Streaming TTS

RunAnywhere.synthesizeStream(
  text: string,
  options?: Partial<TTSOptions>,
  extra?: { signal?: AbortSignal }
): AsyncIterable<TTSOutput>

import { RunAnywhere, AudioFormat } from '@runanywhere/web'
import { AudioPlayback } from '@runanywhere/web/browser'

const player = new AudioPlayback()

for await (const chunk of RunAnywhere.synthesizeStream('One sentence. Then another.')) {
  const samples = toFloat32(chunk)
  if (samples.length > 0) {
    await player.play(samples, chunk.sampleRate)
  }
  if (chunk.isFinal) break
}

player.dispose()

const controller = new AbortController()
const stream = RunAnywhere.synthesizeStream(text, undefined, { signal: controller.signal })
// later
controller.abort()

function toFloat32(chunk: { audioData: Uint8Array; audioFormat: AudioFormat }): Float32Array {
  const buffer = chunk.audioData.buffer.slice(
    chunk.audioData.byteOffset,
    chunk.audioData.byteOffset + chunk.audioData.byteLength
  )
  return chunk.audioFormat === AudioFormat.AUDIO_FORMAT_PCM_S16LE
    ? RunAnywhere.pcm16ToFloat32Samples(buffer)
    : new Float32Array(buffer)
}

import { AudioPlayback } from '@runanywhere/web/browser'

const player = new AudioPlayback({ sampleRate: 22050 })
console.log(player.isPlaying)

await player.play(samples, 22050)

player.pause()
player.resume()
player.setVolume(0.5)
player.stop()
player.dispose()

const output = await RunAnywhere.synthesize('Hello world')
const samples = toFloat32(output)

const audioCtx = new AudioContext()
const buffer = audioCtx.createBuffer(1, samples.length, output.sampleRate)
buffer.getChannelData(0).set(samples)

const source = audioCtx.createBufferSource()
source.buffer = buffer
source.connect(audioCtx.destination)
source.start()

const output = await RunAnywhere.synthesize('Hello world')
const bytes = output.audioData

const wav = RunAnywhere.pcm16ToWav(
  bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength),
  output.sampleRate
)

const url = URL.createObjectURL(new Blob([wav], { type: 'audio/wav' }))
const a = document.createElement('a')
a.href = url
a.download = 'speech.wav'
a.click()
URL.revokeObjectURL(url)

async function speakSequence(texts: string[]) {
  for (const text of texts) {
    await RunAnywhere.speak(text)
  }
}

await speakSequence(['First sentence.', 'Second sentence.', 'And the third.'])