1 · Enable the microphone and pick a device
Open Settings → Microphone. Flip the "Microphone active" switch on. The "Microphone device" dropdown shows all available inputs:
- Built-in laptop mic — works in a pinch, but noticeably thin
- USB condenser (Rode NT-USB+, Shure MV7) — recommended for solo work
- XLR mic via interface (Shure SM7B + Scarlett) — broadcast quality
- Headset mic — practical for long sessions, reduces cable clutter
The microphone permission is requested once by Windows. If the dialog doesn't appear: check Windows Settings → Privacy → Microphone, Hypnotika has to be allowed.
2 · Live VU meter for levelling
Directly below the device dropdown you'll see a live vertical VU meter. Speak normally into your mic:
- Peaks at −18 to −12 dB → ideal, clean headroom
- Peaks from −6 dB upwards → too loud, reduce input gain on the mic or in the Windows mixer
- Peaks below −30 dB → too quiet, ducking won't trigger reliably
The "Test" button records 5 seconds and plays them back — so you hear how your voice sounds in the recording, including room reflections.
3 · What is ducking?
Ducking means: as soon as you speak, the music becomes quieter automatically so your voice stays clear. As soon as you stop, the music gently returns to the original level.
The trick: Hypnotika uses Silero VAD (Voice Activity Detection), a neural model. It detects actual speech — unlike classical threshold-based duckers, which also trigger on paper rustling, coughing or ambient noise. The result: only your actual speech lowers the music.
4 · The three ducking parameters
Under Settings → Microphone → Ducking you'll find three sliders:
Attenuation (dB)
How much the music drops while you speak. Recommendation: −9 dB (default). That's about half the perceived loudness — enough for your voice to dominate, without the music disappearing. For hypnotic inductions with a soft voice, you can reduce it to −6 dB. For louder music or directive techniques, rather −12 dB.
Attack (ms)
How quickly the music is attenuated once speech is detected. Recommendation: 50 ms (default). Too short (< 20 ms) → audible click. Too long (> 150 ms) → the first syllables get masked by the music. 50 ms is inconspicuous in almost every context.
Release (ms)
How long it takes until the music returns to 0 dB after you stop talking. Recommendation: 400 ms (default). Too short (< 200 ms) → after every pause the music audibly "breathes" back up, feels nervous. Too long (> 800 ms) → the music stays unnecessarily quiet, continuity suffers.
5 · Session recording as MP3
With the mic active, Hypnotika automatically records a mixed track: music + voice, with ducking correctly applied. The result is a single MP3 file you can send directly to your client.
Related settings:
- Recording folder — path is configurable, default:
%USERPROFILE%\Documents\Hypnotika\Sessions - Filename pattern —
{date}_{time}_{clientname}.mp3(client name optional per session) - Bitrate — 192 kbps (default, good compromise) or 320 kbps (maximum)
Recording starts with the first track play and ends automatically on Soft Stop. No manual button-pressing needed.
6 · Alternative: recording via OBS
If you'd rather use OBS Studio or an external DAW for recording (e.g. for video sessions or for later editing): turn off Hypnotika's own recording in the settings so nothing is recorded twice. If you want only the music (with the anchor sounds) to go to OBS and not your mic, also turn off the microphone in Hypnotika — then you record your voice directly in OBS. Details in the audio guide.