Struct AudioSamples 

pub struct AudioSamples {
    pub samples: Vec<f32>,
    pub sample_rate: u32,
    pub channels: u16,
}

Raw audio samples produced by TTS synthesis.

Fields

samples: Vec<f32>

Raw PCM samples as f32 in range [-1.0, 1.0].

sample_rate: u32

Sample rate in Hz (e.g. 24000).

channels: u16

Number of audio channels (always 1 for mono).

Implementations

impl AudioSamples

pub fn new(samples: Vec<f32>, sample_rate: u32) -> Self

Create a new AudioSamples instance.

pub fn duration_secs(&self) -> f32

Duration of the audio in seconds.

pub fn len(&self) -> usize

Number of samples.

pub fn is_empty(&self) -> bool

Whether the audio is empty.

pub fn from_wav_bytes(bytes: &[u8]) -> Result<Self, TtsError>

Decode a WAV file from bytes.

Supports RIFF/WAVE PCM integer data (8/16/24/32-bit) and 32-bit float. Multi-channel audio is downmixed to mono to match the library’s output convention.

pub fn from_wav_file(path: impl AsRef<Path>) -> Result<Self, TtsError>

Decode a WAV file from disk.

pub fn from_audio_stream<R>(stream: R) -> Result<Self, TtsError>

where R: Read + Seek + Send + Sync + 'static,

Decode a WAV or MP3 stream into mono PCM samples.

The input format is auto-detected. WAV is decoded directly and MP3 is decoded with the built-in MP3 decoder.

pub fn from_audio_bytes(bytes: &[u8]) -> Result<Self, TtsError>

Decode a WAV or MP3 byte buffer into mono PCM samples.

pub fn from_audio_file(path: impl AsRef<Path>) -> Result<Self, TtsError>

Decode a WAV or MP3 file from disk.

pub fn denoise_audio_stream<R>( stream: R, options: DenoiseOptions, ) -> Result<Self, TtsError>

where R: Read + Seek + Send + Sync + 'static,

Decode a WAV or MP3 stream and apply speech-focused denoising.

pub fn denoise_audio_bytes( bytes: &[u8], options: DenoiseOptions, ) -> Result<Self, TtsError>

Decode a WAV or MP3 byte buffer and apply speech-focused denoising.

pub fn denoise_speech(&self, options: DenoiseOptions) -> Self

Apply speech-focused denoising to the current audio samples.

This is a classical DSP pass, not a learned source-separation model. It works best on mono spoken audio with steady background noise or music.

pub fn to_i16(&self) -> Vec<i16>

Convert samples to i16 PCM (for WAV output).

pub fn get_wav(&self) -> Vec<u8>

Encode the audio as a 16-bit PCM WAV and return the raw bytes.

The returned Vec<u8> contains a complete RIFF WAV file that can be written to disk, sent over the network, or played back directly.

pub fn save_wav(&self, path: impl AsRef<Path>) -> Result<()>

Save the audio to a WAV file (16-bit PCM).

Creates parent directories automatically.

Trait Implementations

impl Clone for AudioSamples

fn clone(&self) -> AudioSamples

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AudioSamples

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

Formats the value using the given formatter.