iOS10新特性之SFSpeechRecognizer的使用

SFSpeechRecognizer是iOS10新特性之一，用来进行语音识别的。

1、使用前需要导入头文件

import Speech

• 1

2、设置请求权限

//请求权限
        SFSpeechRecognizer.requestAuthorization { (state) in
            print(state)
        }

• 1
• 2
• 3
• 4

3、配置info.plist

//添加一条权限
Privacy - Speech Recognition Usage Description   -> “xxx”

• 1
• 2

4、初始化识别器

//初始化一个识别器
let recognizer = SFSpeechRecognizer(locale: Locale(identifier: "zh_CN"))

• 1
• 2
• 3

5、初始化请求

 //path
 let url = Bundle.main.url(forResource: "bg", withExtension: "mp3")

 //初始化一个识别请求
 let request = SFSpeechURLRecognitionRequest(url: url!)

• 1
• 2
• 3
• 4
• 5
• 6

6、开启识别任务

//开始一个识别请求
recognizer?.recognitionTask(with: request, resultHandler: { (result, error) in
     if error != nil{
                print("识别错误")
                return
     }
     else{
       print(result?.bestTranscription.formattedString)
         }
    })

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10

完成代码：

import UIKit
import Speech

class ViewController: UIViewController {

    override func viewDidLoad() {
        super.viewDidLoad()

        //请求权限
        SFSpeechRecognizer.requestAuthorization { (state) in
            print(state)
        }
    }

    override func touchesBegan(_ touches: Set, with event: UIEvent?) {
        //初始化一个识别器
        let recognizer = SFSpeechRecognizer(locale: Locale(identifier: "zh_CN"))

        //path
        let url = Bundle.main.url(forResource: "bg", withExtension: "mp3")

        //初始化一个识别请求
        let request = SFSpeechURLRecognitionRequest(url: url!)

        //开始一个识别请求
        recognizer?.recognitionTask(with: request, resultHandler: { (result, error) in
            if error != nil{
                print("识别错误")
                return
            }
            else{
                print(result?.bestTranscription.formattedString)
            }
        })

    }
}

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33
• 34
• 35
• 36
• 37
• 38
• 39