Timing Sentences and Delays

[vestal edge]

Hello!
I've a project where I need to set timings and delays.
For example:

[
    {
        "text": "We've got all the information you need here, and we explain why",
        "in": 0.37537499999999996,
        "out": 4.546208333333333
    },
    {
        "text": "melting glaciers could threaten millions of people's lives.",
        "in": 4.587916666666667,
        "out": 8.592916666666666
    },
    {
        "text": "If you're struggling to see how renewable energy",
        "in": 10.676333333333332,
        "out": 13.220541666666666
    },
    {
        "text": "could replace our usual energy sources, check out these solar powered aircraft.",
        "in": 13.220541666666666,
        "out": 18.851166666666664
    }
]

I need the voice to start at the in point and end at the out point.

Solutions that I brainstormed about:

• Get Words-Per-Minute to calculate the rate and set the rate to the API.
• Set delay blocks inside the combined text to pause the talking.

But I don't know if the information I need for that are available.
Does anyone have any idea to solve that problem? (I'm using Java, but I can transcode everything, so every language answer is welcome)

[lime oriole]

There is no way to force the timings unfortunately. Your best bet if you don't need low latency would be to hit this https://elevenlabs.io/docs/api-reference/text-to-speech-with-timestamps and then if the response doesn't fit in your time requirements you can regenerate it until it fits. I know it's not perfect but I think it's the best you could get.

[vestal edge]

I see, thank you for your answer!
Another thought is to setup the voice to be more speedy or more slow. Is there any option to get a result like that?

[lime oriole]

Not really, apart from adding punctuation to slow it down but that can add artifacts etc

OR look a the REQUEST STICHING

you could add prompts like "he spoke slowly" etc.

[vestal edge]

Good idea, thank you! I' try that