大家好,我是RickyWei。

语音识别有很多种,今天给大家带来的是,百度语音识别 。

一. 引入编译需要的Framework

1.BDVRClient使用了录音和播放功能,因此需要在Xcode工程中引入AudioToolbox.framework和AVFoundation.framework;BDVRClient还使用到了网络状态检测功能,因此还需要引入SystemConfiguration.framework;为了生成设备UDID,需要引入Security.framework;为了支持gzip压缩,需要引入libz.1.dylib; 网络模块需要引入CFNetwork.framework;某些场景需要获取设备地理位置以提高识别准确度,需引入CoreLocation.framework。

为了支持识别控件,需要引入OpenGLES.framework,QuartzCore.framework,GLKit.framework,CoreGraphics.framework和CoreText.framework。

添加方式:右键点击Xcode中的工程文件,在出现的界面中,选中TARGETS中应用,在出现的界面中选中Build Phase->Link Binary With Libraries,点击界面中的“+”图标,在弹出的界面中选择这些Framework即可,添加完成效果图如图所示。

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

2.将在百度语音开房平台上下载的SDK拖入到项目工程的文件目录下

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

3.删除BDVRClientSample文件夹

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

4.如图 向项目中添加文件夹

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

4.1 将文件夹BDVoiceRecognitionClientResources从项目中移除 再添加

添加模式如下:

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

4.2 修改后文件夹目录如下:

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

二.引入BDVRClient的头文件

1.首先将BDVRClient提供的头文件拷贝到工程目录下,在XCode中添加此文件,引入BDVRClient提供的头文件。

2.1 如果使用识别UI,请添加如下头文件(本文只使用识别UI和接口):1.#import "BDRecognizerViewController.h"

2.#import "BDRecognizerViewDelegate.h"

2.2 如果只使用识别接口,添加如下头文件:1.#import "BDVoiceRecognitionClient.h"

2.3 如果要对音频数据或音频文件直接进行识别,请分别添加如下头文件:1.#import "BDVRRawDataRecognizer.h"

2.#import "BDVRFileRecognizer.h"

添加完后如图:

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

2.4 控件代码如下:

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

apiKey与secretKey 为百度语音识别应用给的key

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation#import "ViewController.h"

#import "BDRecognizerViewController.h"

#import "BDRecognizerViewDelegate.h"

#import "BDVoiceRecognitionClient.h"@interface ViewController (){

//语音界面

BDRecognizerViewController *bdrv;

NSMutableData *allData;

//参数设置 Key 密钥

BDRecognizerViewParamsObject *bdvp;

}//这是直接拉的一个text View

@property (weak, nonatomic) IBOutlet UITextView *textLable;@end

@implementation ViewController- (void)viewDidLoad {

[super viewDidLoad];

//添加一个按钮

UIButton *b = [UIButton buttonWithType:UIButtonTypeRoundedRect];

b.frame = CGRectMake(100, 400, 100, 30);

[b setTitle:@"开始识别" forState:UIControlStateNormal];

[b addTarget:self action:@selector(click) forControlEvents:UIControlEventTouchUpInside];

[self.view addSubview:b];

//主题设置

BDTheme *me = [BDTheme lightOrangeTheme];

bdrv = [[BDRecognizerViewController alloc]initWithOrigin:CGPointMake(20, 10) withTheme:me];

//设置为全屏幕

bdrv.enableFullScreenMode = YES;

bdrv.delegate = self;

bdvp = [[BDRecognizerViewParamsObject alloc]init];

//bdvp.productID 不用设置

bdvp.apiKey = @"0VoiTxX42brGtyp3P330cONO";

bdvp.secretKey = @"5b929494d330b6d57385c7539628cccd";

}-(void)click{

allData = [[NSMutableData alloc]init];

[bdrv startWithParams:bdvp];

}/**

* @brief 录音数据返回

*

* @param recordData 录音数据

* @param sampleRate 采样率

*/

- (void)onRecordDataArrived:(NSData *)recordData sampleRate:(int)sampleRate{

[allData appendData:recordData];

}/**

* @brief 返回中间识别结果

*

* @param results

*            中间识别结果

*/

- (void)onPartialResults:(NSString *)results{

self.textLable.text = results;

}//这个可以选择性删除

- (void)didReceiveMemoryWarning {

[super didReceiveMemoryWarning];

// Dispose of any resources that can be recreated.

}@end

三.添加第三方开源库

3.1     BDVRClient中使用了第三方开源库,包括TTTAttributedLabel和苹果非官方的录音API,  如果产品项目中已经包含其中的部分库,请勿重复添加,否则,请添加这三种第三方开源库到项目中,第三方库文件可以在SDK开发包下的Third-party目录下找到。由于SDK中使用了类别扩展,请在Build Setting中的Other Linker Flags中添加-ObjC。

注意:其中第三方库TTTAttributedLabel需要设置为ARC方式编译。

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

3.2 同时不要忘记要把JSONKit关闭 ARC

ebe92a1aaa54?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation

如上,为本人对百度语音识别SDK小小的理解,希望可以帮到您。

Logo

CSDN联合极客时间,共同打造面向开发者的精品内容学习社区,助力成长!

更多推荐