AudioQueue实现音频流实时播放实战

时间  2019-11-06
标签 audioqueue 实现 音频 实时 播放 实战 繁體版
原文   原文链接

需求

使用Audio Queue实现实时播放音频流数据.这里以一个装着pcm数据的caf文件为例进行播放.node

实现原理

借助数据传输队列,将不管任务数据源的音频数据装入队列中,而后开启audio queue后从队列中循环取出音频数据以进行播放.ios

阅读前提

代码地址 : Audio Queue Player

掘金地址 : Audio Queue Player

简书地址 : Audio Queue Player

博客地址 : Audio Queue Player

整体架构

本例借助队列实现音频数据的中转, 这里用队列是由于audio queue是靠数据驱动以支持播放的,因此有数据回调函数才能持续调用,若是咱们不借助队列,就只能在audio queue的类中从回调函数中取来自音频文件的数据,并且假设之后有别的数据源过来,使得音频播放模块代码耦合度愈来愈高,而这里借助队列的好处是外界不管是音频文件仍是音频流仅仅须要放入队列中就好,开启音频模块后咱们会从音频队列回调函数中取出队列中的数据,而无需关心数据的来源.git

简易流程

  • AudioStreamBasicDescription: 配置传入的音频数据格式
  • AudioQueueNewOutput : 新建audio queue
  • AudioQueueAddPropertyListener : 监听audio queue是否正在工做
  • AudioQueueSetParameter : 设置音量
  • AudioQueueAllocateBuffer : 为audio queue buffer 分配内存
  • 从队列中取出数据装入audio queue buffer,并入队AudioQueueEnqueueBuffer
  • AudioQueueStart : 开启audio queue
  • 进入播放回调函数, 在回调函数中取出队列中存储的音频数据
  • 再次入队, 以播放音频数据,播放完后会自动触发回调函数,依次循环

文件结构

1.file_structure

快速使用

  • 配置音频数据来源的ASBD

下面是本例中的格式,其余文件须要按文件格式自行配置github

// This is only for the testPCM.caf file.
    AudioStreamBasicDescription audioFormat = {
        .mSampleRate         = 44100,
        .mFormatID           = kAudioFormatLinearPCM,
        .mChannelsPerFrame   = 1,
        .mFormatFlags        = kLinearPCMFormatFlagIsSignedInteger | kLinearPCMFormatFlagIsPacked,
        .mBitsPerChannel     = 16,
        .mBytesPerPacket     = 2,
        .mBytesPerFrame      = 2,
        .mFramesPerPacket    = 1,
    };
复制代码
  • 配置audio queue player
// Configure Audio Queue Player
    [[XDXAudioQueuePlayer getInstance] configureAudioPlayerWithAudioFormat:&audioFormat bufferSize:kXDXReadAudioPacketsNum * audioFormat.mBytesPerPacket];
复制代码
  • 配置音频文件模块
// Configure Audio File
    NSString *filePath = [[NSBundle mainBundle] pathForResource:@"testPCM" ofType:@"caf"];
    XDXAudioFileHandler *fileHandler = [XDXAudioFileHandler getInstance];
    [fileHandler configurePlayFilePath:filePath];
复制代码
  • 开始播放

开始播放前先从文件中读取音频数据并放入队列,咱们这里先让队列中缓存5帧音频数据,而后再启动audio queue player. 关于音频文件读取以及队列原理这里不作过多说明.如需帮助请参考上文阅读前提.缓存

// Put audio data from audio file into audio data queue
    [self putAudioDataIntoDataQueue];
    
    // First put 5 frame audio data to work queue then start audio queue to read it to play.
    [NSTimer scheduledTimerWithTimeInterval:0.01 repeats:YES block:^(NSTimer * _Nonnull timer) {
        dispatch_async(dispatch_get_main_queue(), ^{
            XDXCustomQueueProcess *audioBufferQueue = [XDXAudioQueuePlayer getInstance]->_audioBufferQueue;
            int size = audioBufferQueue->GetQueueSize(audioBufferQueue->m_work_queue);
            if (size > 5) {
                [[XDXAudioQueuePlayer getInstance] startAudioPlayer];
                [timer invalidate];
            }
        });
    }];
复制代码

具体实现

1. 定义一个结构体存储音频相关数据

#define kXDXAudioPCMFramesPerPacket 1
#define kXDXAudioPCMBitsPerChannel 16

static const int kNumberBuffers = 3;

struct XDXAudioInfo {
    AudioStreamBasicDescription  mDataFormat;
    AudioQueueRef                mQueue;
    AudioQueueBufferRef          mBuffers[kNumberBuffers];
    int                          mbufferSize;
};
typedef struct XDXAudioInfo *XDXAudioInfoRef;

static XDXAudioInfoRef m_audioInfo;

+ (void)initialize {
    int size = sizeof(XDXAudioInfo);
    m_audioInfo = (XDXAudioInfoRef)malloc(size);
}

复制代码

2. 初始化

在初始化方法中初始化音频队列,由于本例借助另外一个类进行传输,因此这里做为实例对象,以便使用.bash

- (instancetype)init {
    static dispatch_once_t onceToken;
    dispatch_once(&onceToken, ^{
        _instace                  = [super init];
        self->_isInitFinish       = NO;
        self->_audioBufferQueue   = new XDXCustomQueueProcess();
    });
    return _instace;
}
复制代码

3. 配置音频播放器

  • 将传入的音频格式,音频Buffer大小拷贝到实例中
- (void)configureAudioPlayerWithAudioFormat:(AudioStreamBasicDescription *)audioFormat bufferSize:(int)bufferSize {
    memcpy(&m_audioInfo->mDataFormat, audioFormat, sizeof(XDXAudioInfo));
    m_audioInfo->mbufferSize = bufferSize;    
    BOOL isSuccess = [self configureAudioPlayerWithAudioInfo:m_audioInfo
                                                playCallback:PlayAudioDataCallback
                                            listenerCallback:AudioQueuePlayerPropertyListenerProc];
    
    self.isInitFinish = isSuccess;
}

复制代码
  • 建立audio queue对象实例

经过传入的视频数据格式ASBD, 及回调函数名称便可建立一个对应的audio queue对象.这里将本类做为实例传入,以便回调函数与本类交流.架构

注意: 由于回调函数是C语言函数的形式,因此没法直接调用类的实例方法.async

// Create audio queue
    OSStatus status = AudioQueueNewOutput(&audioInfo->mDataFormat,
                                         playCallback,
                                         (__bridge void *)(self),
                                         CFRunLoopGetCurrent(),
                                         kCFRunLoopCommonModes,
                                         0,
                                         &audioInfo->mQueue);
    
    if (status != noErr) {
        NSLog(@"Audio Player: audio queue new output failed status:%d \n",(int)status);
        return NO;
    }
复制代码
  • 监听audio queue工做状态 在回调函数中能够监听audio queue实例工做工做的变化,如正在播放或中止播放.
// Listen the queue is whether working
    AudioQueueAddPropertyListener (audioInfo->mQueue,
                                   kAudioQueueProperty_IsRunning,
                                   listenerCallback,
                                   (__bridge void *)(self));
                                   
......

static void AudioQueuePlayerPropertyListenerProc  (void *              inUserData,
                                                   AudioQueueRef           inAQ,
                                                   AudioQueuePropertyID    inID) {
    XDXAudioQueuePlayer * instance = (__bridge XDXAudioQueuePlayer *)inUserData;
    UInt32 isRunning = 0;
    UInt32 size = sizeof(isRunning);
    
    if(instance == NULL)
        return ;
    
    OSStatus err = AudioQueueGetProperty (inAQ, kAudioQueueProperty_IsRunning, &isRunning, &size);
    if (err) {
        instance->_isRunning = NO;
    }else {
        instance->_isRunning = isRunning;
    }
    
    NSLog(@"The audio queue work state: %d",instance->_isRunning);
}
复制代码
  • 验证设置的ASBD音频格式及设置音量
// Get audio ASBD
    UInt32 size = sizeof(audioInfo->mDataFormat);
    status = AudioQueueGetProperty(audioInfo->mQueue,
                                   kAudioQueueProperty_StreamDescription,
                                   &audioInfo->mDataFormat,
                                   &size);
    if (status != noErr) {
        NSLog(@"Audio Player: get ASBD status:%d",(int)status);
        return NO;
    }
    
    // Set volume
    status = AudioQueueSetParameter(audioInfo->mQueue, kAudioQueueParam_Volume, 1.0);
    if (status != noErr) {
        NSLog(@"Audio Player: set volume failed:%d",(int)status);
        return NO;
    }
复制代码
  • 为audio queue buffer分配内存
// Allocate buffer for audio queue buffer
    for (int i = 0; i != kNumberBuffers; i++) {
        status = AudioQueueAllocateBuffer(audioInfo->mQueue,
                                          audioInfo->mbufferSize,
                                          &audioInfo->mBuffers[i]);
        if (status != noErr) {
            NSLog(@"Audio Player: Allocate buffer status:%d",(int)status);
        }
    }
    
复制代码

4. 启动audio queue

  • 预入队几个buffer以驱动播放

由于audio queue是驱动播放的模式,因此只有数据先入队以后才会继续从回调函数中轮循播放,也就是咱们须要将前面分配好内存的buffer入队来完成播放.函数

播放采用从原始音频数据队列中读取音频数据,以下,先出队,而后将音频数据拷贝到AudioQueueBufferRef实例,取出须要的信息(此队列仍可继续扩展).oop

for (int i = 0; i != kNumberBuffers; i++) {
        [self receiveAudioDataWithAudioQueueBuffer:audioInfo->mBuffers[i]
                                         audioInfo:audioInfo
                                  audioBufferQueue:_audioBufferQueue];
    }
    
    ......
    
- (void)receiveAudioDataWithAudioQueueBuffer:(AudioQueueBufferRef)inBuffer audioInfo:(XDXAudioInfoRef)audioInfo audioBufferQueue:(XDXCustomQueueProcess *)audioBufferQueue {
    XDXCustomQueueNode *node = audioBufferQueue->DeQueue(audioBufferQueue->m_work_queue);
    
    if (node != NULL) {
        if (node->size > 0) {
            UInt32 size = (UInt32)node->size;
            inBuffer->mAudioDataByteSize = size;
            memcpy(inBuffer->mAudioData, node->data, size);
            AudioStreamPacketDescription *packetDesc = (AudioStreamPacketDescription *)node->userData;
            AudioQueueEnqueueBuffer (
                                     audioInfo->mQueue,
                                     inBuffer,
                                     (packetDesc ? size : 0),
                                     packetDesc);

        }

        free(node->data);
        node->data = NULL;
        audioBufferQueue->EnQueue(audioBufferQueue->m_free_queue, node);
    }else {
        AudioQueueStop (
                        audioInfo->mQueue,
                        false
                        );
    }
}
复制代码
  • 开始工做
OSStatus status;
    status = AudioQueueStart(m_audioInfo->mQueue, NULL);
    if (status != noErr) {
        NSLog(@"Audio Player: Audio Queue Start failed status:%d \n",(int)status);
        return NO;
    }else {
        NSLog(@"Audio Player: Audio Queue Start successful");
        return YES;
    }
复制代码

5. 触发回调函数轮循播放

正如前面所说, audio queue的播放模式是数据驱动式,也就是咱们已经预先入队了几个音频队列数据,而后开启audio queue后咱们它会自动播放前面已经入队的数据,每当播放完会自动触发回调函数读取数据以完成下一次播放.

static void PlayAudioDataCallback(void * aqData,AudioQueueRef inAQ , AudioQueueBufferRef inBuffer) {
    XDXAudioQueuePlayer *instance = (__bridge XDXAudioQueuePlayer *)aqData;
    if(instance == NULL){
        return;
    }
    
    /* Debug
    static Float64 lastTime = 0;
    NSTimeInterval currentTime = [[NSDate date] timeIntervalSince1970]*1000;
    NSLog(@"Test duration - %f",currentTime - lastTime);
    lastTime = currentTime;
    */
    
    [instance receiveAudioDataWithAudioQueueBuffer:inBuffer
                                         audioInfo:m_audioInfo
                                  audioBufferQueue:instance->_audioBufferQueue];
}
复制代码

6. 其余

Demo中还有音频队列的暂停, 恢复, 中止, 销毁等功能,较为简单,这里再也不说明.

7. 从音频文件中读取音频数据

  • 经过本地文件路径实例化为CFURLRef对象
- (void)configurePlayFilePath:(NSString *)filePath {
    char path[256];
    [filePath getCString:path maxLength:sizeof(path) encoding:NSUTF8StringEncoding];
    self->m_playFileURL = CFURLCreateFromFileSystemRepresentation (
                                                                   NULL,
                                                                   (const UInt8 *)path,
                                                                   strlen (path),
                                                                   false
                                                                   );
}

复制代码
  • 使用时先打开文件

函数中可配置文件权限及类型,本例中文件类型为caf文件.

OSStatus status;
        status = AudioFileOpenURL(self->m_playFileURL,
                                  kAudioFileReadPermission,
                                  kAudioFileCAFType,
                                  &self->m_playFile);
        if (status != noErr) {
            NSLog(@"open file failed: %d", (int)status);
        }
复制代码
  • 从文件中读取音频数据

首先指定每次读取多少个音频数据包, 该函数会返回最终读取的字节数. 这里经过m_playCurrentPacket记录当前读取的音频包数以便下次继续读取.读取完成后关闭文件.

UInt32 bytesRead = 0;
    UInt32 numPackets = readPacketsNum;
    OSStatus status = AudioFileReadPackets(m_playFile,
                                  false,
                                  &bytesRead,
                                  packetDesc,
                                  m_playCurrentPacket,
                                  &numPackets,
                                  audioDataRef);
    
    if (status != noErr) {
        NSLog(@"read packet failed: %d", (int)status);
    }
    
    if (bytesRead > 0) {
        m_playCurrentPacket += numPackets;
    }else {
        status = AudioFileClose(m_playFile);
        if (status != noErr) {
            NSLog(@"close file failed: %d", (int)status);
        }
        self.isPlayFileWorking = NO;
        m_playCurrentPacket = 0;
    }
复制代码
相关文章
相关标签/搜索
每日一句
    每一个你不满意的现在，都有一个你没有努力的曾经。
本站公众号
   欢迎关注本站公众号,获取更多信息
联系我们 最近搜索 最新文章 沪ICP备13005482号-10 MyBatis教程 SQL 教程 MySQL教程 Java 教程 Thymeleaf 教程 Hibernate教程 Spring教程 Redis教程