Crate veils 

Veils: High-Performance STFT Library

Veils provides a high-performance Short-Time Fourier Transform (STFT) implementation compatible with the one-dimensional, real-valued subset of scipy.signal.ShortTimeFFT documented in the project README.

Features

• SciPy compatibility: Verified against SciPy 1.18.1
• High performance: Optimized Rust implementation using rustfft
• Multiple FFT modes: TwoSided, Centered, OneSided, OneSided2X
• Invertible STFT: Perfect reconstruction with canonical dual window
• Flexible windowing: Support for custom windows and dual windows

Quick Start

use veils::StandaloneSTFT;

// Create a Hann window
let window: Vec<f64> = (0..16)
    .map(|i| 0.5 * (1.0 - (2.0 * std::f64::consts::PI * i as f64 / 15.0).cos()))
    .collect();

// Create STFT instance
let stft = StandaloneSTFT::new(
    window,
    4,      // hop length
    1000.0, // sampling rate
    Some("onesided"),
    None,   // mfft (defaults to window length)
    None,   // dual_win (computed automatically)
    None,   // phase_shift
).unwrap();

// Generate test signal
let signal: Vec<f64> = (0..64)
    .map(|i| (2.0 * std::f64::consts::PI * 5.0 * i as f64 / 1000.0).sin())
    .collect();

// Compute STFT
let stft_result = stft.stft(&signal, None, None, None).unwrap();
println!("STFT shape: {} x {}", stft_result.len(), stft_result[0].len());

Structs

StandaloneSTFT

Short-Time Fourier Transform implementation

Enums

FftMode

FFT mode for STFT computation