# manifest-worklet

> The source of an AudioWorklet processor that plays one of several audio.js manifests (the audio.js of an @audio/* package): the model sent by message, a parameter it reads as it plays changed at once, a model made again for one it is made with, warmed on the last 100 ms and faded in, its state brought back to rest on silence, and a meter of the gain it takes off.

```bash
npx shadcn@latest add @audio-ui/manifest-worklet
```

```ts
import { manifestWorklet } from "@/lib/manifest-worklet";
```

- Files: `lib/manifest-worklet.ts`
- Depends on: nothing; npm: nothing
- Used by: `compander-device`, `compressor-device`, `deesser-device`, `ducker-device`, `expander-device`, `gate-device`, `limiter-device`, `transient-device`, `unlimiter-device`

| Export | Kind | What it is |
| --- | --- | --- |
| `manifestWorklet` | function | The source of an `AudioWorklet` processor that plays one of several audio.js manifests: the `audio.js` of an `@audio/*` package, a processor made from its parameters, which flags `restart` the ones it is made with. |

## Usage

The source of an `AudioWorklet` processor that plays one of several audio.js manifests (the `audio.js` of an `@audio/*` package) as a device's models: a model picked by message, a parameter the manifest reads as it plays changed at once, a change that needs a new instance faded in over the old one, and a meter of the gain it takes off.

```ts
import { manifestWorklet } from "@/lib/manifest-worklet";

// As the compressor's device.ts makes its models' processor:
export const COMPRESSOR_MODELS_WORKLET = manifestWorklet({
  kernels: COMPRESSOR_KERNELS, // the manifests, as the device's kernels.ts vendors them
  models: { fet: "fet", opto: "opto", smooth: "compressor", varimu: "varimu", vca: "vca" },
  processor: "audioui-compressor-models",
});
// then loadWorklet(ctx, COMPRESSOR_MODELS_WORKLET), and an AudioWorkletNode of "audioui-compressor-models" with processorOptions: { model: "smooth" }
```

The compressor, gate, expander, ducker, compander, unlimiter and others play their audio.js models through it.