Hello everyone today I'm going to
show you how easy it is to set up and then generate AI voiceovers through
text to speech in Microsoft Azure. This is available with a free account.
So first of all go to portal.azure.com and sign in with your microsoft account.
you can create one if needed Once signed in, you'll be in the main portal page From here, we go to the search bar
at the top of the page and type in speech and speech services will appear
under "services" in the drop down menu Click
speech and you will eb taken to
"Azure AI Services | Speech Service" Now click "Create"
Choose your subscription, and a resource group. If you don't have any
resource groups yet, create one. The name here is not important, a resource group is just
a container for services. Now, choose a region. Typically here we
want to choose a fairly local region, however, for this example, I'm choosing North
Central US as it has more AI voice offerings. Give your speech service a name. It will be
conve
rted into a URL so does need some uniqueness. You'll be told if the name is already in use.
In Pricing tier, you can choose free to make use of the free services, or, if you have a billing
mechanism, you can choose the standard tier, which offers more options. Be
aware of the pricing details link if you choose the paid-for version.
We'll be using the Free tier in this video. here I click create for some validation and
then click network to visit the network tab. I change to selected networks a
s I want our
speech environment to be restricted to me. As I was editing this video I remembered
that virtual networks may not be available to free users, so, you can leave
this on all networks for now and we can come back to this after the speech service is created.
Click next to proceed. Now click next through identity and tags then
click review and create, followed by create. Azure will now create your speech service.
Once created, click to go to resource group to see your new speech serv
ice in the
list. It may take a refresh to load Once you can see your speech
service, click on it. Here you can see some details about
your speech service. The overview page. Now I'll take you through setting up
the network restrictions. So click on "networking".
on this page we want to choose "selected networks and private endpoints" and where it says
"firewall" at the bottom click the box next to "Add your client IP address"
Then click "save". Now, we can go back
to the overview
screen and "go to speech studio" Ok In the speech studio page, scroll
down to the Text To Speech section and click on "Try out voice gallery"
The voice gallery will now load. If you see any error messages like I have on
the screen, wait a minute or so before reloading the page. It can take a little while
for the firewall changes we just made to update. Refresh the page and eventually
it will load in properly With the page loaded, we'll now see the AI
voice catalog. We can c
lick a voice name to select it and then, over to the right, where it
says default, we can click to play a sample. "The second step will follow at least five weeks
later. Coming no sooner than the 12th April" You'll see that there are a lot of voices to
choose from here, some are marked as OpenAI, we can click one to sample it. These may
not be available in your chosen region. You can sample some of the
other voices here as well. When you've chosen a voice
to use, click Try it out. Here you
can see that the OpenAI voice we have
selected is not available on the free tier. Other voices are though.
So, we'll click Sonia, and we'll type in some text
And we'll choose to play that in the cheerful style "This is a test of
Sonia, she is in the free resource layer" Now let's click "Go to audio content
creation" to generate some real audio
In the Speech Studio, Audio content creation is where we can type or paste
the text we want our AI
voice to say Here I'm typing in a few sentences
Above the text window, we can set a quality level for the output
Over to the right is where we choose from a number of options to change the voice and style
we have to assign a voice to each sentence We can click play and it will play the
sentences from the selected paragraph onwards "Hi this is what I want Sonia to say" "This
is the second sentence" "Thanks for watching" We can introduce punctuation to
affect the cadence of the speech "Hi, thi
s is what I want Sonia to say" "This is
the second sentence" "Thanks for watching" If needed, we can change to a different voice notice the voice change affects
the selected paragraph only "Hi, this is what I want Ryan to
say" "This is the second sentence" I'll change some more text
"Please leave a like" Now, I'll save the current
text to the cloud storage Finally, I'll show you how to export
the sound files. You can export to the cloud audio library, or to local disk on
your computer. I'
ll save to local disk Here you can change the output file format,
as well as choosing how many files the output is broken up in to.
Click export, and the files will be downloaded by your browser.
And, with those files download, you now have your AI voice over
Let's open them up and have a listen "Hi, this is what I want Ryan to say"
"Please leave a like" I hope that this video was helpful, please leave
a like and subscribe for more content. Thanks
Comments