4DHJ6YeLOWI

[Music]
hello there and welcome to another
tutorial my name is tan May bakshi and
this time we're going to be going over
using the Nuance speech kit in Swift
with iOS
well to begin this was actually a
question sent to me by one of my
subscribers
on clip now this is tarabala Christian
one of my subscribers
days ago she on a YouTube comment uh on
my NS speech recognizer tutorial
actually asked me how we can actually
use the Nuance speech SDK for iOS with
swift so that's the tutorial I'm
creating today I hope this helps you out
Tara
season sent by Tara uh let's actually
get into the Mac part
I will be explaining exactly how we can
use the Nuance speech kit with swift too
so uh let's get right into it shall we
so welcome back to the Mac part and now
I'm going to be showing you uh well I
mean just before this super simple and
short code I'm going to show showing you
a demo of the app that I've made for
Tara so let's begin as you can see this
is my iPhone mirrored onto my Mac
alright so as you can see what I'm gonna
do is I am going to say a sentence into
this application and let's see if it can
convert it to text so as you can see
there is a listen button right over at
the bottom here so we're going to click
this this will turn into listening then
the label over here will uh tell us what
the result is uh right as we're done
saying things now this will
automatically detect when you are done
talking and so you don't need to click
any sort of stop button which is just
great again Siri is actually powered by
Nuance and so that's actually a really
neat fun fact But continuing though
let's uh begin so let's actually give it
something pretty simple like
I have a YouTube channel comma in which
I like to teach things period
as you can see it says I have a YouTube
channel in which I like to teach things
alright so that was pretty simple let's
try something else though as you can see
I'm going to click on listen and it
turns into listening
I have created this sample app for one
of my subscribers
as you can see it said I have created
this sample app for one of my
subscribers great as you can see Nuance
is actually pretty accurate now there's
one thing that I am going to save for
the last which is actually just
fascinating uh but then again the best
examples are going to be saved for the
last uh all right so now let's continue
though let's begin with the UI shall we
so as you can see the UI is even simpler
than the code the listen button has an
IB action on the touch up inside which
has button clicked can't get simpler
than that button click gets executed
right as you click a button done okay
referencing Outlets we have an IB Outlet
that's called a record button connected
to the button that's it all right but
the label is a bit different the label
is actually connected only to an IB
Outlet that's called result label okay
that's the label now as you can see in
the code we actually have those two IB
Outlets uh record button which is a UI
button and result label which is a UI
label now as you can see view controller
doesn't just in uh actually let me see
UI view controller uh is a class so it
doesn't just inherit from UI view
controller but it actually conforms to
the SK transaction delegate protocol and
in case you were wondering actually if
you were to click on command on your
keyboard actually I promise I'm clicking
command on my keyboard and if you move
around your mouse you can actually see
these things convert into links whenever
you hover over something and so
basically if I hover over SK transaction
delegate and I click on it and then I
release command and the click as you can
see basically it shows us where this is
defined and as you can see right now
this is actually a public protocol of SK
transaction delegate in the nuanced
speech kit framework and so that's why I
know that this is a protocol so I can
say that this can then new controller
conforms to it not in the hair it's from
if it were to inherit from that would be
a class conforms to linear protocol all
right so I'm going to explain exactly
what the difference is in just a second
but then again that's an entirely
separate video uh but I will explain in
minor detail right now in just a second
then let's look at the IB action
function called button clicked now what
does this do basically we're going to
take the record button and we're going
to set its title to listening and for
State normal you don't need to worry
about that this that's just telling
record button that it's still a normal
button it's it's not weird
um
so basically uh the record buttons title
is now listening instead of listen
because you've clicked on it but then
after that we actually create a session
which is equal to a new speech kit
session or SK session for short if we go
into this this is inside of speech kit
and so basically what we're doing here
is we're creating that SK session with a
URL now basically the hard part is comes
in you want you're going to want to
actually go to the link in the
description uh the link in the
description basically brings you to a
page in Nuance that allows you to
actually register for an API key and so
once you've done that uh you will
actually get something called a Sandbox
credential now basically this is a
development credential uh and when
you're sort of developing your app with
Nuance you want to use these credentials
because they're absolutely free and who
doesn't like free stuff however what
this is not fit for production use okay
you want something that's paid when you
actually actually use it for a sort of
real life when it's on the App Store
basically and so basically uh you're
just going to want to pay when you're
actually uh using it for production but
right now when you're developing it I
guess uh the free credentials work
perfectly fine and they're called
sandbox credentials uh so basically what
you're going to want to do is take those
sandbox credentials and I mean I'm using
speech kit too but you could also use
speech kit one but I don't see why you
would do that since they have newer
versions released now and so basically
what you're going to do is you're going
to take the link that you get on these
speech kit too
um
under speaker two you'll see a link and
you're going to take that link and it'll
look something like this except instead
of tagumanny gmail.com it'll be your own
email and then lots of other random
letters along with the date that you
registered I assume this is I Don't Know
But continuing though it'll just give
you a link and you want to copy that and
paste it into this string okay I have
already done that then it's also going
to give you an application token near
the end uh but before speech kit one and
you're going to want to copy that into
here as well once you're done you're
going to actually type in another line
of code and this will essentially tell
the session that's uh and so the session
is basically connecting to the Nuance
SDK in the cloud and so basically what
you're going to do is you're going to
say you know what start listening
through the microphone for some sort of
uh user voice and so basically we're
going to say session dot recognize with
type SK transaction speech type
dictation what does this mean what do
these Ren letters mean they basically
just mean that Nuance will listen for
users voice and convert it into uh text
it'll dictate it okay
or the user will dictate to it basically
okay so that's what that is you don't
really need to worry about that too much
I don't know why they haven't created
enumeration for this but they haven't
for some reason now they could have done
SK transaction.speech type dot or speech
type dot dictation sorry uh which would
have been a bit cleaner way of doing it
but whatever uh continuing though then
we set the detection to long now this
can also be short or none uh and so
basically long means that the user will
be able to literally say in paragraphs
with punctuation and stuff but you could
also set this to short and that would be
like a sentence for like a quick text
message and you could even set this To
None now the reason I'm not doing none
is because that would require an
entirely separate function from that
delicate that were or that delegate that
we are conforming to SK transaction
delegate and so basically we would have
to sort of say when the user would click
a stop button uh in order for nuance to
actually stop listening for the user's
voice and but with detection it'll
automatically realize when the user is
done saying stuff and will automatically
just sort of stop listening and sort of
tell us the answer so that's why I use
detection and so that's why I also like
to keep it on long because with long you
can really say short stuff as well it
just isn't as good at it but with short
you can't really say it as long of
things so I'm just going to keep it that
long you can always change this if you'd
like and if you need help with the dot
none option please do tell me I will
have my contact info at the end of the
video but continuing dog
and then we set the language to English
USA now again you can use whatever
language you have I assume there is a
link in the documentation for the
nuanced speech kit SDK and so basically
they would describe sort of the
languages that are actually compatible
with nuance and so if they do have a
page like that if I'm able to find it it
will be in the description I hope I find
it for you that's uh basically then
we've set the delicate uh the delegate I
don't know why it's in delicate delegate
uh then we set the delegates of that
session to sell because we are
conforming to that protocol and that
there has to be a reason for it and so
this is why we're actually doing that
all right so that's why yeah that's
basically what we're gonna do there
however
what happens when Nuance is done sort of
getting the answer for us well that's
where the next function comes in now
to a protocol the protocol will give us
optional functions that we can or
mandatory functions that we can include
into our code in order to add some
functionality for example this sessions
delegate has been set to sell and so now
what it's going to do is since we
conform to this delegate and we're
setting ourselves to that sort of a
delegate of the session
it'll know that when we're done we have
to contact view controller
transaction function now there are many
transaction functions in this protocol
however one of them which is uh quite
interesting called didn't receive
recognition this one will actually be
called when the Nuance SDK is done sort
of recognizing what the user said in
transcribing it and so basically what we
do here is we actually get an SK
recognition from the nuance svk
after that what we want to do is we want
to take the highest scoring candidate
from that entire list of uh sort of
candidate answers of what uh the what
the user could have said we want to take
the highest scoring candidate and we
want to put that into a string okay
we're going to put that into one string
and put that into the result labels text
um we're going to set that as the text
of result label which is our UI label
that we uh declared up here all right so
once we've done that we've already given
our user uh the answer for them then
what we're going to do is we're going to
set the record buttons title back to
listen because we're done listening now
we're not listening to the user anymore
and so we're going to set the button
back to listen
that's it all right so that was pretty
much uh all we had to uh describe for
the code but now as I said I'd like to
show you one more thing which is
actually absolutely fascinating if I
were to actually say to this app
okay I have to get ready this is a lot
of stuff to say all right
this app is running on an iPhone comma
which runs iOS comma however comma my
Mac comma which is an iMac comma runs
Mac OS X comma which runs xcode comma
okay great so as you can see uh what has
happened is it says transcribe the
entire thing uh into this app is running
on the iPhone which runs iOS however my
Mac which is an iMac runs Mac Oh Mac OS
X Mac OS X which runs xcode now this
usually works if I were to say
Mac OS X runs on my iMac
and for some reason it gives us m-a-c-k
don't know why it was working a few
minutes ago anyway uh so I mean that's a
problem with the Nuance SDK not with our
code so I mean that would be fixed soon
I guess it keeps changing I guess I mean
it's machine learning so it would have
kept learning from people's uh responses
to the SDK but then again whatever
continuing that's gonna be it for this
tutorial I really hope this uh you
enjoyed and I really hope this was
helpful for your project Tara uh if you
have any more if uh you have any more
questions with your projects uh
regarding the iOS development I'd be
glad to help and everyone else if you
would also require help uh you can
actually comment down below uh with your
question suggestion or feedback and I'd
be happy to get back to you you can even
email me at tajimani gmail.com and you
can also tweet to me at tajimani that's
gonna be it for this video If you
enjoyed please leave a like subscribe to
my channel if you really liked it and
even share the video that would really
help out a lot uh in fact I actually
just hit 1 000 subscribers a few days
ago hooray but uh continuing though hope
you enjoyed Please Subscribe that's
gonna be it for this video goodbye
[Music]