使用iOS自带AAC编码器

AAC（Advanced Audio Coding），中文名：高级音频编码，出现于1997年，基于MPEG-2的音频编码技术。由Fraunhofer IIS、杜比实验室、AT&T、Sony等公司共同开发，目的是取代MP3格式。2000年，MPEG-4标准出现后，AAC重新集成了其特性，加入了SBR技术和PS技术，为了区别于传统的MPEG-2 AAC又称为MPEG-4 AAC。

iOS平台支持AAC编码器，主要使用AudioToolbox中的AudioConverter API。之所以做AAC编码器是因为在做一个HLS的功能，HLS要求的TS文件，需要视频采用H264编码，音频采用AAC编码。H264可以使用硬件或软件编码器，前面已经介绍。AAC也可以使用硬件或者软件编码，iOS全都支持。

首先需要创建一个Converter，也就是一个AAC Encoder，使用如下接口：

extern OSStatus
AudioConverterNew(      const AudioStreamBasicDescription*  inSourceFormat,
                        const AudioStreamBasicDescription*  inDestinationFormat,
                        AudioConverterRef*                  outAudioConverter)      __OSX_AVAILABLE_STARTING(__MAC_10_1,__IPHONE_2_0);

输入参数分别是源和目的的数据格式。

在AAC编码的场景下，源格式就是采集到的PCM数据，目的格式就是AAC。

    AudioStreamBasicDescription inAudioStreamBasicDescription;
    
//    FillOutASBDForLPCM()
    inAudioStreamBasicDescription.mFormatID = kAudioFormatLinearPCM;
    inAudioStreamBasicDescription.mSampleRate = 44100;
    inAudioStreamBasicDescription.mBitsPerChannel = 16;
    inAudioStreamBasicDescription.mFramesPerPacket = 1;
    inAudioStreamBasicDescription.mBytesPerFrame = 2;
    inAudioStreamBasicDescription.mBytesPerPacket = inAudioStreamBasicDescription.mBytesPerFrame * inAudioStreamBasicDescription.mFramesPerPacket;
    inAudioStreamBasicDescription.mChannelsPerFrame = 1;
    inAudioStreamBasicDescription.mFormatFlags = kLinearPCMFormatFlagIsPacked | kLinearPCMFormatFlagIsSignedInteger | kLinearPCMFormatFlagIsNonInterleaved;
    inAudioStreamBasicDescription.mReserved = 0;
    
    AudioStreamBasicDescription outAudioStreamBasicDescription = {0}; // Always initialize the fields of a new audio stream basic description structure to zero, as shown here: ...
    
    outAudioStreamBasicDescription.mChannelsPerFrame = 1;
    outAudioStreamBasicDescription.mFormatID = kAudioFormatMPEG4AAC;
    UInt32 size = sizeof(outAudioStreamBasicDescription);
    AudioFormatGetProperty(kAudioFormatProperty_FormatInfo, 0, NULL, &size, &outAudioStreamBasicDescription);

    OSStatus status = AudioConverterNew(&inAudioStreamBasicDescription, &outAudioStreamBasicDescription, &_audioConverter);
    if(status != 0) {NSLog(@"setup converter failed: %d", (int)status);}

这样就创建了AAC编码器，默认情况下，Apple会创建一个硬件编码器，如果硬件不可用，会创建软件编码器。

经过我的测试，硬件AAC编码器的编码时延很高，需要buffer大约2秒的数据才会开始编码。而软件编码器的编码时延就是正常的，只要喂给1024个样点，就会开始编码。

那么如何在创建的时候指定使用软件编码器呢？需要用到下面的接口：

- (AudioClassDescription *)getAudioClassDescriptionWithType:(UInt32)type
                                           fromManufacturer:(UInt32)manufacturer
{
    static AudioClassDescription desc;
    
    UInt32 encoderSpecifier = type;
    OSStatus st;
    
    UInt32 size;
    st = AudioFormatGetPropertyInfo(kAudioFormatProperty_Encoders,
                                    sizeof(encoderSpecifier),
                                    &encoderSpecifier,
                                    &size);
    if (st) {
        NSLog(@"error getting audio format propery info: %d", (int)(st));
        return nil;
    }
    
    unsigned int count = size / sizeof(AudioClassDescription);
    AudioClassDescription descriptions[count];
    st = AudioFormatGetProperty(kAudioFormatProperty_Encoders,
                                sizeof(encoderSpecifier),
                                &encoderSpecifier,
                                &size,
                                descriptions);
    if (st) {
        NSLog(@"error getting audio format propery: %d", (int)(st));
        return nil;
    }
    
    for (unsigned int i = 0; i < count; i++) {
        if ((type == descriptions[i].mSubType) &&
            (manufacturer == descriptions[i].mManufacturer)) {
            memcpy(&desc, &(descriptions[i]), sizeof(desc));
            return &desc;
        }
    }
    
    return nil;
}

AudioClassDescription *desc = [self getAudioClassDescriptionWithType:kAudioFormatMPEG4AAC
 fromManufacturer:kAppleSoftwareAudioCodecManufacturer];
 OSStatus status = AudioConverterNewSpecific(&inAudioStreamBasicDescription, &outAudioStreamBasicDescription, 1, desc, &_audioConverter);

如果要正确的编码，编码码率参数是必须设置的。否则编码时会返回560226676错误码（!dat）。

        UInt32 ulBitRate = 64000;
        UInt32 ulSize = sizeof(ulBitRate);
        status = AudioConverterSetProperty(_audioConverter, kAudioConverterEncodeBitRate, ulSize, &ulBitRate);

需要注意，AAC并不是随便的码率都可以支持。比如如果PCM采样率是44100KHz，那么码率可以设置64000bps，如果是16K，可以设置为32000bps。

创建完成Converter和设置完Bitrate之后，可以查询一下最大编码输出的大小，后续会用到。

    UInt32 value = 0;
    size = sizeof(value);
    AudioConverterGetProperty(_audioConverter, kAudioConverterPropertyMaximumOutputPacketSize, &size, &value);

获取出来的Value表示编码器最大输出的包大小。

然后调用AudioConverterFillCOmplexBuffer进行编码：

        AudioBufferList outAudioBufferList = {0};
        outAudioBufferList.mNumberBuffers = 1;
        outAudioBufferList.mBuffers[0].mNumberChannels = 1;
        outAudioBufferList.mBuffers[0].mDataByteSize = value;//value是上面查询到的值
        outAudioBufferList.mBuffers[0].mData = new int8[value];
                
        UInt32 ioOutputDataPacketSize = 1;
        status = AudioConverterFillComplexBuffer(_audioConverter, inInputDataProc, (__bridge void *)(self), &ioOutputDataPacketSize, &outAudioBufferList, NULL);

编码接口中，inInputDataProc是一个输入数据的回调函数。用来喂PCM数据给Converter，ioOutputDataPacketSize为1表示编码产生1帧数据即返回。outAudioBufferList用来存放编码后的数据。

inInputDataProc中的处理如下：

static OSStatus inInputDataProc(AudioConverterRef inAudioConverter, UInt32 *ioNumberDataPackets, AudioBufferList *ioData, AudioStreamPacketDescription **outDataPacketDescription, void *inUserData)
{
    
    AACEncoder *encoder = (__bridge AACEncoder *)(inUserData);
    UInt32 requestedPackets = *ioNumberDataPackets;
    uint8_t *buffer;
    uint32_t bufferLength = requestedPackets * 2;
    uint32_t bufferRead;
    bufferRead = [encoder.pcmPool readBuffer:&buffer withLength:bufferLength];
    if (bufferRead == 0) {
        *ioNumberDataPackets = 0;
        return -1;
    }

    ioData->mBuffers[0].mData = buffer;
    ioData->mBuffers[0].mDataByteSize = bufferRead;
    ioData->mNumberBuffers = 1;
    ioData->mBuffers[0].mNumberChannels = 1;
    
    *ioNumberDataPackets = bufferRead >> 1;
    return noErr;
}

pcmPool是一个用于存放PCM数据的环形缓冲区。

因为采集输入每次不一定有1024样点，所以可以将数据缓存起来，再满足1024样点时再调用编码。

另外，对于TS文件来说，每个AAC数据需要增加一个adts头，adts头是一个7bit的数据，通过adts可以得知AAC数据的编码参数，方便解码器进行解码。

adts头的计算方法如下：

- (NSData*) adtsDataForPacketLength:(NSUInteger)packetLength {
    int adtsLength = 7;
    char *packet = (char *)malloc(sizeof(char) * adtsLength);
    // Variables Recycled by addADTStoPacket
    int profile = 2;  //AAC LC
    //39=MediaCodecInfo.CodecProfileLevel.AACObjectELD;
    int freqIdx = 8;  //16KHz
    int chanCfg = 1;  //MPEG-4 Audio Channel Configuration. 1 Channel front-center
    NSUInteger fullLength = adtsLength + packetLength;
    // fill in ADTS data
    packet[0] = (char)0xFF;	// 11111111  	= syncword
    packet[1] = (char)0xF9;	// 1111 1 00 1  = syncword MPEG-2 Layer CRC
    packet[2] = (char)(((profile-1)<<6) + (freqIdx<<2) +(chanCfg>>2));
    packet[3] = (char)(((chanCfg&3)<<6) + (fullLength>>11));
    packet[4] = (char)((fullLength&0x7FF) >> 3);
    packet[5] = (char)(((fullLength&7)<<5) + 0x1F);
    packet[6] = (char)0xFC;
    NSData *data = [NSData dataWithBytesNoCopy:packet length:adtsLength freeWhenDone:YES];
    return data;
}

adts头的计算需要几个参数：profile／frequency／channels／length，具体可参考 http://wiki.multimedia.cx/index.php?title=ADTS

参考文献：

Audio Converter Services Reference