Audio Dequantization for High Fidelity Audio Generation in Flow-based Neural Vocoder
This addresses a specific bottleneck for real-time speech generation systems, but it is incremental as it adapts an existing technique from image generation to audio.
The paper tackles the problem of degraded performance in flow-based neural vocoders due to training on discrete audio data by proposing audio dequantization methods, resulting in improved audio generation quality with better harmonic structure and fewer artifacts.
In recent works, a flow-based neural vocoder has shown significant improvement in real-time speech generation task. The sequence of invertible flow operations allows the model to convert samples from simple distribution to audio samples. However, training a continuous density model on discrete audio data can degrade model performance due to the topological difference between latent and actual distribution. To resolve this problem, we propose audio dequantization methods in flow-based neural vocoder for high fidelity audio generation. Data dequantization is a well-known method in image generation but has not yet been studied in the audio domain. For this reason, we implement various audio dequantization methods in flow-based neural vocoder and investigate the effect on the generated audio. We conduct various objective performance assessments and subjective evaluation to show that audio dequantization can improve audio generation quality. From our experiments, using audio dequantization produces waveform audio with better harmonic structure and fewer digital artifacts.