mergers. Using a unified inspiral–merger–post-merger (IPM) waveform prescription and the same astrophysical population model for both calculations, we evaluate representative third-generation detector configurations involving ET, CE, and a dedicated kilohertz detector. In the baseline IPM calculation, the all-source broadband BNS SGWB exceeds the adopted stochastic threshold for the representative detector-network configurations considered across the local-rate interval. By contrast, the kilohertz PM-band SGWB remains below the adopted detection threshold. For individual events, over $R_0=7.6-250~{\rm Gpc}^{-3}\,{\rm yr}^{-1}$, the baseline IPM model predicts approximately $5-174$ PM-band candidates per year for the dedicated kilohertz detector, while the numerical-relativity cases considered here give approximately $0.07-7~{\rm yr}^{-1}$.