iOS 讯飞语音的使用

一、下载讯飞的SDK

二、集成流程

第一步：获取appid（进入控制台查看）

第二步：工程配置

1、将模板项目的lib文件拉入项目，并加入依赖库

注意： 添加iflyMSC.framework时，请检查工程BuildSetting中的framwork path的设置，如果出现找不到framework的情况，可以将path清空，在Xcode中删除framework，然后重新添加。 如果使用了离线识别，还需要增加libc++.tbd。 如果使用aiui，需要添加libc++.tbd，libicucord.tbd。

2、设置Bitcode

在Xcode 7,8默认开启了Bitcode，而Bitcode 需要工程依赖的所有类库同时支持。MSC SDK暂时还不支持Bitcode，可以先临时关闭。后续MSC SDK支持Bitcode 时，会在讯飞开放平台上进行SDK版本更新，QQ支持群中同时会有相关提醒，请关注。关闭此设置，只需在Targets - Build Settings 中搜索Bitcode 即可，找到相应选项，设置为NO。

3、用户隐私权限配置

iOS 10发布以来，苹果为了用户信息安全，加入隐私权限设置机制，让用户来选择是否允许。 隐私权限配置可在info.plist 新增相关privacy字段，MSC SDK中需要用到的权限主要包括麦克风权限、联系人权限和地理位置权限：

NSMicrophoneUsageDescription

NSLocationUsageDescription

NSLocationAlwaysUsageDescription

NSContactsUsageDescription

即在Info.plist 中增加下图设置：

第1种设置方法：

第2种设置方法：

第三步：初始化

初始化示例：

//可在启动时添加，Appid是应用的身份信息，具有唯一性，初始化时必须要传入Appid。

NSString *initString = [[NSString alloc] initWithFormat:@"appid=%@", @"YourAppid"];

[IFlySpeechUtility createUtility:initString];

注意： 初始化是一个异步过程，可放在App启动时执行初始化，具体代码可以参照Demo的MSCAppDelegate.m。

还需要在工程中添加结果解析文件，否则编译会报错。

文件在Demo中的位置：MSCDemo -> tools -> jsoncpp

还需要从模板项目中拖拽IATConfig类和ISRDataHelper类到自己的项目中

、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、

三、下面就是具体代码操作

1、语音在线合成

本示例对应Demo的TTSUIController文件，为在线合成的代码示例。

//包含头文件
#import "iflyMSC/IFlyMSC.h"
//需要实现IFlySpeechSynthesizerDelegate合成会话的服务代理
@interface TTSUIController : UIViewController<IFlySpeechSynthesizerDelegate>
@property (nonatomic, strong) IFlySpeechSynthesizer *iFlySpeechSynthesizer;
@end

//获取语音合成单例
_iFlySpeechSynthesizer = [IFlySpeechSynthesizer sharedInstance];
//设置协议委托对象
_iFlySpeechSynthesizer.delegate = self;
//设置合成参数
//设置在线工作方式
[_iFlySpeechSynthesizer setParameter:[IFlySpeechConstant TYPE_CLOUD]  forKey:[IFlySpeechConstant ENGINE_TYPE]];
//设置音量，取值范围 0~100
[_iFlySpeechSynthesizer setParameter:@"50" forKey: [IFlySpeechConstant VOLUME]];
//发音人，默认为”xiaoyan”，可以设置的参数列表可参考“合成发音人列表”
[_iFlySpeechSynthesizer setParameter:@" xiaoyan "
 forKey: [IFlySpeechConstant VOICE_NAME]];
//保存合成文件名，如不再需要，设置为nil或者为空表示取消，默认目录位于library/cache下
[_iFlySpeechSynthesizer setParameter:@" tts.pcm" forKey: [IFlySpeechConstant TTS_AUDIO_PATH]];
//启动合成会话
[_iFlySpeechSynthesizer startSpeaking: @"你好，我是科大讯飞的小燕"];

//IFlySpeechSynthesizerDelegate协议实现
//合成结束
- (void) onCompleted:(IFlySpeechError *) error {}
//合成开始
- (void) onSpeakBegin {}
//合成缓冲进度
- (void) onBufferProgress:(int) progress message:(NSString *)msg {}
//合成播放进度
- (void) onSpeakProgress:(int) progress beginPos:(int)beginPos endPos:(int)endPos {}

2、语音听写（带界面的例子） IFlySpeechRecognizer是不带界面的语音听写控件，IFlyRecognizerView是带界面的控件，此处仅介绍不带界面的语音听写控件。在ViewController.m中写的，可以直接测试

使用示例如下所示：

//
//  ViewController.m
//  文字语音读写
//
//  Created by User on 2017/7/7.
//  Copyright © 2017年 User. All rights reserved.
//

#import "ViewController.h"
#import "IATConfig.h"
#import "ISRDataHelper.h"

#import "iflyMSC/iflyMSC.h"

@interface ViewController ()<IFlyRecognizerViewDelegate>

@property (nonatomic, strong) IFlySpeechRecognizer *iFlySpeechRecognizer;//语音转文字

//（讯飞自带的界面View）
@property (nonatomic, strong) IFlyRecognizerView *iflyRecognizerView;//Recognition control with view

@property (nonatomic, strong) NSMutableArray *resultARR;

@end

@implementation ViewController

#pragma mark - Initialization

/**
 initialize recognition conctol and set recognition params
 **/
-(void)initRecognizer
{
    NSLog(@"%s",__func__);
    
        //recognition singleton with view
        if (_iflyRecognizerView == nil) {
            
            _iflyRecognizerView= [[IFlyRecognizerView alloc] initWithCenter:self.view.center];
        }
        
        [_iflyRecognizerView setParameter:@"" forKey:[IFlySpeechConstant PARAMS]];
        
        //set recognition domain
        [_iflyRecognizerView setParameter:@"iat" forKey:[IFlySpeechConstant IFLY_DOMAIN]];
        
        
        _iflyRecognizerView.delegate = self;
        
        if (_iflyRecognizerView != nil) {
            IATConfig *instance = [IATConfig sharedInstance];
            //set timeout of recording
            [_iflyRecognizerView setParameter:instance.speechTimeout forKey:[IFlySpeechConstant SPEECH_TIMEOUT]];
            //set VAD timeout of end of speech(EOS)
            [_iflyRecognizerView setParameter:instance.vadEos forKey:[IFlySpeechConstant VAD_EOS]];
            //set VAD timeout of beginning of speech(BOS)
            [_iflyRecognizerView setParameter:instance.vadBos forKey:[IFlySpeechConstant VAD_BOS]];
            //set network timeout
            [_iflyRecognizerView setParameter:@"20000" forKey:[IFlySpeechConstant NET_TIMEOUT]];
            
            //set sample rate, 16K as a recommended option
            [_iflyRecognizerView setParameter:instance.sampleRate forKey:[IFlySpeechConstant SAMPLE_RATE]];
            
            //set language
            [_iFlySpeechRecognizer setParameter:instance.language forKey:[IFlySpeechConstant LANGUAGE]];
            //set accent
            [_iFlySpeechRecognizer setParameter:instance.accent forKey:[IFlySpeechConstant ACCENT]];
            //set whether or not to show punctuation in recognition results
            [_iflyRecognizerView setParameter:instance.dot forKey:[IFlySpeechConstant ASR_PTT]];
        
        if([[IATConfig sharedInstance].language isEqualToString:@"en_us"]){
            if([IATConfig sharedInstance].isTranslate){
                [self translation:NO];
            }
        }
        else{
            if([IATConfig sharedInstance].isTranslate){
                [self translation:YES];
            }
        }
        
    }

}



-(void)translation:(BOOL) langIsZh
{
    
    if ([IATConfig sharedInstance].haveView == NO) {
        [_iFlySpeechRecognizer setParameter:@"1" forKey:[IFlySpeechConstant ASR_SCH]];
        
        if(langIsZh){
            [_iFlySpeechRecognizer setParameter:@"cn" forKey:@"orilang"];
            [_iFlySpeechRecognizer setParameter:@"en" forKey:@"translang"];
        }
        else{
            [_iFlySpeechRecognizer setParameter:@"en" forKey:@"orilang"];
            [_iFlySpeechRecognizer setParameter:@"cn" forKey:@"translang"];
        }
        
        [_iFlySpeechRecognizer setParameter:@"translate" forKey:@"addcap"];
        
        [_iFlySpeechRecognizer setParameter:@"its" forKey:@"trssrc"];
    }
    else{
        [_iflyRecognizerView setParameter:@"1" forKey:[IFlySpeechConstant ASR_SCH]];
        
        if(langIsZh){
            [_iflyRecognizerView setParameter:@"cn" forKey:@"orilang"];
            [_iflyRecognizerView setParameter:@"en" forKey:@"translang"];
        }
        else{
            [_iflyRecognizerView setParameter:@"en" forKey:@"orilang"];
            [_iflyRecognizerView setParameter:@"cn" forKey:@"translang"];
        }
        
        [_iflyRecognizerView setParameter:@"translate" forKey:@"addcap"];
        
        [_iflyRecognizerView setParameter:@"its" forKey:@"trssrc"];
    }
    
    
}



- (void)viewDidLoad {
    [super viewDidLoad];
    _resultARR = [[NSMutableArray alloc]init];
    
    if(_iflyRecognizerView == nil)
    {
        [self initRecognizer ];
    }
    
    //Set microphone as audio source
    [_iflyRecognizerView setParameter:IFLY_AUDIO_SOURCE_MIC forKey:@"audio_source"];
    
    //Set result type
    [_iflyRecognizerView setParameter:@"plain" forKey:[IFlySpeechConstant RESULT_TYPE]];
    
    //Set the audio name of saved recording file while is generated in the local storage path of SDK,by default in library/cache.
    [_iflyRecognizerView setParameter:@"asr.pcm" forKey:[IFlySpeechConstant ASR_AUDIO_PATH]];
    

    UIButton *btn = [UIButton buttonWithType:UIButtonTypeCustom];
    btn.backgroundColor = [UIColor redColor];
    btn.frame = CGRectMake(200, 300, 100, 60);
    [btn addTarget:self action:@selector(btnClick1:) forControlEvents:UIControlEventTouchUpInside];
    btn.selected = NO;
    [self.view addSubview:btn];
    
    [self setExclusiveTouchForButtons:self.view];

}

-(void)setExclusiveTouchForButtons:(UIView *)myView
{
    for (UIView * button in [myView subviews]) {
        if([button isKindOfClass:[UIButton class]])
        {
            [((UIButton *)button) setExclusiveTouch:YES];
        }
        else if ([button isKindOfClass:[UIView class]])
        {
            [self setExclusiveTouchForButtons:button];
        }
    }
}


- (void)btnClick1:(UIButton *)btn{
    [_resultARR removeAllObjects];
    //启动识别服务
    [_iflyRecognizerView start];

}

/*!
 *  回调返回识别结果
 *
 *  @param resultArray 识别结果，NSArray的第一个元素为NSDictionary，NSDictionary的key为识别结果，sc为识别结果的置信度
 *  @param isLast      -[out] 是否最后一个结果
 */
- (void)onResult:(NSArray *)resultArray isLast:(BOOL) isLast{
    NSMutableString *resultString = [[NSMutableString alloc] init];
    NSDictionary *dic = [resultArray objectAtIndex:0];
    NSLog(@"%@",dic);
    for (NSString *key in dic) {
        [resultString appendFormat:@"%@",key];
    }
    
    NSString * resultFromJson =  nil;
    if([IATConfig sharedInstance].isTranslate){
        
        NSDictionary *resultDic  = [NSJSONSerialization JSONObjectWithData:    //The result type must be utf8, otherwise an unknown error will happen.
                                    [resultString dataUsingEncoding:NSUTF8StringEncoding] options:kNilOptions error:nil];
        if(resultDic != nil){
            NSDictionary *trans_result = [resultDic objectForKey:@"trans_result"];
            
            NSString *src = [trans_result objectForKey:@"src"];
            NSLog(@"src=%@",src);
            resultFromJson = [NSString stringWithFormat:@"%@\nsrc:%@",resultString,src];
        }
    }
    else{
        resultFromJson = [NSString stringWithFormat:@"%@",resultString];//;[ISRDataHelper stringFromJson:resultString];
    }
    [_resultARR addObject:resultFromJson];
    
    
}

/*!
 *  识别结束回调
 *
 *  @param error 识别结束错误码
 */
- (void)onError: (IFlySpeechError *) error{

    NSLog(@"errorCode:%d",[error errorCode]);
    NSString *resultStr = [_resultARR componentsJoinedByString:@" "];
    UIAlertView *alertV = [[UIAlertView alloc]initWithTitle:@"" message:resultStr delegate:self cancelButtonTitle:@"cancle" otherButtonTitles:nil];
    [alertV show];

}


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


@end

效果图：

////////////

最后，为了提高语音识别精确读，在程序启动时加入以下代码

#pragma mark----------语音转文字时需要精确辨别的词----------
- (void)setWordsIUsed{
    
    //用户词表
#define USERWORDS   @"{\"userword\":[{\"name\":\"iflytek\",\"words\":[\"指动生活\",\"指动\",\"生活\",\"温馨提示\",\"温馨\",\"提示\",\"家园活动\",\"家园\",\"活动\",\"智能门禁\",\"智能\",\"门禁\",\"投诉报修\",\"投诉\",\"报修\",\"物业缴费\",\"交费\",\"缴费\"]}]}"
    
    //创建上传对象
    _uploader = [[IFlyDataUploader alloc] init];
    IFlyUserWords *iFlyUserWords = [[IFlyUserWords alloc] initWithJson:USERWORDS ];
    //设置上传参数
    [_uploader setParameter:@"uup" forKey:@"sub"];
    [_uploader setParameter:@"userword" forKey:@"dtt"];
    //启动上传（请注意name参数的不同）
    [_uploader uploadDataWithCompletionHandler:^(NSString * grammerID, IFlySpeechError *error){
        //
    }name: @"userwords" data:[iFlyUserWords toString]];
    
    
}

百度云盘：讯飞语音使用