CoolFace
Modelpublic

aoiandroid/IndexTTS-Rust

sourceHugging Facemitupdated 4mo agoView on Hugging Face
0likes1downloads
lib.rs60 linesDownload Raw Back to src
1//! IndexTTS - High-performance Text-to-Speech Engine in Pure Rust2//!3//! This is a Rust implementation of the IndexTTS system, providing4//! zero-shot multi-lingual text-to-speech synthesis with emotion control.5//!6//! # Features7//! - High-performance audio processing with SIMD optimizations8//! - Multi-language support (Chinese, English, mixed)9//! - Emotion control via vectors or text10//! - Speaker voice cloning from reference audio11//! - Efficient memory usage with zero-copy operations12//!13//! # Example14//! ```no_run15//! use indextts::{IndexTTS, Config};16//! use indextts::pipeline::SynthesisOptions;17//!18//! let config = Config::load("config.yaml").unwrap();19//! let tts = IndexTTS::new(config).unwrap();20//!21//! let options = SynthesisOptions::default();22//! tts.synthesize("Hello world", "speaker.wav", &options).unwrap();23//! ```24 25pub mod audio;26pub mod config;27pub mod error;28pub mod model;29pub mod pipeline;30pub mod quality;31pub mod text;32pub mod vocoder;33 34pub use config::Config;35pub use error::{Error, Result};36pub use pipeline::IndexTTS;37 38// Re-export Marine quality validation39pub use quality::{40    ComfortLevel, ConversationAffectSummary, MarineProsodyConditioner, MarineProsodyVector,41};42 43/// Library version44pub const VERSION: &str = env!("CARGO_PKG_VERSION");45 46/// Default sample rate for audio processing47pub const SAMPLE_RATE: u32 = 22050;48 49/// Default number of mel filterbank channels50pub const N_MELS: usize = 80;51 52/// Default FFT size53pub const N_FFT: usize = 1024;54 55/// Default hop length for STFT56pub const HOP_LENGTH: usize = 256;57 58/// Default window size59pub const WIN_LENGTH: usize = 1024;60