This paper describes new concepts on adaptive, impulsive noise detection & reduction algorithms using the well known autoregressive signal modeling theory and LMS based adaptive techniques. The concept of simultaneous detection and elimination of impulsive noises in the same process is first examined, resulting in simple but computationally expensive LPC based algorithm. In order to reduce computational cost, the separation of the detection and elimination processes, is then investigated yielding on a novel modified LMS algorithm which identifies the positions of the impulsive noises with much better performance over simple LMS. The elimination of the impulse by using various techniques is then described. The proposed technques are examined in various real - life audio recordings.
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=6362
(354KB)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
