This paper focuses on the classical additive noise signal restoration problem. The proposed time domain denoising method iteratively removes outliers. The proposed denoising filter incorporates a threshold operation to determine which sample values are outliers. This method is compared with wavelet soft/hard thresholding and empirical mode decomposition interval thresholding. The proposed method is shown to be a promising method to denoise signals where a frequency decomposition may not be a robust representation of the noise free signal. The paper provides a discussion on threshold selection and proposes future work to include applying noise estimation to automatically determine thresholds, possible stopping (convergence) criteria, and possible inclusion of an weighted mean computation to improve denoising performance.