aiqbench analysis
This research introduces TVBC, a novel architecture that combines a lightweight neural controller with an interpretable cascade of IIR filters for real-time speech enhancement on hearing aids. By reducing parameters by roughly 29x while maintaining performance comparable to massive black-box models, the work demonstrates a viable path toward efficient, transparent AI deployment in resource-constrained assistive devices. The approach effectively bridges the gap between high-performance deep learning and the need for explainable, low-latency signal processing in consumer audio hardware.
An Interpretable, Controllable Time-Varying IIR Denoiser for On-Device Assistive Hearing
arXiv:2603.02794v4 Announce Type: replace-cross Abstract: We present TVBC (Time-Varying Biquad Cascade), an interpretable, low-latency speech enhancement model for real-time, on-device assistive hearing. A lightweight neural controller predicts, in real time, the coefficients of a differentiable cascade of 35 second-order IIR filters (biquads), so the model tracks non-stationary noise while keeping a fully interpretable processing chain: every spectral modification is an explicit, adjustable equalizer curve rather than an opaque `black-box' transform. Because the biquad cascade carries the signal processing, the controller can be made very small, driving the cascade with only 24k parameters at a 10.7ms algorithmic latency, within hearing-aid budgets, and running entirely on-device so that audio never leaves the device. We also expose the suppression-versus-preservation trade-off as an explicit control: it can be set during training through the loss weighting, and adjusted at inference, with no retraining, by mixing the noisy input with the denoised output. On hearing-aid metrics (HASPI/HASQI) the 24k model stays within about 0.02 of DFNet3 (2.3M parameters, almost two orders of magnitude larger) while using about 29X fewer multiply-accumulates, although larger black-box models still lead on reference metrics such as PESQ. We present TVBC as a proof of concept for a compact, interpretable, and controllable denoiser for on-device assistive hearing.
Abstract:We present TVBC (Time-Varying Biquad Cascade), an interpretable, low-latency speech enhancement model for real-time, on-device assistive hearing. A lightweight neural controller predicts, in real time, the coefficients of a differentiable cascade of 35 second-order IIR filters (biquads), so the model tracks non-stationary noise while keeping a fully interpretable processing chain: every spectral modification is an explicit, adjustable equalizer curve rather than an opaque `black-box' transform. Because the biquad cascade carries the signal processing, the controller can be made very small, driving the cascade with only 24k parameters at a 10.7ms algorithmic latency, within hearing-aid budgets, and running entirely on-device so that audio never leaves the device. We also expose the suppression-versus-preservation trade-off as an explicit control: it can be set during training through the loss weighting, and adjusted at inference, with no retraining, by mixing the noisy input with the denoised output. On hearing-aid metrics (HASPI/HASQI) the 24k model stays within about 0.02 of DFNet3 (2.3M parameters, almost two orders of magnitude larger) while using about 29X fewer multiply-accumulates, although larger black-box models still lead on reference metrics such as PESQ. We present TVBC as a proof of concept for a compact, interpretable, and controllable denoiser for on-device assistive hearing.
Sound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
arXiv:2603.02794 [cs.SD]
(or arXiv:2603.02794v4 [cs.SD] for this version)
https://doi.org/10.48550/arXiv.2603.02794 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 302 words
Related coverage
Game Engines Explained: How They Work, Why They Cost So Much, and Where Theyre Going Next
wccftech.com · Sep 30, 2026
2026.9: There's room on this bus
Home Assistant · Sep 2, 2026
Concrete Evaluations to Investigate the OpenAI Model That Hacked Hugging Face
AI Alignment Forum · Aug 3, 2026
I Did This Without Your Knowledge
LessWrong · Oct 17, 2026