Create Custom AI Voices With Google AI Studio (Full Transcript)

[00:00:00] Speaker 1: In this video, I'll show you how to use Google AI Studio's free AI Voice Generator to turn text into realistic speech. Like this.

[00:00:09] Speaker 2: We almost ended a friendship over candy corn.

[00:00:13] Speaker 3: It has its place.

[00:00:15] Speaker 2: Sure, we just disagree on where that place is.

[00:00:18] Speaker 1: And neither of those voices is a real person speaking. I'll show you how to generate your own voices, customize how they sound, and download the finished audio. I'm Elizabeth. Let's take a look. First, head to AI Studio. You can find the link in the video description below. For this example, we'll create a voicemail greeting for a fictional candy store called Elizabeth's Candy Connection. I'll start by creating a new dialogue and typing my desired greeting here in the text box.

[00:00:51] Speaker 2: Thank you for calling Elizabeth's Candy Connection. Please leave your name, number, and a brief message and we'll get back to you soon.

[00:00:58] Speaker 1: And there's our first greeting. We've gone from a few lines of text to an audio recording. Now let's customize how this greeting sounds. I can either choose a new speaker in its entirety.

[00:01:10] Speaker 4: Observe how the metaphor develops across the third chapter.

[00:01:14] Speaker 1: Observe how the metaphor develops across the third chapter. Or I can describe my desired style here. For a professional greeting, I want the voice to sound warm and welcoming. Think of this like giving directions to a voice actor.

[00:01:29] Speaker 2: Thank you for calling Elizabeth's Candy Connection. Please leave your name, number, and a brief message and we'll get back to you soon.

[00:01:37] Speaker 1: That sounded warm and welcoming, but let's try it again with a completely different personality.

[00:01:44] Speaker 2: Thank you for calling Elizabeth's Candy Connection. Please leave your name, number, and a brief message and we'll get back to you soon.

[00:01:51] Speaker 1: Once you're happy with how your voice sounds, select Run on the bottom right-hand corner and then you can download your new voice directly to your computer. The same steps work for a YouTube voiceover, a training video, or a product demonstration. Just replace the greeting with your own script. Now let's create a conversation between two speakers. We'll make a short trailer for a podcast called Elizabeth's Candy Connection. I'll open a two-speaker template. We'll start by assigning a different voice to each speaker. For my first voice, I want someone who's warm. I'll select Cleo for now. For the second voice, I want something contrasting so that they're easy to tell apart. Let's go with the witty podcast host. Now let's replace the sample text with our script. Now this is my favorite part of the Google AI Studio because you'll notice in each one of these speaker boxes, you can adjust the tone. For this very first speaker, I want an upbeat, enthusiastic podcast intro with a bright vocal smile and a brisk tempo. While in the second speaker box, I'm asking for a more animated, warm agreement with an eager conversational momentum. You can adjust this tone for each one of these speaker boxes, but you don't have to. Now something else we haven't touched on is this expression button here. This is where you can add some natural expression throughout your dialogue. Here I want to add a little cackle before my first speaker says it has its place. Let's click on run and see how it sounds.

[00:03:33] Speaker 2: Welcome to Elizabeth's Candy Connection, where we take our sweets and our candy debates very seriously.

[00:03:40] Speaker 5: Very seriously. We almost ended a friendship over candy corn. It has its place. Sure. I disagree on where that place is.

[00:03:50] Speaker 2: Join us as we taste new treats, revisit childhood favorites, and settle the biggest debates in candy.

[00:03:56] Speaker 6: Starting with whether you actually have to share.

[00:03:59] Speaker 1: If you're happy with how everything sounds, great, click on download. I do want to make a change to this. For this particular speaker, it has its place. It sounds more like it's asking a question versus being playful, so I can go ahead and change the tone. I'm going to remove the curious and leaning into the mic park and just leave it at playful.

[00:04:32] Speaker 7: It has its place.

[00:04:34] Speaker 1: Much better. But something still sounds a little bit off to me. I'm going to try a new expression here to see if that sounds more desirable. I removed the first expression of the cackle and let's insert a chuckle instead.

[00:04:51] Speaker 2: It has its place.

[00:04:53] Speaker 1: There we go. That's exactly how I want this podcast to sound. Another useful setting when dealing with multiple speakers is to automatically have the AI studio add filler words. These are conversational interjections, natural background noises, and things that typically happen during a conversation like using hmm or ah. However, if you don't want this, you can easily toggle it off. You can get started with Google AI studio for free, although usage limits do apply. If you need to generate more audio, Google also offers paid options. Give it a try and let me know what you think in the comments below. And while you're at it, tell me what your favorite childhood candy is. I'm going to eat a Twix. Thanks for watching.