Tips on Getting Great Voice Clones from ElevenLabs

Today, I final­ly made a voice clone of my own voice to use on this blog.

I had actu­al­ly done this process once before using the pro­fes­sion­al voice cloning option from Eleven­Labs. But I didn’t end up using that ver­sion — it just didn’t sound quite right to me. 

The mis­take I made was my own fault. To cre­ate a pro­fes­sion­al voice clone you need at least 30 min­utes of clean audio. I dis­cov­ered that with­out being very judi­cious in my selec­tion of audio, I had uploaded audio that includ­ed unnat­ur­al paus­es or sen­tence frag­ments and those aspects of my audio end­ed up in the voice clone as well. It turns out that to get a good result with pro­fes­sion­al voice cloning, you need real­ly high-qual­i­ty voice sam­ples, and a lot of them. 

This time around I used the Instant Voice Cloning fea­ture.

It still requires a good voice sam­ple, but not as many min­utes of it. Find­ing or cre­ate­ing a min­i­mum of two min­utes of clean autio is much eas­i­er to accom­plish. I’d encour­age you to start with this option if you’re exper­i­ment­ing with cre­at­ing your own voice clone.

Here’s my key tip if you’re plan­ning to try it:

Use your own writ­ing as the source mate­r­i­al when record­ing your voice sam­ple.

If you plan to use your voice clone to narrate your own writ­ing, this will make a big dif­fer­ence. That’s because the way you write and the way you speak are usu­al­ly tied togeth­er. Your nat­ur­al into­na­tion, rhythm, and empha­sis come through more eas­i­ly when you’re read­ing some­thing you wrote your­self. That makes the train­ing sam­ple more expres­sive — and that expres­sive­ness car­ries through when Eleven­Labs gen­er­ates new audio from your future writ­ing.

In short:

1. Choose the Instant Voice Cloning option.
2. Record your­self read­ing some­thing you wrote.
3. Use that as your train­ing sam­ple.

You’ll like­ly end up with a voice clone that sounds more nat­ur­al, more expres­sive, and more you.