Struct AudioStack 

pub struct AudioStack {
    pub decoder: AudioVaeDecoder,
    pub out_rate: usize,
    /* private fields */
}

The whole audio tail: latent → spectrogram → 16 kHz → 48 kHz.

Fields

decoder: AudioVaeDecoder

out_rate: usize

Implementations

impl AudioStack

pub fn from_cmf(model: &Arc<CmfModel>) -> Result<AudioStack, String>

pub fn decode(&self, latent: &Grid, pool: Option<&Pool>) -> Sig

[8, T, 16] latent → stereo waveform at out_rate.

pub fn decode_from_mel(&self, mel: &Grid, pool: Option<&Pool>) -> Sig

The vocoder half on its own, from a [2, frames, mel] log-mel.