TTS(Text To Speech 文本转语音缩写)

简介

涉及到AVFoundatio框架中的AVSpeechSynthesizer、AVSpeechUtterance、AVSpeechSynthesisVoice三个类，这三个类的声明都在AVSpeechSynthesis.h里，这三个类比较简单。

需要用到的一些枚举与宏

• 播放暂停时是立即停止播放语音还是播放完当前词再暂停。

  typedef NS_ENUM(NSInteger, AVSpeechBoundary) {
      AVSpeechBoundaryImmediate,
      AVSpeechBoundaryWord
  }
  

• 语音质量

  typedef NS_ENUM(NSInteger, AVSpeechSynthesisVoiceQuality) {
      AVSpeechSynthesisVoiceQualityDefault = 1,
      AVSpeechSynthesisVoiceQualityEnhanced
  }
  

• 速率

  AVF_EXPORT const float AVSpeechUtteranceMinimumSpeechRate NS_AVAILABLE_IOS(7_0);
  AVF_EXPORT const float AVSpeechUtteranceMaximumSpeechRate NS_AVAILABLE_IOS(7_0);
  AVF_EXPORT const float AVSpeechUtteranceDefaultSpeechRate NS_AVAILABLE_IOS(7_0);
  

• 注意：设置属性字符串时，AVSpeechSynthesisIPANotationAttribute的正确使用，最后会有Demo传送门

  AVF_EXPORT NSString *const AVSpeechSynthesisIPANotationAttribute;
  

AVSpeechUtterance

要开始说话，请指定AVSpeechSynthesisVoice和要说出的字符串，然后根据需要随意更改速率，音高或音量。

四种初始化方法

+ (instancetype)speechUtteranceWithString:(NSString *)string;
+ (instancetype)speechUtteranceWithAttributedString:(NSAttributedString *)string API_AVAILABLE(ios(10.0), watchos(3.0), tvos(10.0));

- (instancetype)initWithString:(NSString *)string;
- (instancetype)initWithAttributedString:(NSAttributedString *)string API_AVAILABLE(ios(10.0), watchos(3.0), tvos(10.0));

通过以上创建出来的实例，其语音、语速、音调、音量都是默认的，可以通过属性进行定制，但是在入队后设置这些值将不起作用。其中rate的值应该介于AVSpeechUtteranceMinimumSpeechRate和AVSpeechUtteranceMaximumSpeechRate之间

@property(nonatomic, retain, nullable) AVSpeechSynthesisVoice *voice;
@property(nonatomic) float rate;           
@property(nonatomic) float pitchMultiplier;  // [0.5 - 2] Default = 1
@property(nonatomic) float volume;           // [0-1] Default = 1
@property(nonatomic) NSTimeInterval preUtteranceDelay;    // Default is 0.0
@property(nonatomic) NSTimeInterval postUtteranceDelay;   // Default is 0.0

有两个属性可以获得当前播放的字符串或属性字符串

@property(nonatomic, readonly) NSString *speechString;
@property(nonatomic, readonly) NSAttributedString *attributedSpeechString

AVSpeechSynthesisVoice

通过指定应该在其中说出文本的语言代码来检索语音，或者通过使用voiceWithIdentifier来获取已知语音标识符。

现支持的语音语言37种,可以用类方法speechVoices获取

+ (NSArray<AVSpeechSynthesisVoice *> *)speechVoices;

获取设备当前语言语音码

+ (NSString *)currentLanguageCode;

根据指定的语言语音码获取语音

+ (nullable AVSpeechSynthesisVoice *)voiceWithLanguage:(nullable NSString *)languageCode;

根据指定的语言语音标识码获取语音 iOS9.0之后可用

+ (nullable AVSpeechSynthesisVoice *)voiceWithIdentifier:(NSString *)identifier NS_AVAILABLE_IOS(9_0);

获取某个语音的信息，只读

@property(nonatomic, readonly) NSString *language;
@property(nonatomic, readonly) NSString *identifier NS_AVAILABLE_IOS(9_0);
@property(nonatomic, readonly) NSString *name NS_AVAILABLE_IOS(9_0);
@property(nonatomic, readonly) AVSpeechSynthesisVoiceQuality quality NS_AVAILABLE_IOS(9_0);

AVSpeechSynthesizer

合成器

是否正在播放、暂停

@property(nonatomic, readonly, getter=isSpeaking) BOOL speaking;
@property(nonatomic, readonly, getter=isPaused) BOOL paused;

将已加入或正在讲话的相同AVSpeechUtterance加入会引发异常。

- (void)speakUtterance:(AVSpeechUtterance *)utterance;

以下方法只对正在讲话的语音进行操作才有效。 如果成功则返回YES，失败则返回NO。

- (BOOL)stopSpeakingAtBoundary:(AVSpeechBoundary)boundary;
- (BOOL)pauseSpeakingAtBoundary:(AVSpeechBoundary)boundary;
- (BOOL)continueSpeaking;

指定要用于合成语音的音频通道，如AVAudioSession当前路径中的通道描述所述。
语音音频将复制到每个指定的频道。默认值为nil，表示系统默认值。

@property(nonatomic, retain, nullable) NSArray<AVAudioSessionChannelDescription *> *outputChannels;

学习资源

#import "ViewController.h"
#import <AVFoundation/AVFoundation.h>

@interface ViewController ()

@property(nonatomic, strong) AVSpeechSynthesizer *speechSynthesizer;
@property(nonatomic, strong) AVSpeechUtterance *utterance;
@property(nonatomic, strong) UILabel *label;

@end

@implementation ViewController

- (void)viewDidLoad {
    [super viewDidLoad];
    // Do any additional setup after loading the view, typically from a nib.
    self.speechSynthesizer = [[AVSpeechSynthesizer alloc] init];
    [self commonInit];
}


- (void)didReceiveMemoryWarning {
    [super didReceiveMemoryWarning];
    // Dispose of any resources that can be recreated.
}

- (void)commonInit {
    UISwitch *senderSwitch = [[UISwitch alloc] init];
    senderSwitch.on = YES;
    senderSwitch.frame = CGRectMake(0, 0, 100, 20);
    senderSwitch.center = self.view.center;
    [senderSwitch addTarget:self action:@selector(updateUtterance:) forControlEvents:UIControlEventValueChanged];
    [self.view addSubview:senderSwitch];
    
    self.label = [[UILabel alloc] initWithFrame:CGRectMake((self.view.bounds.size.width - 200)/2.0f, CGRectGetMaxY(senderSwitch.frame) - 80, 200, 20)];
    self.label.textAlignment = NSTextAlignmentCenter;
    [self.view addSubview:self.label];
    
    UIButton *btn = [UIButton buttonWithType:UIButtonTypeCustom];
    btn.frame = CGRectMake((self.view.bounds.size.width - 128)/2.0f, CGRectGetMaxY(senderSwitch.frame) + 20, 128, 128);
    [btn setImage:[UIImage imageNamed:@"audio"] forState:UIControlStateNormal];
    [btn addTarget:self action:@selector(btnClick:) forControlEvents:UIControlEventTouchUpInside];
    [self.view addSubview:btn];
    
    [self updateUtterance:senderSwitch];
}

- (void)updateUtterance:(UISwitch *)sender {
    sender.on = !sender.isOn;
    if (sender.on) {
        NSMutableAttributedString *attString = [[NSMutableAttributedString alloc] initWithString:@"Tsutsumi"];
        // 这里的国际音标可以随意修改 eg:"tən.tən.mə"  "tən.mə" ... 可以多试试其他的情况
        [attString addAttribute:AVSpeechSynthesisIPANotationAttribute value:@"tən.tən.mi" range:NSMakeRange(0, attString.length)];
        self.utterance = [AVSpeechUtterance speechUtteranceWithAttributedString:attString];
        
        self.label.text = @"Tsutsumi(Attributed)";
    } else {
        self.utterance = [AVSpeechUtterance speechUtteranceWithString:@"Tsutsumi"];
        self.label.text = @"Tsutsumi";
    }
}

- (void)btnClick:(UIButton *)sender {
    if (self.speechSynthesizer.isSpeaking) {
        NSLog(@"正在朗读中....");
        return;
    }
    [self.speechSynthesizer speakUtterance:self.utterance];
}

@end

想看Demo的同学请移驾-->TTS