INTERVIEW LEVELER 1.0
A Nyquist process effect for Audacity 3.x

INSTALL
1. Extract this ZIP. Keep the .ny extension on Interview Leveler.ny.
2. In Audacity, open Tools > Nyquist Plugin Installer and choose the .ny file.
3. Relaunch Audacity. If Interview Leveler does not appear in the Effect menu,
   open Plugin Manager and enable it. Menu grouping can vary by version.

Official instructions:
https://manual.audacityteam.org/man/nyquist_plug_in_installer.html

USE
1. Save your project or duplicate the track so you can A/B the result.
2. Select 20 to 30 seconds with both voices, then open Interview Leveler.
3. Apply and listen. If it works, Undo the test, select the whole interview
   track and apply the effect once. It processes the selection offline.

STARTING SETTINGS
- Maximum quiet-voice boost: 15 dB. Lower it if room noise comes forward.
- Loud voice bypass level: -25 dB RMS. Lower it (e.g. -29) if the host gets
  too much upward boost; raise it a few dB if quiet turns get too little.
- Output gain: 5 dB. Lower this for an already loud recording.
- Limiter ceiling: -2 dBFS (sample peak). Leave it there to give encoding
  some headroom. This is not a fixed LUFS target or a true-peak meter.

WHAT IT DOES
A slow level detector lifts quieter phrases. It avoids extra quiet-voice
boost in very quiet pauses. A 75 Hz high-pass, light 2:1 compression and
a peak limiter keep the result controlled. Stereo channels remain linked.
The plugin works locally and does not open or send any files.

This is a reusable version of the workflow, not the exact settings from
one recording. It cannot identify speakers or separate overlapping voices.
If both people overlap, or if louder background sound triggers the detector,
use Audacity's Envelope Tool for that section. If you have isolated mic
tracks, level those tracks independently instead.
