Changelog

0.6.2 (2026-03-21)

  • Added pre-built manylinux_aarch64 wheels (linux/arm64), enabling pip install pyminidsp inside Docker on Apple Silicon Macs.

0.6.1 (2026-03-21)

  • Upgraded miniDSP C library from v0.5.1 to v0.5.2 (bug fix affecting VAD).

0.6.0 (2026-03-21)

  • Upgraded miniDSP C library from v0.4.0 to v0.5.1.

  • New: VAD class — stateful voice activity detector with adaptive feature normalization and onset/hangover smoothing. Supports frame-by-frame processing (process_frame()) and batch processing (process()).

  • New: VAD_NUM_FEATURES constant (value 5).

0.5.0 (2026-03-18)

  • Upgraded miniDSP C library from v0.3.1 to v0.4.0.

  • New: MiniDSPError exception — C-level errors now raise Python exceptions instead of silently returning default values.

  • New: ERR_NULL_POINTER, ERR_INVALID_SIZE, ERR_INVALID_RANGE, ERR_ALLOC_FAILED error code constants.

0.4.0

  • Upgraded miniDSP C library from v0.1.0 to v0.3.1.

  • New: design_lowpass_fir() — Kaiser-windowed sinc lowpass FIR filter design.

  • New: lowpass_brickwall() — FFT-based ideal lowpass filter.

  • New: kaiser_window() — Kaiser window with configurable beta parameter.

  • New: resample() and resample_output_len() — polyphase sinc resampler.

  • New: bessel_i0() and sinc() math utility functions.

  • New: STEG_SPECTEXT constant for hybrid LSB + spectrogram text steganography.

  • Includes upstream bug fixes (division-by-zero in minidsp_core.c).

0.3.0

  • Published on PyPI with Linux x86-64 and macOS ARM64 wheels.

0.1.0 (unreleased)

  • Initial release.

  • Python bindings for the full miniDSP C library API via CFFI.

  • Signal generators: sine, noise, impulse, chirp, square, sawtooth, Shepard tone.

  • Spectral analysis: magnitude spectrum, PSD, phase spectrum, STFT, mel filterbank, mel energies, MFCCs.

  • Signal analysis: RMS, zero-crossing rate, autocorrelation, peak detection, F0 estimation.

  • Effects: delay/echo, tremolo, comb reverb.

  • FIR filtering: time-domain convolution, moving average, FIR filter, FFT overlap-add convolution.

  • Biquad IIR filters (LPF, HPF, BPF, notch, PEQ, low/high shelf).

  • DTMF tone generation and detection.

  • GCC-PHAT delay estimation.

  • Audio steganography (LSB and frequency-band methods).

  • Spectrogram text art synthesis.