Struct PpgConverter 

pub struct PpgConverter { /* private fields */ }

PPG (Phonetic Posteriorgram) based voice converter.

Uses ASR-derived phonetic features as the content bottleneck:

• More explicit phonetic representation
• Better for cross-lingual conversion
• Requires pre-trained ASR model for PPG extraction

Implementations

impl PpgConverter

pub fn new(config: VoiceConversionConfig, num_phonemes: usize) -> PpgConverter

Create a new PPG-based converter

pub fn default_ppg() -> PpgConverter

Create with default PPG config

pub fn num_phonemes(&self) -> usize

Get number of phoneme classes

Trait Implementations

impl Clone for PpgConverter

fn clone(&self) -> PpgConverter

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for PpgConverter

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

Formats the value using the given formatter.

impl Default for PpgConverter

fn default() -> PpgConverter

Returns the “default value” for a type.

impl VoiceConverter for PpgConverter

fn config(&self) -> &VoiceConversionConfig

Get the configuration

fn convert( &self, source_audio: &[f32], _source_embedding: Option<&SpeakerEmbedding>, target_embedding: &SpeakerEmbedding, ) -> Result<ConversionResult, VoiceError>

Convert voice using speaker embeddings.

fn extract_content(&self, audio: &[f32]) -> Result<Vec<Vec<f32>>, VoiceError>

Extract content features from audio (speaker-independent).

fn synthesize( &self, content: &[Vec<f32>], _speaker: &SpeakerEmbedding, ) -> Result<Vec<f32>, VoiceError>

Synthesize audio from content features and speaker embedding.