Aliasing Reduction in Neural Amp Modeling by Smoothing Activations

TL;DR

This paper explores how smoothing activation functions in neural networks can significantly reduce aliasing artifacts in digital emulations of analog audio hardware, improving audio quality without sacrificing accuracy.

Contribution

It introduces a novel aliasing metric (ASR) and demonstrates that smoother activation functions effectively lower aliasing in neural audio models.

Findings

Smoother activation functions lead to lower aliasing levels.

Aliasing reduction does not compromise modeling accuracy.

The ASR metric accurately measures aliasing in neural models.

Abstract

The increasing demand for high-quality digital emulations of analog audio hardware, such as vintage tube guitar amplifiers, led to numerous works on neural network-based black-box modeling, with deep learning architectures like WaveNet showing promising results. However, a key limitation in all of these models was the aliasing artifacts stemming from nonlinear activation functions in neural networks. In this paper, we investigated novel and modified activation functions aimed at mitigating aliasing within neural amplifier models. Supporting this, we introduced a novel metric, the Aliasing-to-Signal Ratio (ASR), which quantitatively assesses the level of aliasing with high accuracy. Measuring also the conventional Error-to-Signal Ratio (ESR), we conducted studies on a range of preexisting and modern activation functions with varying stretch factors. Our findings confirmed that activation functions with smoother curves tend to achieve lower ASR values, indicating a noticeable reduction in aliasing. Notably, this improvement in aliasing reduction was achievable without a substantial increase in ESR, demonstrating the potential for high modeling accuracy with reduced aliasing in neural amp models.