Phase Repair for Time-Domain Convolutional Neural Networks in Music Super-Resolution

TL;DR

This paper identifies phase distortion as the cause of artifacts in time-domain CNNs for music super-resolution and introduces a phase repair method using a neural vocoder to significantly enhance perceptual audio quality.

Contribution

It is the first to demonstrate phase distortion causes artifacts in TD-CNNs and proposes a neural vocoder-based phase repair method that improves audio quality across different models.

Findings

Phase distortion causes artifacts in TD-CNN outputs.

Neural vocoder-based phase repair improves perceptual quality.

Method is effective across various TD-CNN architectures.

Abstract

Audio Super-Resolution (SR) is an important topic as low-resolution recordings are ubiquitous in daily life. In this paper, we focus on the music SR task, which is challenging due to the wide frequency response and dynamic range of music. Many models are designed in time domain to jointly process magnitude and phase of audio signals. However, prior works show that approaches using Time-Domain Convolutional Neural Network (TD-CNN) tend to produce annoying artifacts in their waveform outputs, and the cause of the artifacts is yet to be identified. To the best of our knowledge, this work is the first to demonstrate the artifacts in TD-CNNs are caused by the phase distortion via a subjective experiment. We further propose Time-Domain Phase Repair (TD-PR), which uses a neural vocoder pre-trained on the wide-band data to repair the phase components in the waveform outputs of TD-CNNs. Although the vocoder and TD-CNNs are independently trained, the proposed TD-PR obtained better mean opinion score, significantly improving the perceptual quality of TD-CNN baselines. Since the proposed TD-PR only repairs the phase components of the waveforms, the improved perceptual quality in turn indicates that phase distortion has been the cause of the annoying artifacts of TD-CNNs. Moreover, a single pretrained vocoder can be directly applied to arbitrary TD-CNNs without additional adaptation. Therefore, we apply TD-PR to three TD-CNNs that have different architecture and parameter amount. Consistent improvements are observed when TD-PR is applied to all three TD-CNN baselines. Audio samples are available on the demo page.