语音合成 iOS SDK

Architectures in the fat file: libBDSpeechTTSBaseKit.a are: x86_64 arm64

// 创建语音合成器实例
    self.synthesizer = [[BDSSpeechSynthesizer alloc] initWithDelegate:self];
    // 设置离线参数资源文件
    NSString *textData = [[NSBundle mainBundle] pathForResource:@"bd_etts_common_text_txt_all_mand_eng_middle_big_v6.0.0_20240731" ofType:@"dat"];
    NSString *speechData = [[NSBundle mainBundle] pathForResource:@"bd_etts_common_speech_duxiaoyao_mand_eng_high_am-tac-csubgan16k_v4.9.0_20251028_20251031153737" ofType:@"dat"];

    // 正确的离线模型配置
    NSDictionary *offlineModel = @{
            BDSSpeechSynthesizerParams.ttsTextModelFile: textData,
            BDSSpeechSynthesizerParams.ttsSpeechModelFile: speechData
        };
    NSData *offlineModeData = [NSJSONSerialization dataWithJSONObject:offlineModel options:0 error:nil];
    NSString *offlineModelString = [[NSString alloc] initWithData:offlineModeData encoding:NSUTF8StringEncoding];
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramOfflineModel value:offlineModelString];

#import "TTSViewController.h"
#import "AuthManager.h"
#import <AVFAudio/AVAudioSession.h>

#import "BDSpeechTTSBaseKit.h"
#import "BDSSpeechSynthesizerResponse.h"
#import "BDSSpeechLogger.h"
#import "BDSSpeechSynthesizerParams.h"
#import "Reachability.h"

- (void)initOnlineSDK {
    // 创建语音合成器实例
    self.synthesizer = [[BDSSpeechSynthesizer alloc] initWithDelegate:self];
    // 申请获取token鉴权
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAccessToken value:@"平台申请获取的token"];
    // 申请获取IAMKEY鉴权
    //  [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAccessToken value:@"平台申请获取的IAMKEY"];
    // 设置在线发音人
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramOnlineSpeaker value:@"4100"];
//     加载在线引擎
    BDSSpeechResult *result = [self.synthesizer loadOnlineEngine];
    if (result.code != 0) {
        NSLog(@"加载在线引擎失败: %ld", (long)result.code);
    }
}

/**
 token鉴权方式
 token：申请的token
 expirationDate：token过期时间戳（毫秒）
 */
- (void)setTTSLicenseWithAuthToken:(NSString *)token AndExpirationDate:(long long)expirationDate;

/**
 iamkey方式初始化鉴权
 iamkey：申请的iamkey
 expirationDate：iamkey过期时间（毫秒）
 */
- (void)setTTSLicenseWithIAMKey:(NSString *)iamkey AndExpirationDate:(long long)expirationDate;

// 初始化离线合成SDK
[self initOfflineSDK];

- (void)initOfflineSDK {
    // 初始化日志，仅 DEBUG 调试打开此设置
    [[BDSSpeechLogger sharedInstance] setLogLevel:BDSSpeechLogLevelInfo path:nil callback:nil];
    
    // 创建语音合成器实例
    self.synthesizer = [[BDSSpeechSynthesizer alloc] initWithDelegate:self];
    // 设置离线参数资源文件
    NSString *textData = [[NSBundle mainBundle] pathForResource:@"bd_etts_common_text_txt_all_mand_eng_middle_big_v6.0.0_20240731" ofType:@"dat"];
    NSString *speechData = [[NSBundle mainBundle] pathForResource:@"bd_etts_common_speech_duxiaoyao_mand_eng_high_am-tac-csubgan16k_v4.9.0_20251028_20251031153737" ofType:@"dat"];

    // 正确的离线模型配置
    NSDictionary *offlineModel = @{
            BDSSpeechSynthesizerParams.ttsTextModelFile: textData,
            BDSSpeechSynthesizerParams.ttsSpeechModelFile: speechData
        };
    NSData *offlineModeData = [NSJSONSerialization dataWithJSONObject:offlineModel options:0 error:nil];
    NSString *offlineModelString = [[NSString alloc] initWithData:offlineModeData encoding:NSUTF8StringEncoding];
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramOfflineModel value:offlineModelString];
    
    // 设置离线授权文件
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAppId value:@"平台申请获取的APPID"];
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAuthSerialNumber value:@"平台申请获取的SN"];
    
    // 加载离线引擎
    BDSSpeechResult *result = [self.synthesizer loadOfflineEngine];
    result = [self.synthesizer loadOfflineEngine];
    if (result.code != 0) {
        NSLog(@"加载离线引擎失败: %ld", (long)result.code);
    }
}

- (void)initCommonSDK {
    //主要通过函数实现 参考demo对应部分
}

步骤一：
// 遵循鉴权sdk内置协议，初始化设置协议
[AuthManager sharedInstance].temporaryTokenAndKeyDelegate = self;

// 分别实现协议传入token、iamkey
/**
 客户实现get token 协议方法
 */
- (void)getToken {
    ......
    //客户业务侧处理获取token的逻辑
    // demo示例发送请求
     [self sendPostRequest];
}

/**
 客户实现get iamkey 协议方法
 */
- (void)getIAMKey {
    ......
    //客户业务侧处理获取iamkey的逻辑
    // demo示例发送请求
     [self sendPostRequest];
}

步骤二：
// 初始化调用在线函数进行SDK模型参数初始化配置
// 在线TTS
 [self initCommonSDK]; // 完成模型初始化参数配置相关操作
//获取网络权限判断逻辑
    ......
    ......
    ......
// 检查当前网络状态并尝试网络请求         
[self checkNetworkStatusAndMakeRequest];       
          
          
  // 检查网络权限是否生效
- (void)checkNetworkStatusAndMakeRequest {
    // 生效 即可做post请求获取token、有效期
    ......
    ......
    ......
    // 调用本地获取token的请求逻辑
     [self sendPostRequest];
}

  // post请求 仅供参考
-(void)sendPostRequest {
 // 请求逻辑具体参考demo代码
  ......
  ......
  请求拿到token后调用API接口完成token鉴权
  // token 鉴权   成功：正常完成合成主流程鉴权通过。 失败：触发上述步骤一中的协议回调 重新获取有效token（此处以token为例） 
  [[BDSpeechTTSBaseKit sharedInstance] setTTSLicenseWithAuthToken:@"平台获取的token" AndExpirationDate:"token有效期时间"];
}

逻辑整理 1.遵循鉴权sdk内置协议 2.完成初始化请求token获取[self sendPostRequest]; 并在请求结果中设置对应的鉴权方式如 [[BDSpeechTTSBaseKit sharedInstance] setTTSLicenseWithAuthToken:@"平台获取的token" AndExpirationDate:"token有效期时间"];获取token。3.成功后正常继续下一步，异常或请求token失败、token有效期到期等 SDK代理会监听到 触发getToken、getIamkey代理 进行重新处理逻辑 demo示例中做了重新获取token的操作 仅供参考 
   
注：使用该方式为了满足 「鉴权信息保存在server端，app通过请求server端获取access token」的需求，语音SDK需要提供 鉴权信息回调点，并由用户最终实现 回调接口。

- (void)initOfflineSDK {
    // 初始化日志，仅 DEBUG 调试打开此设置
    [[BDSSpeechLogger sharedInstance] setLogLevel:BDSSpeechLogLevelInfo path:nil callback:nil];
    
    // 创建语音合成器实例
    self.synthesizer = [[BDSSpeechSynthesizer alloc] initWithDelegate:self];
    // 设置离线参数资源文件
    NSString *textData = [[NSBundle mainBundle] pathForResource:@"bd_etts_common_text_txt_all_mand_eng_middle_big_v6.0.0_20240731" ofType:@"dat"];
    NSString *speechData = [[NSBundle mainBundle] pathForResource:@"bd_etts_common_speech_duxiaoyao_mand_eng_high_am-tac-csubgan16k_v4.9.0_20251028_20251031153737" ofType:@"dat"];

    // 正确的离线模型配置
    NSDictionary *offlineModel = @{
            BDSSpeechSynthesizerParams.ttsTextModelFile: textData,
            BDSSpeechSynthesizerParams.ttsSpeechModelFile: speechData
        };
    NSData *offlineModeData = [NSJSONSerialization dataWithJSONObject:offlineModel options:0 error:nil];
    NSString *offlineModelString = [[NSString alloc] initWithData:offlineModeData encoding:NSUTF8StringEncoding];
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramOfflineModel value:offlineModelString];
    
    // 设置离线授权文件
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAppId value:@"平台申请获取的APPID"];
    [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAuthSerialNumber value:@"平台申请获取的SN"];
    
    // 加载离线引擎
    BDSSpeechResult *result = [self.synthesizer loadOfflineEngine];
    result = [self.synthesizer loadOfflineEngine];
    if (result.code != 0) {
        NSLog(@"加载离线引擎失败: %ld", (long)result.code);
    }
}

// 初始化 - (instancetype)initWithDelegate:(id<BDSSpeechSynthesizerDelegate>)delegate;
// 释放 - (void)releaseEngine;

/**
* @brief 合成器参数设置接⼝ 
* @param param 参数对应的值 
* @param key 合成器预定义的参数名称，参⻅附录及头⽂件BDSSpeechSynthesizerParams.h. 
*/ - (BDSSpeechResult *)setParam:(NSString *)key value:(NSString *)value;

/**
* @brief 为在线合成设置认证信息
* @param AccessToken 在百度开发者中⼼注册应⽤获得
*/
[self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAccessToken value:@"xxx"];

// 加载在线引擎
BDSSpeechResult *result = [self.synthesizer loadOnlineEngine];
if (result != 0) {
NSLog(@"加载在线引擎失败: %ld", (long)result.code); } 

// 设置离线资源 (⽅式⼀，适合业务⾃⾏管理，云端下发资源⽂件)
NSString *textData = [[NSBundle mainBundle] pathForResource:@"xxx" ofType:@"dat"];
NSString *speechData = [[NSBundle mainBundle] pathForResource:@"xxx" ofType:@"dat"];
NSDictionary *offlineModel = @{BDSSpeechSynthesizerParams.ttsTextModelFile: @([textData UTF8String]),
BDSSpeechSynthesizerParams.ttsSpeechModelFile: @([speechData UTF8String])};
NSData *offlineModelData = [NSJSONSerialization dataWithJSONObject:offlineModel options:0 error:nil];
NSString *offlineModelString = [[NSString alloc] initWithData:offlineModelData encoding:NSUTF8StringEncoding]; [self.synthesizer setParam:BDSSpeechSynthesizerParams.paramOfflineModel value:offlineModelString];

// 设置您的鉴权 sn
[self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAuthSerialNumber value:@"your auth sn")]; 
[self.synthesizer setParam:BDSSpeechSynthesizerParams.paramAppId value:@"xxx"];

// 加载离线引擎
result = [self.synthesizer loadOfflineEngine];
if (result.code != 0) {
NSLog(@"加载离线引擎失败: %d", result.code); }

/**
* 合成并播放语⾳ * @param entity 语⾳合成实体 
* @return 播放结果 
*/ 
- (BDSSpeechResult *)speak:(BDSSpeechEntity *)entity;

/**
* 合成语⾳ * @param entity 语⾳合成实体 
* @return 合成结果 
*/ 
- (BDSSpeechResult *)synthesize:(BDSSpeechEntity *)entity;

/**
* 停⽌播放/合成 
* @return 停⽌结果 
*/
- (BDSSpeechResult *)stop;

/**
* 暂停播放
* @return 暂停结果 
*/ 
- (BDSSpeechResult *)pause;

/**
* 恢复播放
* @return 恢复结果 
*/ 
- (BDSSpeechResult *)resume;

- (void)onSynthesizerResponse:(BDSSpeechSynthesizerResponse *)synthesizerResponse {
switch (response.synthesizeType) {
case BDSSpeechSynthesizerTypeStart:
NSString *sn = synthesizerResponse.sn; // 此次SDK侧传递的⽂本合成的标识，建议保存，后续排查服务器等问题需要使⽤。
NSString *utteranceId = synthesizerResponse.utteranceId;// 此次业务侧传递的⽂本合成的标识，⽤来做相关逻辑判断。
break; ...... }

BDSSpeechEntity *entity = [[BDSSpeechEntity alloc] initWithText:text
                                                               mode:BDSSpeechModeOnline
                                                        utteranceId:[[NSUUID UUID] UUIDString]
                                                           params:nil];

[self.synthesizer setParam:BDSSpeechSynthesizerParams.paramPlayerVolume value:@"0-1"];

// 业务⽅根据需求场景⾃定义设置，此处仅做举例
- (void)setAudioSessionConfig {
NSString *category = AVAudioSessionCategoryPlayback;
NSUInteger categoryOptions = (AVAudioSessionCategoryOptionMixWithOthers|
AVAudioSessionCategoryOptionDuckOthers);
NSError *error = nil; [[AVAudioSession sharedInstance] setCategory:category withOptions:categoryOptions error:&error];
if (error) { [self appendLog:[NSString stringWithFormat:@"设置⾳频会话失败: %@", error.localizedDescription]]; } }

// 初始化⽇志，仅 DEBUG 调试打开此设置 
[[BDSSpeechLogger sharedInstance] setLogLevel:BDSSpeechLogLevelInfo path:nil];

bash
strings -a BDTTSClientSDK | grep TTS_SDK_VERSION // SDK 版本号
strings -a BDTTSClientSDK | grep ETTS_UNIQUE // 引擎 版本号