Audear dynamic speech-band instrument audio never leaves this device

Audear

A frequency window that breathes β€” slowly widening and narrowing around the speech band, the way the public SSP patent describes. Shape it, hear it, verify it.

f_low(t) f_high(t)
Not SSP

This is an experimental audio transformation inspired by publicly described acoustic-processing principles in US Patent US10029068B2. It is not the Safe and Sound Protocol, not SSP Core, not clinically equivalent to SSP, and not medically validated. The commercial implementation is proprietary and may differ in undisclosed ways.

01

Source

waiting for audio

What the track looks like before anything touches it β€” and anything that will shape how the filter should treat it.

02

Shape

waiting for audio

The window states come from the patent; everything else on these sliders is an engineering value, not a claimed SSP parameter. The charts redraw as you move them.

Window trajectory β€” f_low(t) & f_high(t)

f_lowf_high

Applied spectral gain over time

cut  boost

Filter cross-section

at the narrow state
Renders the strongest vocal passage with the current settings β€” switch between processed and original while it plays.
03

Render

waiting for audio

Processes the full track with the current settings, matches loudness to the source (makeup capped at +6 dB), and holds true peak at βˆ’1 dBTP. Output is 24-bit WAV.

0%

Processed track

Download WAV

A/B comparison

Download WAV
04

Verify

waiting for a render

Evidence, not assertion β€” the rendered file is re-measured, and the two spectrograms let you see the window breathing in the processed audio.

Original β€” spectrogram

Processed β€” spectrogram, window overlaid

Β§

Where every number comes from

Four categories, kept deliberately unblurred.

A Β· in the patent
  • 800–1500 Hz preferred narrow state
  • example windows to 500–3500 Hz
  • cycles begin & end narrow
  • stages ~10–30 s; cycles ~60–300 s
B Β· inferred
  • the running order of the window states
  • symmetric expand β†’ contract cycles
  • this track’s stage & cycle lengths
C Β· engineering choice
  • every gain, slope & transition value
  • log-frequency raised-cosine interpolation
  • STFT 4096/75% Hann; mask smoothing Ο„ 0.35 s
  • loudness match & βˆ’1 dBTP ceiling
D Β· unknown / proprietary
  • SSP’s actual filter realisation & slopes
  • emphasis vs. true band-limiting
  • per-track sequences and timing
  • any dynamics or psychoacoustic stages