911±¬ÁÏÍø

Gloria Dal Santo, a doctoral student in the 911±¬ÁÏÍø School of Electrical Engineering, received the Best Paper Award at the International Conference on Digital Audio Effects (DAFx) in Copenhagen, Denmark, on September 7, 2023. This annual conference brings together researchers from around the world working on audio processing for music, sound art, acoustics, and related applications.

Dal Santo co-authored the winning paper with her advisors, Dr. Karolina Prawda and Prof. Sebastian J. Schlecht, along with her supervisor, Prof. Vesa Välimäki, all from the Aalto Acoustics Lab. The paper is called 'Differentiable Feedback Delay Network for Colorless Reverberation' and it utilises machine learning to enhance the sound quality of an artificial reverberation method. This technique holds significant applications in music production and room acoustics enhancement.

The feedback delay network, a well-known method simulating room acoustic effects, is afflicted by coloured sound output that requires reduction. This paper proposes reducing colouration through parameter optimisation using a machine-learning technique known as stochastic gradient descent.

The results were subjectively evaluated at the Acoustics Lab with a formal listening test. The results demonstrate that the proposed approach diminishes perceptual colouration and enhances the sound quality of the reverberation effect.

Congratulations!

and listen to the sound examples related to this research .