Tinnitus Freq Finder (Multi-Layer)

Sample rate Duration (s) Tip: test with 10–30s; long exports are heavy.

Multi-Layer Synth (up to 3 layers)

Layer 1

Tone + band / cluster

Frequency (Hz) Bandwidth (Hz)

Tone mix Noise mix Level

Noise color Texture AM rate (Hz) AM depth

Layer 2

Optional 2nd band

Frequency (Hz) Bandwidth (Hz)

Tone mix Noise mix Level

Noise color Texture AM rate (Hz) AM depth

Layer 3

Optional 3rd band

Frequency (Hz) Bandwidth (Hz)

Tone mix Noise mix Level

Noise color Texture AM rate (Hz) AM depth

Global Muffle, Gating & Safety

Muffle cutoff (Hz) 0 = off Muffle order

Slice ON (ms) Slice OFF (ms) (Set ON > 0 to enable; OFF can be 0)

Output level dB (capped) Routing Disable HF pad >10 kHz

Shift+Drag for fine control • Live update for layer Freq, BW, Mix/Level, AM, Muffle cutoff, Output. Changes to Texture, Slice, Muffle order, Duration apply on next start.

Files include settings in the filename.

Notes

Texture: “Noise” = classic filtered noise (grainy); “Smooth hiss” = many micro-sines (fine, steady); “Hum” = few detuned sines (soft, tonal).
Change Texture then press Stop → Play to rebuild the source (the switch also autostops/plays now).

How to dial it in

For soft hum/hiss:
Texture = Smooth hiss, Noise mix ~0.6–0.9, Tone mix ~0.1–0.4, BW ~80–250 Hz.
If you want even silkier: bump K in buildGaussHiss to 96 or 128.

For a hummier base:
Texture = Hum, BW ~40–120 Hz, Noise mix ~0.6, Tone mix ~0.3.

Pitch-Anchor Sweep (Quick Method)

Setup (one time for anchoring)

Layer 1 only: enabled.
Tone mix = 1.00, Noise mix = 0.00 (pure sine).
Level modest (e.g., 0.5–0.7).
BW can stay at 0 (only affects noise/hum; with Noise mix = 0 it’s irrelevant).
AM depth = 0, Muffle off, Slice off.
Route to the ear with the louder tinnitus.

Sweep

Coarse sweep around your expected region (e.g., ~2.6 kHz):
Start at 1 kHz, then jump to 1.5 → 2.0 → 2.3 → 2.6 → 2.9 → 3.2 → 3.5 → 4.0 kHz.
Listen for where the external tone seems to “sit inside” your tinnitus and beating/roughness is minimal.
Zoom in around the best spot:
Use ~100 Hz steps (e.g., 2400, 2500, 2600, 2700).
Then 10 Hz steps (e.g., 2580–2620).
Use Shift+drag for fine control.
Micro-tune if needed:
Use 1–2 Hz nudges to land on the “least beating / most fused” point.
That number is your center f₀.

Only after you’ve got f₀…

Bring Noise mix up (e.g., 0.3–0.8) and pick Texture (“Smooth hiss” for fine hiss, “Hum” for soft tonality).
Now BW matters — open BW until the character (hiss/hum body) matches your tinnitus.
Report four numbers to the team:
low = f₀ − BW/2, center = f₀, high = f₀ + BW/2, width = BW.

Key clarifications

Don’t set both Tone and Noise to 0 — you won’t hear anything. For anchoring, use Tone=1, Noise=0.
You don’t need to repeat the sweep with BW=10 or 20 for anchoring; BW only matters once you add noise/hum to shape the body.
Keep volume low and stable during the sweep; avoid just after shower/exercise/meditation to reduce short-term shifts.

Tinnitus Freq Finder — User Guide

0) What it is

Plays a tone + band‑limited noise centered at your chosen frequency
Adds optional muffle (treble roll‑off)
Adds optional slice/gating (ON/OFF ms) for choppy/raspy percepts
Exports 16‑bit WAV files you can use elsewhere

1) Setup (60 seconds)

Open this page in Chrome or Edge. Safari/Firefox usually work, but can be slower with long renders.
Use decent headphones/earbuds or monitors. Start at low volume.

Browsers block auto‑play. You must click Play / Monitor to hear anything.

2) Quick Start (2 minutes)

Set Sample rate (48,000 Hz is fine).
Set Duration (10–30 s for testing; render longer only when exporting).
Set Frequency near your perceived tone (e.g., 8000 Hz).
Set Bandwidth: 0–400 narrow/clean; 600–1200 broader/muffled; >1500 hissier.
Adjust Tone mix vs Noise mix (auto‑normalized): tonal 0.7/0.3; muffled 0.3/0.7.
If dull/muffled, enable Muffle: cutoff 5–7 kHz, order 2–4.
If choppy/raspy, enable Slice: try ON 2 ms / OFF 8 ms.
Click Play / Monitor. The button turns red Stop while playing and flips back when done.
When it blends with your tinnitus best, click Export WAV.

3) Controls Cheat Sheet

Frequency (Hz): center pitch.
Bandwidth (Hz): width around the center (≈ Q = f0 / BW). Bigger = noisier.
Tone/Noise mix: relative weights (keeps overall level sane).
Muffle: smooth high‑frequency roll‑off; higher order = more “under a pillow.”
Slice ON/OFF (ms): amplitude gating; ON>0 enables it. Duty ≈ ON / (ON+OFF).
Duration: playback length and export length.

Fine control: Shift+drag sliders.

4) Recommended Starting Presets

Clean, tonal: Freq≈guess; BW 200–400; Tone 0.7 / Noise 0.3; Muffle off; Slice off.
Muffled: BW 600–1200; Tone 0.3 / Noise 0.7; Muffle cutoff 5–7 kHz, Order 2–4; Slice off.
Choppy/raspy: Start from “Muffled,” then Slice 2/8 ms. Try 3/7, 4/6, 1/9 if 2/8 isn’t right.
Very high‑pitch hiss: Freq near your pitch; BW 1200–2000; Tone 0.2 / Noise 0.8; Muffle cutoff 7–10 kHz, Order 1–2.

5) Finding a Perceptual Match (fast workflow)

Coarse pass: sweep Frequency (±5–10%), BW at 400, 800, 1200.
Character pass: adjust Tone/Noise and Muffle to get timbre right.
Texture pass: if choppy, add Slice (start 2/8 ms) and tweak ON/OFF.
Export a 30–60 s candidate. If it still matches, export therapy length.

6) Exporting WAVs

16‑bit PCM mono at chosen sample rate.
Very long renders (10–30 min) can be heavy. If it fails, export a short loop and stitch it in a DAW.
Filenames encode all settings.

7) Troubleshooting

No sound: you must click Play / Monitor; also check tab/system volume.
Distortion: turn down output; the app has headroom but your DAC can still clip.
Export fails: reduce Duration; export a short loop and stitch in a DAW.
Firefox sluggish: use Chrome/Edge.
iOS: use a desktop browser for exporting long files.

8) Safety & Protocol

Volume: start low. Don’t blast your ears.
Trial rules: follow your study protocol; don’t upload files to a device if it violates the agreement.
This tool is not a medical device—just an audio lab.

Runs locally in your browser; no network calls or analytics.

9) Built‑in Safety Features (what the app enforces)

Output cap: master output slider is limited to −24…−6 dBFS. Exports use the same cap.
Limiter: a fast dynamics limiter after the master catches brief transients around ≈ −6 dBFS.
Soft start/stop: 300 ms fade‑in/out to avoid thumps.
HF pad: automatic −6 dB padding when Frequency > 10 kHz (badge shows). You can override with “Disable HF pad >10 kHz.”
Peak monitor + auto‑backoff: LED + live dBFS readout. If the peak approaches −1 dBFS, the app reduces output by 3 dB automatically.
Panic stop: press Space or Esc to stop immediately.
Ear routing: Stereo / Left‑only / Right‑only so you can favor one ear if your protocol calls for it.
Click‑train caution: when Slice is enabled the app shows a headphones warning; keep volumes conservative because gating creates sharper transients.
Session time badge: the header shows Today: X/Y min with color cues and prompts at 75% & 100% of your chosen budget. (Not SPL‑calibrated.)

Limitations: the app cannot control your system/device volume and is not a dosimeter. Peak dBFS is not the same as ear‑canal dB SPL.

9) Built‑in Safety Features (what the app enforces)

Output cap: master output slider is limited to −24…−6 dBFS. Exports use the same cap.
Limiter: a fast dynamics limiter after the master catches brief transients around ≈ −6 dBFS.
Soft start/stop: 300 ms fade‑in/out to avoid thumps.
HF pad: automatic −6 dB padding when Frequency > 10 kHz (badge shows). You can override with “Disable HF pad >10 kHz.”
Peak monitor + auto‑backoff: LED + live dBFS readout. If the peak approaches −1 dBFS, the app reduces output by 3 dB automatically.
Panic stop: press Space or Esc to stop immediately.
Ear routing: Stereo / Left‑only / Right‑only so you can favor one ear if your protocol calls for it.
Click‑train caution: when Slice is enabled the app shows a headphones warning; keep volumes conservative because gating creates sharper transients.
Session time badge: the header shows Today: X min so you can tally daily exposure time. (Not SPL‑calibrated.)

Limitations: the app cannot control your system/device volume and is not a dosimeter. Peak dBFS is not the same as ear‑canal dB SPL.

🎧 Tinnitus Freq Finder

Multi-Layer Synth (up to 3 layers)

Layer 1

Layer 2

Layer 3

Global Muffle, Gating & Safety

Notes

How to dial it in

Pitch-Anchor Sweep (Quick Method)

Setup (one time for anchoring)

Sweep

Only after you’ve got f0…

Key clarifications

Tinnitus Freq Finder — User Guide

0) What it is

1) Setup (60 seconds)

2) Quick Start (2 minutes)

3) Controls Cheat Sheet

4) Recommended Starting Presets

5) Finding a Perceptual Match (fast workflow)

6) Exporting WAVs

7) Troubleshooting

8) Safety & Protocol

9) Built‑in Safety Features (what the app enforces)

9) Built‑in Safety Features (what the app enforces)

Only after you’ve got f₀…