Acoustic noise is one of the major problems faced by the industrial workforce and causes physiological and psychological issues. The first step in reducing noise pollution is to identify and localize the noise source before studying its characteristics and mitigation. Since the early 1980s, the development of beamforming and near-field acoustic holography techniques have provided impetus to sound source localization. While the beamforming methods can only predict the direction of the source, acoustic holography methods can localize the source and predict the source strengths. Due to hardware challenges, both methods are restricted by the number of microphones used for sampling. Therefore, recently several equivalent source method (ESM) based acoustic holography methods have been proposed based on compressive sensing, such as one norm convex optimization (L1CVX) and iteratively reweighted least squares (IRLS). This study examines the performance of a proposed sparsity based method, multipath orthogonal matching pursuit (MPOMP), numerically and experimentally. MPOMP achieved better source localization and pressure field prediction than L1CVX and IRLS. Also, in terms of computational time, MPOMP is much more efficient than L1CVX and IRLS. This work has been published in Applied Acoustics. See Link: https://doi.org/10.1016/j.apacoust.2021.108501