Struct HifiGanVocoder 

pub struct HifiGanVocoder { /* private fields */ }

HiFi-GAN vocoder.

High-fidelity generative adversarial network for mel-to-audio.

Note

This is a stub - requires model weights.

Implementations

impl HifiGanVocoder

pub fn new(sample_rate: u32, n_mels: usize) -> HifiGanVocoder

Create new HiFi-GAN vocoder

pub fn default_config() -> HifiGanVocoder

Create with default settings (22050Hz, 80 mels)

Trait Implementations

impl Debug for HifiGanVocoder

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Vocoder for HifiGanVocoder

fn vocalize(&self, mel: &[Vec<f32>]) -> Result<Vec<f32>, SpeechError>

Convert mel spectrogram to audio.

fn sample_rate(&self) -> u32

Get output sample rate

fn n_mels(&self) -> usize

Get expected number of mel channels