Automatic Audio Equalization with Semantic Embeddings

Moliner E, Välimäki V, Drosos K, Hämäläinen M. Automatic Audio Equalization with Semantic Embeddings. In: 2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio. Audio Engineering Society; 2025. Paper 7. Available from: https://aes.org/publications/elibrary-page/?id=22996

@inproceedings{Moliner2025_22996,
  author = {Moliner, Eloi and Välimäki, Vesa and Drosos, Konstantinos and Hämäläinen, Matti},
  title = {{Automatic Audio Equalization with Semantic Embeddings}},
  booktitle = {2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio},
  note = {Paper 7},
  year = {2025},
  month = sep,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=22996}
}

TY  - CPAPER
TI  - Automatic Audio Equalization with Semantic Embeddings
AU  - Moliner, Eloi
AU  - Välimäki, Vesa
AU  - Drosos, Konstantinos
AU  - Hämäläinen, Matti
T2  - 2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio
M1  - Paper 7
PY  - 2025
DA  - 2025/09/02
UR  - https://aes.org/publications/elibrary-page/?id=22996
PB  - Audio Engineering Society
LA  - en
AB  - This paper presents a data-driven approach to automatic blind equalization of audio by predicting log-mel spectral features and deriving an inverse filter. The method uses a deep neural network, where a pre-trained model provides semantic embeddings as a backbone, and only a lightweight head is trained. This design improves training-time efficiency and generalization. Trained on both music and speech, the model is robust to noise and reverberation. An objective evaluation confirms its effectiveness, whereas a subjective test shows a performance comparable to an oracle that uses true log-mel spectral features, demonstrating its potential for real-world applications.
ER  -