@audio/denoise 0.3.8 → 0.3.9
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +57 -0
- package/index.js +2 -0
- package/package.json +5 -3
package/README.md
CHANGED
|
@@ -19,7 +19,9 @@ Single-pass noise reduction. 13 specialised methods + an auto-classifier.
|
|
|
19
19
|
| [deplosive](#deplosive) | time | LF bursts | ★★★ | low | mic plosives (p, b) |
|
|
20
20
|
| [deesser](#deesser) | time | sibilance | ★★★★ | low | voice (s, sh) |
|
|
21
21
|
| [debreath](#debreath) | time | inter-word noise | ★★★ | low | breath / hiss in pauses |
|
|
22
|
+
| [desilence](#desilence) | time | pauses | ★★★ | low | remove / shorten / trim silence, split by pause |
|
|
22
23
|
| [dereverb](#dereverb) | freq | late reverb | ★★ | medium | moderate room reverb |
|
|
24
|
+
| [dewow](#dewow) | time + freq | pitch drift | ★★★ | medium | tape / vinyl / cassette wow & flutter |
|
|
23
25
|
|
|
24
26
|
For broader DSP needs use [stretch](https://github.com/audiojs/stretch), [shift](https://github.com/audiojs/shift), [pitch](https://github.com/audiojs/pitch), [beat](https://github.com/audiojs/beat).
|
|
25
27
|
|
|
@@ -290,6 +292,34 @@ dereverb(data, { t60: 0.6, predelay: 0.04 })
|
|
|
290
292
|
**Not for:** heavy reverb or convolutive distortion — use multi-channel WPE (out of scope).
|
|
291
293
|
|
|
292
294
|
|
|
295
|
+
## Pitch drift
|
|
296
|
+
|
|
297
|
+
### `dewow`
|
|
298
|
+
|
|
299
|
+
Wow & flutter correction. Estimates the transport speed curve — from stable spectral partials tracked with a phase vocoder (McAulay–Quatieri linking, amplitude-weighted median across tracks), from a known reference tone (mains hum, calibration tone, tape bias residual), or from monophonic pitch — then reads the signal back through a variable-rate windowed-sinc resampler along the integrated curve. Whole-signal (needs the full curve), length-preserving by default. The classical counterpart of Celemony Capstan; `wowFlutter()` alone is a wow & flutter meter.
|
|
300
|
+
|
|
301
|
+
```js
|
|
302
|
+
dewow(data, { fs }) // partial tracking (default)
|
|
303
|
+
dewow(data, { fs, mode: 'reference', refFreq: 50 }) // lock to 50 Hz hum
|
|
304
|
+
dewow(data, { fs, wow: true, flutter: false }) // remove slow drift only
|
|
305
|
+
wowFlutter(data, { fs }) // → { speed, times, wow, flutter, confidence }
|
|
306
|
+
```
|
|
307
|
+
|
|
308
|
+
| Param | Default | |
|
|
309
|
+
|---|---|---|
|
|
310
|
+
| `mode` | `'partial'` | `'partial' \| 'reference' \| 'pitch'` |
|
|
311
|
+
| `refFreq` | — | Hz — the known tone for `reference` mode |
|
|
312
|
+
| `smooth` | `0.05` | s — flutter-band smoothing of the speed curve |
|
|
313
|
+
| `wow` / `flutter` | `true` / `true` | correct the < 6 Hz / 6–43 Hz bands |
|
|
314
|
+
| `maxDeviation` | `0.05` | clamp on the speed ratio (±5 %) |
|
|
315
|
+
| `keepLength` | `true` | output length equals input |
|
|
316
|
+
|
|
317
|
+
Measured on synthetic 2 % wow @ 0.8 Hz + 0.4 % flutter @ 30 Hz over a sustained chord: curve correlation 0.98, residual deviation of the 440 Hz partial 1.4 % → 0.18 %; clean input passes at 44 dB SNR.
|
|
318
|
+
|
|
319
|
+
**Use when:** tape, cassette, vinyl and film transfers with audible pitch wobble; material with sustained tones or a reference tone.<br>
|
|
320
|
+
**Not for:** dropouts, azimuth/time-skew, material with no stable partials (percussion-only); flutter above `fs/(2·hopSize)` ≈ 43 Hz at defaults needs a smaller `hopSize`.
|
|
321
|
+
|
|
322
|
+
|
|
293
323
|
## Gates & inter-word
|
|
294
324
|
|
|
295
325
|
### `gate`
|
|
@@ -315,6 +345,33 @@ debreath(data, { range: -10 }) // -10 dB on non-s
|
|
|
315
345
|
**Use when:** breath, mouth noise, hiss in pauses on a voiceover.
|
|
316
346
|
|
|
317
347
|
|
|
348
|
+
### `desilence`
|
|
349
|
+
|
|
350
|
+
VAD-driven silence editing. Runs `vad` once, folds frames into speech/pause segments, then cuts pauses per `mode` with equal-power crossfades at every splice — never a hard cut. Length-changing, so it is a batch call (and a `silence` stat atom), not a streaming processor.
|
|
351
|
+
|
|
352
|
+
```js
|
|
353
|
+
desilence(data, { mode: 'shorten', maxSilence: 0.25 }) // Overcast "Smart Speed": long pauses → 0.25 s
|
|
354
|
+
desilence(data, { mode: 'remove', minSilence: 0.5, pad: 0.1 }) // cut pauses > 0.5 s, keep 0.1 s around speech
|
|
355
|
+
desilence(data, { mode: 'trim' }) // leading/trailing silence only
|
|
356
|
+
silenceSegments(data, { fs }) // → { speech: [{start, end}], silence: [...] }
|
|
357
|
+
splitSilence(data, { fs }) // → one Float32Array per phrase
|
|
358
|
+
```
|
|
359
|
+
|
|
360
|
+
| Param | Default | |
|
|
361
|
+
|---|---|---|
|
|
362
|
+
| `mode` | `'shorten'` | `'shorten' \| 'remove' \| 'trim'` |
|
|
363
|
+
| `minSilence` | `0.5` | s — pauses shorter than this are never touched |
|
|
364
|
+
| `maxSilence` | `0.25` | s — `shorten` target length of any remaining pause (cut from the middle, onsets kept) |
|
|
365
|
+
| `pad` | `0.1` | s — silence kept on each side of speech in `remove` mode |
|
|
366
|
+
| `merge` | `0.15` | s — speech gaps shorter than this join one segment |
|
|
367
|
+
| `fade` | `0.01` | s — crossfade at every cut |
|
|
368
|
+
|
|
369
|
+
Returns `{ data, segments, removed, map }`; `project(map, t)` re-times markers/subtitles (`@audio/subtitle` `project()` takes the same map).
|
|
370
|
+
|
|
371
|
+
**Use when:** podcast/lecture pause tightening, split-by-silence, auto-trim.<br>
|
|
372
|
+
**Not for:** music (a rest is a pause to this VAD); overlapping speech.
|
|
373
|
+
|
|
374
|
+
|
|
318
375
|
## Quality measurement
|
|
319
376
|
|
|
320
377
|
```js
|
package/index.js
CHANGED
|
@@ -16,6 +16,8 @@ export { default as debreath } from '@audio/denoise-debreath'
|
|
|
16
16
|
export { default as dereverb } from '@audio/denoise-dereverb'
|
|
17
17
|
export { default as denoise, classify, deesser } from '@audio/denoise-detect'
|
|
18
18
|
export { default as repair } from '@audio/denoise-repair'
|
|
19
|
+
export { default as desilence, segments as silenceSegments, split as splitSilence } from '@audio/denoise-desilence'
|
|
20
|
+
export { default as dewow, analyze as wowFlutter } from '@audio/denoise-dewow'
|
|
19
21
|
|
|
20
22
|
export { snr, segSnr, lsd, nrr, speechAttenuation } from '@audio/quality'
|
|
21
23
|
export { vad, spp, ddSnr } from '@audio/vad'
|
package/package.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@audio/denoise",
|
|
3
|
-
"version": "0.3.
|
|
3
|
+
"version": "0.3.9",
|
|
4
4
|
"type": "module",
|
|
5
|
-
"description": "Noise reduction
|
|
5
|
+
"description": "Noise reduction — gate, spectral subtraction, Wiener, OM-LSA, dehum, declick, decrackle, declip, deesser, dereverb",
|
|
6
6
|
"sideEffects": false,
|
|
7
7
|
"exports": {
|
|
8
8
|
".": "./index.js",
|
|
@@ -63,7 +63,9 @@
|
|
|
63
63
|
"@audio/dynamics-gate": "^0.1.4",
|
|
64
64
|
"@audio/denoise-repair": "^0.1.0",
|
|
65
65
|
"@audio/quality": "^1.0.0",
|
|
66
|
-
"@audio/stft": "^1.0.3"
|
|
66
|
+
"@audio/stft": "^1.0.3",
|
|
67
|
+
"@audio/denoise-desilence": "^0.1.0",
|
|
68
|
+
"@audio/denoise-dewow": "^0.1.0"
|
|
67
69
|
},
|
|
68
70
|
"devDependencies": {
|
|
69
71
|
"audio-lena": "^3.0.0",
|