We study the properties of the gravitational-wave (GW) emission between 10-5 and 50 Hz (which we refer to as low-frequency emission) from core-collapse supernovae, in the context of studying such signals in laser interferometric data as well as performing multimessenger astronomy. We pay particular attention to the GW linear memory, which is when the signal amplitude does not return to zero after the GW burst. Based on the long-term simulation of a core-collapse supernova of a solar-metallicity star with a zero-age main sequence mass of 15 solar masses, we discuss the spectral properties, the memory's dependence on observer position, and the polarization of low-frequency GWs from non- (or slowly) rotating core-collapse supernovae. We make recommendations on the angular spacing of the orientations needed to properly produce results that are averaged over multiple observer locations by investigating the angular dependence of the GW emission. We propose semianalytical models that quantify the relationship between the bulk motion of the supernova shock wave and the GW memory amplitude. We discuss how to extend neutrino-generated GW signals from numerical simulations that are terminated before the neutrino emission has subsided. We discuss how the premature halt of simulations and the nonzero amplitude of the GW memory can induce artifacts during the data analysis process. Lastly, we also investigate potential solutions and issues in the use of taperings for both ground- and space-based interferometers.