Large discrepancies are found between observational estimates and theoretical predictions when exploring the characteristics of dust formed in the ejecta of a core-collapse supernovae. We revisit the scenario of dust production in typical supernova ejecta in the first 3000 days after explosion, with an improved understanding of the evolving physical conditions and the distribution of the clumps. The generic, nonuniform distribution of dust within the ejecta was determined and using that, the relevant opacities and fluxes were calculated. The dependence of the emerging fluxes on the viewing angle was estimated for an anisotropic, ellipsoidal geometry of the ejecta that imitate SN 1987A. We model the He core from the center to its outer edge as 450 stratified, clumpy, annular shells, uniquely identified by their distinct velocities and characterized by their variations in abundances, densities, and gas and dust temperatures. We find that the formation of dust starts between day 450 and day 550 post-explosion, and it continues until about day 2800, although the first 1600 days are the most productive. The total dust mass evolves from similar to 10(-5) M-? at day 500 to 10(-3) M-? at day 800, finally saturating at about 0.06 M-?. The masses of the O-rich dust (silicates, alumina) dominates the C-rich dust (amorphous carbon, silicon carbide) at all times; the formation of carbon dust is delayed beyond 2000 days post-explosion. We show that the opacities are largest between days 800 and 1600, and the characteristic spectral features of O-rich dust species are suppressed at those times. The fluxes emerging along the smallest axes of the ellipsoidal ejecta are found to be the most obscured, while a viewing angle between 16 to 21(?) with that axis appears to be in best agreement with the fluxes from SN 1987A at days 615 and 775.