Resample audio to 16kHz - Remotion

:::warning We are phasing out Remotion WebCodecs and are moving to Mediabunny! :::

For using Whisper, Whisper.cpp, @remotion/install-whisper-cpp and Whisper.wasm, you need to provide a 16-bit, 16KHz, WAVE audio file.

This page describes approaches for converting your audio in the browser and on the server.

In the browser

You can use convertMedia() to convert any of the supported video and audio formats and resample it to 16kHz.

import {LicenseDisclaimer} from './LicenseDisclaimer';

<details> <summary>💼 Important `@remotion/webcodecs` License Disclaimer</summary> <LicenseDisclaimer /> </details>

tsx

import {convertMedia, canReencodeAudioTrack} from '@remotion/webcodecs';

const output = await convertMedia({
  src: 'https://example.com/input.mp4',
  container: 'wav',
  onAudioTrack: async ({track}) => {
    if (
      await canReencodeAudioTrack({
        audioCodec: 'wav',
        track,
        // Ignore this, bitrate is not used for WAV files
        bitrate: 128000,
        sampleRate: 16000,
      })
    ) {
      return {
        type: 'reencode',
        audioCodec: 'wav',
        bitrate: 128000,
        sampleRate: 16000,
      };
    }

    // If this conversion is not supported, return an error
    return {
      type: 'fail',
    };
  },
});

const blob = await output.save(); // returns a `Blob`

On the server

You can use ffmpeg to convert your audio to a 16-bit, 16KHz, WAVE file.

bash

ffmpeg -i /path/to/audio.mp4 -ar 16000 /path/to/audio.wav -y

If you don't want to install FFmpeg, you can also use the smaller FFmpeg binary provided by Remotion.

bash

npx remotion ffmpeg -i input.mp4 -ar 16000 output.wav -y