With the appearance of machine studying and synthetic intelligence, the iOS SDK already comes with a variety of frameworks for builders to develop apps with machine learning-related options. On this tutorial, let’s discover two built-in ML APIs for changing textual content to speech and performing language detection.
Utilizing Textual content to Speech in AVFoundation
Let’s say, if we’re constructing an app that reads customers’ enter message, it’s worthwhile to implement some form of text-to-speech features.
The AVFoundation
framework has include some text-to-speech APIs. To make use of these APIs, now we have to first import the framework:
Subsequent, create an occasion of AVSpeechSynthesizer
:
let speechSynthesizer = AVSpeechSynthesizer() |
To transform the textual content message to speech, you possibly can write the code like this:
speechSynthesizer.communicate(utterance)
let utterance = AVSpeechUtterance(string: inputMessage) utterance.pitchMultiplier = 1.0 utterance.price = 0.5 utterance.voice = AVSpeechSynthesisVoice(language: “en-US”)
speechSynthesizer.communicate(utterance) |
You create an occasion of AVSpeechUtterance
with the textual content for the synthesizer to talk. Optionally, you possibly can configure the pitch, price, and voice. For the voice parameter, we set the language to English (U.S.). Lastly, you cross the utterance
object to the speech synthesizer to learn the textual content in English.
The built-in speech synthesizer is able to talking a number of languages comparable to Chinese language, Japanese, and French. To inform the synthesizer the language to talk, you need to cross the proper language code when creating the occasion of AVSpeechSynthesisVoice
.
To search out out all of the language codes that the machine helps, you possibly can name up the speechVoices()
methodology of AVSpeechSynthesisVoice
:
for voice in voices {
print(voice.language)
}
let voices = AVSpeechSynthesisVoice.speechVoices()
for voice in voices { print(voice.language) } |
Listed here are among the supported language codes:
- Japanese – ja-JP
- Korean – ko-KR
- French – fr-FR
- Italian – it-IT
- Cantonese – zh-HK
- Mandarin – zh-TW
- Putonghua – zh-CN
In some circumstances, you could must interrupt the speech synthesizer. You’ll be able to name up the stopSpeaking
methodology to cease the synthesizer:
speechSynthesizer.stopSpeaking(at: .instant) |
Performing Language Identification Utilizing Pure Language Framework
As you possibly can see within the code above, now we have to determine the language of the enter message earlier than the speech synthesizer can convert the textual content to speech appropriately. Wouldn’t or not it’s nice if the app can mechanically detect the language of the enter message?
The NaturalLanguage
framework supplies a wide range of pure language processing (NLP) performance together with language identification.
To make use of the NLP APIs, first import the NaturalLanguage
framework:
You simply want a pair strains of code to detect the language of a textual content message:
let languageRecognizer = NLLanguageRecognizer() languageRecognizer.processString(inputMessage) |
The code above creates an occasion of NLLanguageRecognizer
after which invokes the processString
to course of the enter message. As soon as processed, the language recognized is saved within the dominantLanguage
property:
if let dominantLanguage = languageRecognizer.dominantLanguage { print(dominantLanguage.rawValue) } |
Here’s a fast instance:

For the pattern, NLLanguageRecognizer
acknowledges the language as English (i.e. en
). In case you change the inputMessage
to Japanese like under, the dominantLanguage
turns into ja
:
let inputMessage = “これはテストです” |
The dominantLanguage
property could don’t have any worth if the enter message is like this:
let inputMessage = “1234949485🎃😸💩🥸” |
Wrap Up
On this tutorial, now we have walked you thru two of the built-in machine studying APIs for changing textual content to speech and figuring out the language of a textual content message. With these ML APIs, you possibly can simply incorporate the text-to-speech characteristic in your iOS apps.
Founding father of AppCoda. Writer of a number of iOS programming books together with Starting iOS Programming with Swift and Mastering SwiftUI. iOS App Developer and Blogger. Observe me at Fb, Twitter and Google+.