Audio enhancement is a signal processing method that improves the listening experience. Although most audio devices provide a variety of sound-enhancing effects, it is reported that very few people are active users of this feature. This lack of usability comes from insufficient sound improvement because of concerns about scene-rendering mismatch, which means that the processing applied to an unintended target may even damage the sound quality. The key solution to this problem is sound intelligence that provides an optimal sound effect with very low latency. The authors propose a real-time audio enhancement system based on a highly precise audio scene classifier using convolutional neural networks. The entire computation including convolutions is optimized for digital signal processing--level implementation, resulting in enhanced audio outputs for every audio frame.

Authors: Hwang, Inwoo; Kim, Kibeom; Kim, Sunmin
Affiliations: Sound Laboratory, Visual Display Division, Samsung Electronics, Suwon, South Korea; Sound Laboratory, Visual Display Division, Samsung Electronics, Suwon, South Korea; Sound Laboratory, Visual Display Division, Samsung Electronics, Suwon, South Korea(See document for exact affiliation information.)
JAES Volume 71 Issue 10 pp. 719-728; October 2023
Publication Date: October 10, 2023 Import into BibTeX
Permalink: https://www.aes.org/e-lib/browse.cfm?elib=22243

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 and would like to subscribe to the E-Library then Join the AES!

This paper costs $33 for non-members and is free for AES members and E-Library subscribers.

Learn more about the AES E-Library

E-Library Location: (CD JAES71)   /jaes71/10/pg719.pdf

DOI: https://doi.org/10.17743/jaes.2022.0093

Start a discussion about this report!