Functions§
- apply_
gain_ simd - SIMD-optimized gain application for a single gain value
- apply_
per_ channel_ gain_ simd - SIMD-optimized per-channel gain application
- blend_
simd - SIMD-optimized linear interpolation (blend) between two buffers
- complex_
mul_ add_ simd - SIMD-optimized complex multiply-accumulate
- complex_
mul_ add_ simd_ chunk - complex_
mul_ inplace_ simd - SIMD-optimized in-place complex multiplication
- complex_
mul_ simd - SIMD-optimized complex multiplication (without accumulation)
- compute_
covariance_ simd - SIMD-accelerated covariance calculation for ERB bands
- deinterleave_
stereo - Deinterleave stereo buffer into separate L/R channels
- enable_
ftz_ daz - Enable FTZ (Flush To Zero) and DAZ (Denormals Are Zero) CPU flags.
- fast_
inv_ sqrt - Fast inverse square root (1/sqrt(x)) using one Newton-Raphson iteration. ~0.1% relative error — suitable for normalizing all-pass filter magnitudes.
- find_
max_ abs_ simd - SIMD-optimized peak detection (maximum absolute value)
- flush_
denormals_ complex_ inplace - Flush denormals in complex buffer (applies to both real and imaginary parts)
- flush_
denormals_ inplace - Flush denormal numbers to zero to prevent CPU performance spikes and audio glitches.
- interleave_
stereo - Interleave separate L/R channels into stereo buffer
- scale_
add_ simd - SIMD-optimized multiply-accumulate for overlap-add synthesis (no window)
- scale_
add_ simd_ inplace - SIMD-optimized in-place scaling.
- window_
mul_ simd - SIMD-optimized windowed copy (for FFT input preparation)
- window_
mul_ simd_ inplace - SIMD-optimized in-place window multiplication.