THIRDEAR 1.4.0   auditory distortion product instrument   databurn

Windows, 64-bit. Two formats, pick the one your host loads:

  thirdear.clap   copy to  C:\Program Files\Common Files\CLAP\
  thirdear.vst3   copy to  C:\Program Files\Common Files\VST3\

Rescan plug-ins in the host. Presets you save land in
%APPDATA%\databurn\thirdear\presets\ as one file each.

Quick start:
  1. Put ThirdEar on an instrument track with nothing after it. Use
     headphones, and turn your system volume down first.
  2. Load the preset "Tune  Passion Chorale" and hold one key. The tune
     plays from the key you hold.
  3. Raise the volume slowly until you hear a low melody under the high
     tones, and stop there: play at the lowest level at which the
     phantom is audible.
  4. The melody is made by your own cochlea. ThirdEar outputs nothing at
     its pitch, and the meter under the note says "register empty".
     A/B switches between the note your ear makes and B, a real tone at
     the same pitch. CANCEL opens the bench, where you cancel the
     phantom with a real tone of your own and can record what you find.

Put nothing that distorts after ThirdEar: a saturator, clipper, limiter
or clipping master bus makes a real tone at the phantom's pitch.

New in 1.4.0:
  Safety: no single action raises Level. Loading a preset, the reset in
  the settings menu and a double-click on LEVEL keep the lower of the
  two levels. Norm loud's boost fades out above 4 kHz (none from 8 kHz),
  so high tones are not made louder, and a CDT pair whose lower tone
  would fall under 20 Hz is not played.
  The bench keeps its answers. RECORD writes each null to readings.csv
  beside the presets. GROWTH -6 dB and -3 dB measure how your phantom
  grows with level. FIND LEVEL finds the lowest level at which you hear
  it. COPY STIMULUS puts the sounding tones and the model's predictions
  on the clipboard as CSV. HEADPHONES names the pair the readings were
  made on.
  A/B's second setting is now called "real tone".
  In CDT, pattern steps are legato: the pair glides to the next note,
  and a repeated note holds. With one note at a time, letting go of the
  sounding key brings back the most recent key still held.
  The meter sizes its frame from the nearest tone and says "unresolved"
  when a tone is too close to tell apart. The cochlea view draws the
  emission back out to the ear canal.
  Rest the mouse on a control and the footer explains it in one line.
  Esc closes the About card, the bench and any open list. The theme is
  remembered between sessions.
  Five new research presets: Tartini's third tone, Kendall's harmonics
  5 to 8, an umngqokolo 9:8 pair, the quintina, and high pairs after
  Estelle Schorpp. Each names the key to play.
  In the host, Rate is now "Pattern rate" and Phase is "Start phase".
  Saved automation still works.
  The manual is rewritten as an operator's manual: setup, the panel,
  the controls with their defaults, the presets and the bench, before
  the science.

New in 1.3.6:
  The sound is unchanged. The manual and the descriptions are reworded
  to say plainly what happens: the note you hear is a distortion
  product your ear makes from the tones, and the plugin outputs nothing
  at its frequency. The manual's cover names the release it belongs to.

New in 1.3.5, fixes:
  Dimmed labels keep every letter: the Partials label lost its I when
  the current mode left that control out of use.
  The Order list fits the long tune names on its face; it cut them off
  mid-word. The open list still shows the full names.
  The I IV V I preset no longer shows a red warning. A cadence holds
  each chord for about two seconds and is not judged as a melody. Tunes
  now count only the notes that sound, and Gymnopedie 1 plays a little
  quicker, 3.9 steps a second, to stay inside the melody range.
  After Amacher into a third asks for a 100 ms attack, the longest the
  knob allows; it asked for 200.

New in 1.3.4, fixes:
  ThirdEar no longer has an audio input. Nothing used it, and audio sent
  into it shared the safety clip with the synth, where a loud input could
  clip the tones and put a real note in the signal.
  Loud norm no longer boosts the lowest CDT notes: below 40 Hz the boost
  fades out, so a C0 pair is not pushed at the woofer for a phantom
  nobody can hear.
  The manual is corrected and clarified throughout.

New in 1.3.3, fixes:
  Raising Partials during a note keeps the phantom at full strength.
  The added tones used to come in out of step: the phantom was up to
  10 dB weaker, and CANCEL could not null it until the next note.
  Chord plays combs in either Mode. In CDT a chord put its own notes
  into the signal as real tones.
  Switching Norm, changing Partials and adding a chord note no longer
  click under the phantom.
  The panel's meter no longer causes dropouts at high sample rates.

New in 1.3.2, fixes:
  The host's stop gestures (all notes off, choke) now stop ThirdEar
  with Chord or Pattern on; they did not. The Carrier and Ratio glide
  is the same in every host and at every buffer size, other parameter
  changes no longer restart it, and a knob turned while nothing plays
  no longer glides on the next note. Lighter on the audio thread with
  the panel closed.

New in 1.3.1:
  Norm has a third setting, loud, now the default: each note is set so
  its tones carry the same A-weighted power (a standard stand-in for
  loudness), so the two modes sit at one level across most of the
  keyboard; the lowest CDT notes are boosted by at most 6 dB. Level
  starts at -15 dB. The preset list scrolls.

New in 1.3.0:
  SPLIT sends every other tone of the comb to each ear. A distortion
  product needs its tones in one cochlea, so each ear now makes its own,
  an octave up in QDT, and none at all in CDT. Headphones only.
  CHORD plays every held key as its own phantom, up to four, tuned just
  or equal. Tuned just, the combs' cross products land on the chord's
  root; switch to equal and hear them scatter.
  COCHLEA, on the graph, shows the basilar membrane unrolled: where the
  tones are felt, where the product is made, and where it is heard.
  The tunes are now the Passion Chorale, Satie's Gymnopedie No. 1 and
  Ave maris stella, and three After Amacher presets follow her
  documented technique.

thirdear-note.pdf is the manual: setup, the panel, the controls, the
presets and the bench, then the model and the literature it was checked
against (document DB-WP-002).

https://databurn.org
