Output Destinations

VoiSona Talk supports three types of output destinations for synthesized speech.

1. File Output (Default)

The synthesized speech is saved as a WAV file on your local filesystem.

const result = await client.synthesizeAndWait({
  text: 'ファイルに出力します。',
  language: 'ja_JP',
  destination: 'file',
  output_file_path: 'C:\\path\\to\\output.wav',
  can_overwrite_file: true
});

• Pros: Easy to use for simple scripts.
• Cons: Requires filesystem access; overhead of file I/O.

2. Memory Output (New in v0.9.1)

Synthesize speech directly into memory and retrieve it as an ArrayBuffer. This is ideal for web servers or applications that need to process the audio data in-memory.

The library provides a convenience method synthesizeToBuffer for this purpose.

// Synthesize to memory and get ArrayBuffer directly
const buffer = await client.synthesizeToBuffer({
  text: 'メモリに出力します。',
  language: 'ja_JP'
});

// Use the buffer (e.g., send as HTTP response)
// ...

• Pros: No temporary files; faster for real-time applications.
• Cons: Consumes system memory for large batches.

3. Audio Device

Output the speech directly through the default audio device configured in VoiSona Talk.

await client.requestSpeechSynthesis({
  text: 'スピーカーから出力します。',
  language: 'ja_JP',
  destination: 'audio_device'
});

Retrieving Audio Device Info

You can check which audio device is currently selected as the default:

const device = await client.getDefaultAudioDevice();
console.log(`Using device: ${device.device_name} (${device.device_type})`);

---

Which one to choose?

Use Case	Recommended Destination
Simple scripts / Saving for later	file
Web APIs / Real-time playback	memory
Direct interactive output	audio_device