Is it possible to adjust the volume or speed of audio in the realtime api?

API
, , ,
1

I want to make the AI voice speak louder and faster. Is it possible? I know I can mess with the audio that I get from the api but that adds more latency to do it in realtime.

1 Like
2

You can update the speed, but not the volume. The volume can be modified using something as follows …

        const source = this.audioContext.createBufferSource();
        const gainNode = this.audioContext.createGain();

        const arrayBuffer = await audioBlob.arrayBuffer();
        const audioBuffer = await this.audioContext.decodeAudioData(arrayBuffer);

        source.buffer = audioBuffer;
        gainNode.gain.value = 2.0; // Increase volume by 2x

        source.connect(gainNode);
        gainNode.connect(this.audioContext.destination);
        source.addEventListener('ended', () => {
          this.isSpeaking = false;
          this.isWaitingForTTS = false;
        });

The code is not complete, but I suspect you will understand the primary parts of it. Interestingly, that code was (mostly) generated by O1 … :wink:

3

ahh I was wondering if there was a way to do so through the LLM. I’m trying to avoid adding more latency to a realtime system

4

The code you’ve put only solves the volumen issue, but how can i set the speech speed? I’m looking for some commands to send to OpenAI based on some slider, but can’t find anything…