Lightweight Neural Prediction for MDCT-Based Perceptual Audio Coding

Abate N, Soule G, Bosi M. Lightweight Neural Prediction for MDCT-Based Perceptual Audio Coding. In: AES Show 2026 Nashville. Audio Engineering Society; 2026.pp. 50:1-10. Available from: https://aes.org/publications/elibrary-page/?id=23480

@inproceedings{Abate2026_23480,
  author = {Abate, Niccolo and Soule, Gabriel and Bosi, Marina},
  title = {{Lightweight Neural Prediction for MDCT-Based Perceptual Audio Coding}},
  booktitle = {AES Show 2026 Nashville},
  pages = {50:1-10},
  url = {https://aes.org/publications/elibrary-page/?id=23480},
  year = {2026},
  month = oct,
  publisher = {Audio Engineering Society},
  address = {Gaylord Opryland Resort}
}

TY  - CPAPER
TI  - Lightweight Neural Prediction for MDCT-Based Perceptual Audio Coding
AU  - Abate, Niccolo
AU  - Soule, Gabriel
AU  - Bosi, Marina
T2  - AES Show 2026 Nashville
SP  - 1
EP  - 10
M1  - Paper 50
PY  - 2026
DA  - 2026/10/30
UR  - https://aes.org/publications/elibrary-page/?id=23480
PB  - Audio Engineering Society
CY  - Gaylord Opryland Resort
LA  - en
AB  - Perceptual audio codecs often use predictive coding to further reduce signal redundancy. Rather than learning the entire codec pipeline end-to-end, we insert a small neural network as a drop-in, deterministic predictor in an otherwise conventional codec: the network estimates the next Modified Discrete Cosine Transform (MDCT) coefficient block from a short context of reconstructed blocks, and the encoder transmits the quantized residual. We compare Multilayer Perceptron (MLP) and Convolutional Neural Network (CNN) predictors against a nonpredictive baseline and a per-line Normalized Least Mean Squares (NLMS) linear predictor (similar to the MPEG-2 AAC Main Profile predictor). The CNN consistently outperforms the MLP, the linear predictor, and the baseline in signal-to-noise ratio (SNR), noise-to-mask ratio (NMR), and MUSHRA evaluation scores. Using a 512-sample analysis window (w = 512), our neural predictors recover, and the CNN exceeds, the coded quality obtained by the non-predictive baseline after quadrupling the analysis-window length to w = 2048, at substantially lower algorithmic delay. The CNN uses only 0.13M parameters (roughly 20 fewer than the MLP) and has a decoder real-time factor (RTF) of 0.10 on a consumer CPU.
ER  -