MelGAN
Introduced by Kumar et al. in MelGAN: Generative Adversarial Networks for Conditional Waveform SynthesisMelGAN is a non-autoregressive feed-forward convolutional architecture to perform audio waveform generation in a GAN setup. The architecture is a fully convolutional feed-forward network with mel-spectrogram $s$ as input and raw waveform $x$ as output. Since the mel-spectrogram is at a 256× lower temporal resolution, the authors use a stack of transposed convolutional layers to upsample the input sequence. Each transposed convolutional layer is followed by a stack of residual blocks with dilated convolutions. Unlike traditional GANs, the MelGAN generator does not use a global noise vector as input.
To deal with 'checkerboard artifacts' in audio, instead of using PhaseShuffle, MelGAN uses kernel-size as a multiple of stride.
Weight normalization is used for normalization. A window-based discriminator, similar to a PatchGAN is used for the discriminator.
Source: MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis
Latest Papers
| PAPER | DATE |
|---|---|
|
Improve GAN-based Neural Vocoder using Pointwise Relativistic LeastSquare GAN
• • • |
2021-03-26 |
|
Universal MelGAN: A Robust Neural Vocoder for High-Fidelity Waveform Generation in Multiple Domains
• • |
2020-11-19 |
|
StyleMelGAN: An Efficient High-Fidelity Adversarial Vocoder with Temporal Adaptive Normalization
• • |
2020-11-03 |
|
SpeedySpeech: Efficient Neural Speech Synthesis
• |
2020-08-09 |
|
VocGAN: A High-Fidelity Real-time Vocoder with a Hierarchically-nested Adversarial Network
• • • • |
2020-07-30 |
|
Adversarial representation learning for private speech generation
• • • • |
2020-06-16 |
|
SE-MelGAN -- Speaker Agnostic Rapid Speech Enhancement
• |
2020-06-13 |
|
MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis
• • • • • • • • |
2019-10-08 |
Tasks
| TASK | PAPERS | SHARE |
|---|---|---|
| Speech Synthesis | 4 | 80.00% |
| Speech Enhancement | 1 | 20.00% |
Average Pooling
Convolution
GAN Feature Matching
GAN Hinge Loss
Leaky ReLU
MelGAN Residual Block
Tanh Activation
Weight Normalization
Window-based Discriminator