This paper presents a framework for improving performance of the popular MUSIC algorithm using blind source separation. Whether it is used for direction of arrival (DOA) estimation or for spectral estimation, the performance of the MUSIC algorithm can be improved via BSS. This paper shows the impact of BSS on spectral estimation via MUSIC and examines scenarios where the sources are corrupted with additive Gaussian noise, differ significantly in strength, and scenarios where the sources are correlated. This paper also examines the impact of spatial smoothing prior to BSS on the resolution performance of correlated sources. BSS does not offer the spectral resolution attained by MUSIC but does enhance its performance in noise.