A cleaner voice clone starts with the source recording

I'm on the FreeVoiceClone team. Before adjusting a generated voice, it helps to check the recording used to create it.

For a first sample, record about 30 seconds of your own voice, or a voice you have permission to use. Choose a quiet room, keep one speaker in the recording, and speak at a natural pace. Avoid background music or strong effects that compete with the voice.

Listen back before uploading. If words are clipped or room noise masks consonants, record another take. In FreeVoiceClone, you can upload an audio file or use the Record option in the source-audio step.

Then compare a short generated passage with the source. Use the same test sentence when comparing takes, and listen for pronunciation, pacing, and whether the voice still sounds like the intended speaker. Check the public/private voice setting before creating your clone.

The attached screenshot shows the source-audio step on our English cloning page.

Create a voice: https://freevoiceclone.com/clone

When preparing a voice sample, is finding a quiet recording space or choosing the right text the harder part?