FreGrad: Lightweight and Fast Frequency-aware Diffusion Vocoder
This addresses the need for efficient audio generation in applications like speech synthesis, though it appears incremental by optimizing existing diffusion methods.
The paper tackles the problem of generating realistic audio with a lightweight and fast diffusion-based vocoder, achieving 3.7 times faster training, 2.2 times faster inference, and a model size of only 1.78M parameters without quality loss.
The goal of this paper is to generate realistic audio with a lightweight and fast diffusion-based vocoder, named FreGrad. Our framework consists of the following three key components: (1) We employ discrete wavelet transform that decomposes a complicated waveform into sub-band wavelets, which helps FreGrad to operate on a simple and concise feature space, (2) We design a frequency-aware dilated convolution that elevates frequency awareness, resulting in generating speech with accurate frequency information, and (3) We introduce a bag of tricks that boosts the generation quality of the proposed model. In our experiments, FreGrad achieves 3.7 times faster training time and 2.2 times faster inference speed compared to our baseline while reducing the model size by 0.6 times (only 1.78M parameters) without sacrificing the output quality. Audio samples are available at: https://mm.kaist.ac.kr/projects/FreGrad.