@picovoice/koala-web

Koala Binding for Web

Koala Noise Suppression Engine

Made in Vancouver, Canada by Picovoice

Koala is an on-device noise suppression engine. Koala is:

• Private; All voice processing runs locally.
• Cross-Platform:
  • Linux (x86_64), macOS (x86_64, arm64), and Windows (x86_64)
  • Android and iOS
  • Chrome, Safari, Firefox, and Edge
  • Raspberry Pi (4, 3) and NVIDIA Jetson Nano

Compatibility

• Chrome / Edge
• Firefox
• Safari

Restrictions

IndexedDB is required to use Koala in a worker thread. Browsers without IndexedDB support (i.e. Firefox Incognito Mode) should use Koala in the main thread.

Installation

Package

Using Yarn:

yarn add @picovoice/koala-web

or using npm:

npm install --save @picovoice/koala-web

AccessKey

Koala requires a valid Picovoice AccessKey at initialization. AccessKey acts as your credentials when using Koala SDKs. You can get your AccessKey for free. Make sure to keep your AccessKey secret. Signup or Login to Picovoice Console to get your AccessKey.

Usage

For the web packages, there are two methods to initialize Koala.

Public Directory

NOTE: Due to modern browser limitations of using a file URL, this method does not work if used without hosting a server.

This method fetches the model file from the public directory and feeds it to Koala. Copy the model file into the public directory:

cp ${KOALA_MODEL_FILE} ${PATH_TO_PUBLIC_DIRECTORY}

Base64

NOTE: This method works without hosting a server, but increases the size of the model file roughly by 33%.

This method uses a base64 string of the model file and feeds it to Koala. Use the built-in script pvbase64 to base64 your model file:

npx pvbase64 -i ${KOALA_MODEL_FILE} -o ${OUTPUT_DIRECTORY}/${MODEL_NAME}.js

The output will be a js file which you can import into any file of your project. For detailed information about pvbase64, run:

npx pvbase64 -h

Koala Model

Koala saves and caches your model file in IndexedDB to be used by WebAssembly. Use a different customWritePath variable to hold multiple models and set the forceWrite value to true to force re-save a model file.

Either base64 or publicPath must be set to instantiate Koala. If both are set, Koala will use the base64 model.

const koalaModel = {
  publicPath: ${MODEL_RELATIVE_PATH},
  // or
  base64: ${MODEL_BASE64_STRING},

  // Optionals
  customWritePath: "koala_model",
  forceWrite: false,
  version: 1,
}

Initialize Koala

Set processErrorCallback to handle errors if an error occurs while enhancing audio. Create a processCallback function to get the streaming results from the engine:

// Optional
const options = {
  processErrorCallback: (error) => {}
}

function processCallback(enhancedPcm: Int16Array) {
  // do something with enhancedPcm
}

Create an instance of Koala on the main thread:

const handle = await Koala.create(
  ${ACCESS_KEY},
  processCallback,
  koalaModel,
  options // optional options
);

Or create an instance of Koala in a worker thread:

const handle = await KoalaWorker.create(
  ${ACCESS_KEY},
  processCallback,
  koalaModel,
  options // optional options
);

Process Audio Frames

The process function will send the input frames to the engine. The enhanced audio is received from processCallback as mentioned above. In case the next audio frame does not follow the previous one, call reset before calling process.

function getAudioFrame(): Int16Array {
  ... // function to get a frame of audio
  return new Int16Array();
}

await handle.reset();
for (;;) {
  await handle.process(getAudioFrame());
  // break on some condition
}

Clean Up

Clean up used resources by Koala or KoalaWorker:

await handle.release();

Terminate (Worker only)

Terminate KoalaWorker instance:

await handle.terminate();

Demo

For example usage refer to our Web demo application.