Registration for the INTERSPEECH 2025 special session on Non-intrusive Objective Speech Quality Assessment is now open! Use the link below to learn more about the challenge. tea-lab.qq.com/conferencingsp…

#tbt Be sure to check out our work with Tencent AI Lab on "Multi-channel multi-frame ADL-MVDR for Target Speech Separation" ieeexplore.ieee.org/document/96234…. A pre-print version is also available on the arXiv: arxiv.org/abs/2012.13442

Congratulations to Khandokar Md. Nayem Khandokar Nayem for passing is Ph.D. Dissertation Proposal today! He is now one step closer to being Dr. Nayem.

PROGRESS returns to #ICASSP2022, aiming to prepare students and junior researchers (especially women and underrepresented minority groups) to pursue academic careers in signal processing. Head to their website now to learn more and register today! bit.ly/3Ok26rM

Three of our group's papers are being presented at INTERSPEECH 2025. Be sure to watch the videos and ask any questions.

Check out this short article about Dr. Williamson's return to CSE OSUengineering. cse.osu.edu/news/2022/10/w…

As part of the Sound Demixing Challenge SDX 2023 Sound Demixing Challenge, we have released our code for the "cocktail fork problem", separating soundtracks into music/speech/sound fx. This includes pre-trained model weights and full training and evaluation pipelines. github.com/merlresearch/c…

A pre-print of our approach that uses acoustic parameter estimation to help dereverberation performance is available on the arXiv. Congratulations to Yuying and Yuchen: arxiv.org/abs/2302.04932

The pre-print version of Yuying Li and Yuchen Liu's research has been accepted by IEEE/ACM TASLP and it is now available: ieeexplore.ieee.org/document/10048…

The SPS Webinar Series continues on 13 March when Dr. Donald Williamson presents "Utilizing Full Signal Reconstruction and Leveraging Perception for Deep Learning-Based Noisy Speech Enhancement." Register now: ieee.webex.com/weblink/regist…

Be sure to virtually attend the IEEE Signal Processing Society webinar this afternoon, where Dr. Williamson will be discussing some of our work in the area of speech enhancement.

Check out Yuchen Liu's survey paper about privacy attacks and defenses for speech and audio. It's a good summary of what has been developed recently. arxiv.org/abs/2309.15087

We're seeing how human feedback can be used to directly optimize LLMs. For speech enhancement, we show how human feedback can indirectly be used to improve performance. Check out our new IEEE paper about this. ieeexplore.ieee.org/document/10301…

Privacy and security are becoming increasingly more important, especially with the advent of AI in our devices. Our recent survey paper summaries different types of attacks and defense measures related to audio/speech privacy. Take a look: arxiv.org/abs/2309.15087

Congratulations to Khandokar Nayem for successfully passing his Ph.D. dissertation defense today!

Check out a recent article about our work @OSU. news.osu.edu/ai-can-use-hum…

AI can use human perception to help tune out noisy audio news.osu.edu/ai-can-use-hum… OSUengineering CSE

Have you ever wondered if you value certain sound frequencies more than others, as they pertain to speech quality? If so, check out Junyi Fan's recent JASA journal publication. tinyurl.com/JASA-Frequency…

Congratulations to Ada Lamba for being a recipient of a CCBS Summer Graduate Research Award! cog.osu.edu/ccbs-summer-gr…

Congratulations to Sabid and the team on this great accomplishment! Winning the Best Paper Award ACM CUI is truly an honor! We're glad to be a part of this.