Hi everyone, Kevin here. Today we're going
to look at how you can use the most realistic sounding text-to-speech software that currently
exists. Here's an example of what it sounds like. Kevin Stratvert has a very small and growing
YouTube channel. He sounds a little robotic, but he delivers solid content. You should
consider subscribing. Now I don't know about you, but for me personally, that sounds
like an actual human read that text. The vocal emotion, the intonation,
it all just sounds so realistic. And in fact, it probably sounds even more
realistic or more human-like than I do. Let's check out how you can use this for free
right now. To start turning text into speech, head to the Eleven Labs homepage. You can click on the
card in the top right-hand corner, or you can also click on the link down below in the description.
Right here on the homepage, you don't even have to set up an account. You can just type in some
text and right here, you can select a voice to narrate that text. Here you have all of these
different pre-made voices and later on, we'll look at how you can even create your own customized
voice. You have a mix of males and also females. Once you're ready to generate your speech, you
can click on this play icon right down below. On the homepage, you are limited by how
many characters you can turn into speech. To be able to turn more text into speech, up
above, let's click on the pricing tab. This drops us onto the pricing page and here we see all
of the available plans. Now the great news is the base plan is entirely free. So here you could
test out speech synthesis to see how it works. Right down here on the free plan, you get
up to 10,000 characters per month that you can convert into speech, and I found that that
works out to about 10 minutes' worth of speech. One of the drawbacks of the free plan is you
cannot use it commercially and you also have to attribute back to Eleven Labs. The next plan up
is the starter plan and this one's attractively priced. For the first month, it's only $1 and then
it jumps up to $5, and the key difference is you get up to 30,000 characters per month that you can
convert into speech, and that works out to about 30 minutes. And one of the really great parts of
this is you get instant voice cloning. What this means is I can upload, let's say myself talking,
and then I can type, and it'll speak back using my voice. That's pretty incredible. In the top
right-hand corner, I'll click on sign in, but if you don't yet have an account, feel free to sign
up for a free account. Once you finish setting up your account, here we drop into speech synthesis
and this is where we can type in some text, and it'll generate speech based on that text. The
benefit of going through an account is you get a much larger quota to convert text into speech.
Right up above, just like we saw on the homepage, here you could select one of these voices and here
you could even preview what these voices sound like by clicking on the preview icon. You can go
through and listen to all the different options. Below that, you also have voice settings
and here you could adjust things like the stability. If you go down to the left where
it's more variable, this will cause the voice to sound more expressive. Here, you could also
modify the clarity and similarity enhancement. You could try adjusting these different values
to see how that affects the voice. Down below, here's where you can type in your text and the
more text you type in, the better the model will work, and the reason why is it adjusts the
delivery based on the context of the text. Here, for instance, I'll type in text.
Unfortunately, I have some bad news to report, so it's bad news, so the voice will
probably be a little bit more downbeat, and it'll factor that into the way it delivers the
voice. It's pretty incredible. Right down below, all I need to do is click on generate and then
we could listen to the voice. Unfortunately, I have some bad news to report. Mrs. Fields won the
best cookie of the year and not the Kevin Cookie Company. That sounded really human-like. I am
still just blown away by how good this technology has gotten. Right up above, I can change the
message and maybe we go with something a little bit more positive or upbeat. Actually, we were
just kidding. The Kevin Cookie Company won. Right up here, I can choose a different voice. Let's
try Elli to see how that one sounds. Down below, I'll click on generate. Haha, just kidding. Great
news to report. The Kevin Cookie Company has the best chocolate chip cookie. Yay, this is fantastic
news. That too also sounds fantastic. If you like the voice, but you weren't quite satisfied with
the delivery, up above, you could click into voice settings and you could try modifying these
different values or you can simply go back down to generate and you can simply regenerate, and you'll
get a slightly different variation. Let's listen again. Haha, just kidding. Great news to report.
The Kevin Cookie Company has the best chocolate chip cookie. Yay, this is fantastic news. So,
there you get the same voice, but it's slightly different and you could go through and generate
again and again until you find one that matches exactly how you want it to sound. Along with
the pre-made voices, if you scroll up the page, you also have the option to add your very own
voice. You can click on this or up on top, you can click into the voice lab. Within the voice lab,
you have two different options for making a new synthetic voice from scratch. Over here, you could
design a voice, or you can also clone a voice. Let's test this out by clicking on add voice.
Let's start by looking at voice design. This opens up a prompt where I can now generate my very
own voice. At the top, I can select a gender. I'll select male. I can also select the age. I'll
go with old. Right here, I can also select the accent. I'll select British and here I can define
the accent strength. Down below, I'll type in my text. I can confirm the Kevin Cookie Company does
indeed have the best cookies in the world. And down here, I'll click on generate. I can confirm
the Kevin Cookie Company does indeed have the best cookies in the world. How would I know? I'm old,
wise, and a British accent speaks with authority. We're going to have to use this in our next
marketing campaign. Here, I'll minimize this section on the bottom and then I can select use
voice. Here, I can give my voice a name. I'll type in a name and then click on create voice. Now,
anytime that I want to synthesize text to speech, I can use this specific voice. So, this might
become the voice of our marketing campaigns, and this way I can ensure that we always have the
same exact voice. Let's now add another voice, but this time we're going to use the instant
voice cloning. Here, I'll click on add voice. And right over here, let's select instant
voice cloning. I can now add my very own voice by uploading sample audio and what better
voice to try than my very own voice. Here, I'll type in Kevin for the voice name and here I
need to drag in sample audio. The good news is I record my voice a lot anytime I pull together
videos. So, I have lots of sample audio. Here, I’ll drag a file in. They recommend uploading at
least five minutes' worth of audio, providing more doesn't really help the model that much. Right
down here, I could also apply different labels and I need to enter a description. So, I might as well
say a handsome, amazing, incredible, brilliant young man with a beautiful sounding voice. I mean,
that's an accurate description, right? Down here, you also have to confirm that you have the rights
to this voice. This is my own voice. So of course, I have the rights to it. Down at the very bottom,
I'll click on add voice. And just like that, it has now added my voice and it went really fast. It
probably only took about 10 or 15 seconds to add my voice. Right down here, I can now use my voice.
Let's test this out. This brings me back to the speech synthesis page, and right here in the voice
list, we now see two new voices. Here, I have my voice, the Kevin voice, and then I also have that
other voice, the marketing campaign voice that I can choose from. I'll select my voice. I'll leave
the speech settings as is, and then right here, I'll type in my text, and I'll use the intro
of this video to see if this AI text to speech could potentially replace me. Down here, let's
click on generate and hear what it sounds like. Hi, everyone. Kevin here. Today, we are going to
look at how you can use the most realistic text to speech software that currently exists. Here's
an example of what it sounds like. Now personally, I don't think it sounds like me that much, but let
me know what you think down below in the comments. Did that sound like me or did it sound like a
computer was speaking it? Once you've converted text into speech, up at the very top, click into
the history tab, and here you can see a history of all of your different generated samples. You can
play them back again to hear what they sound like, and if you'd like to download any of the generated
speech, you can click on this icon and then you can download the selected speech. I am just blown
away by how good text to speech has gotten. The next time you listen to an audio book, will it be
an actual human or a computer narrating the book? And will you be able to tell the difference? I
don't know. To watch more videos like this one, please consider subscribing and
I’ll see you in the next video.