In this paper, we investigate the spherical t-designs with large value of t for function approximation, construction of spherical framelets, and the important task of spherical signal processing. Based on the spherical framelet systems and the fast framelet transform algorithms, we propose an effective denoising scheme for spherical signal denoising that utilizes the nice properties of spherical t-designs with large t value. We provide numerical results of signal/image denoising on several data sets.