From 17 to 21 August 2025, several members of our "Communications Engineering" work group attended the Interspeech Conference in Rotterdam, Netherlands.
They were represented with three contributions:
- Synthesising Speech with Selected Perceptual Voice Qualities - A Case Study with Creaky Voice
- Towards Frame-level Quality Predictions of Synthetic Speech
- Spatio-Spectral Diarisation of Meetings by Combining TDOA-based Segmentation and Speaker Embedding-based Clustering
The 26th Interspeech Conference, the world's largest scientific conference dedicated exclusively to speech science and speech technology, had the theme "Fair and Inclusive Speech Science and Technology". It provided a platform to discuss and exchange new approaches and technologies in the field of human language diversity. The aim is to make speech science and technology fairer, more robust and more personalised.
Further information can be found here.
