Supplementary Material of HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis

Details of the Model Architecture The detailed architecture of the generator and MPD is depicted in Figure 4. The configuration of three variants of the generator is listed in Table 5. In the ResBlock of V1 and V2, 2 convolution layers and 1 residual connection are stacked 3 times. In the Resblock of V3, 1 convolution layer and 1 residual connection are stacked 2 times. Therefore, V3 consists of a much smaller number of layers than V1 and V2. Periodic signal discrimination experiments We conducted additional experiments similar to training a discriminator using a simple dataset to verify the ability of MPD to discriminate periodic signals.