SRS 代码分析【保存AAC音频文件】

SRS 对AAC音频文件的保存主要是调用SrsAacTransmuxer::write_audio(),函数定义如下：

int SrsAacTransmuxer::write_audio(int64_t timestamp, char* data, int size)
{
    int ret = ERROR_SUCCESS;
    
    srs_assert(data);
    
    timestamp &= 0x7fffffff;
    
    SrsBuffer* stream = tag_stream;
    if ((ret = stream->initialize(data, size)) != ERROR_SUCCESS) {
        return ret;
    }
    
    // audio decode
    if (!stream->require(1)) {
        ret = ERROR_AAC_DECODE_ERROR;
        srs_error("aac decode audio sound_format failed. ret=%d", ret);
        return ret;
    }
    
    //从音频Tag中读取AAC Sequence header
    //SoundFormat 4bits | SoundRate 2bits | SoundSize 1bit | SoundType 1bits| AACPacketType 8bits|  
    
    // @see: E.4.2 Audio Tags, video_file_format_spec_v10_1.pdf, page 76
    int8_t sound_format = stream->read_1bytes();//读取SoundFormat
    
    //int8_t sound_type = sound_format & 0x01;
    //int8_t sound_size = (sound_format >> 1) & 0x01;
    //int8_t sound_rate = (sound_format >> 2) & 0x03;
    sound_format = (sound_format >> 4) & 0x0f;
    
    if ((SrsAudioCodecId)sound_format != SrsAudioCodecIdAAC) {//根据SoundFormat判断是否为AAC
        ret = ERROR_AAC_DECODE_ERROR;
        srs_error("aac required, format=%d. ret=%d", sound_format, ret);
        return ret;
    }
    
    if (!stream->require(1)) {
        ret = ERROR_AAC_DECODE_ERROR;
        srs_error("aac decode aac_packet_type failed. ret=%d", ret);
        return ret;
    }
    
    SrsAudioAacFrameTrait aac_packet_type = (SrsAudioAacFrameTrait)stream->read_1bytes();//读取包类型
    if (aac_packet_type == SrsAudioAacFrameTraitSequenceHeader) {                        //根据包类型判断后面的数据是否为AudioSpecificConfig
    	//接着从音频TAG中提取2字节的AUDIO SPECIFIC CONFIG，后面会将其转换成7字节的ADTS HEADER
    	//AudioSpecificConfig(2byte)结构定义：
    	//|audioObjectType:5bit|samplingFrequencyIndex:4bit|channelConfiguration:4bit|frameLengthFlag:1bit |dependsOnCoreCoder:1bit|extensionFlag:1bit|  

 
        // AudioSpecificConfig
        // 1.6.2.1 AudioSpecificConfig, in ISO_IEC_14496-3-AAC-2001.pdf, page 33.
        //
        // only need to decode the first 2bytes:
        // audioObjectType, 5bits.
        // samplingFrequencyIndex, aac_sample_rate, 4bits.
        // channelConfiguration, aac_channels, 4bits
        if (!stream->require(2)) {
            ret = ERROR_AAC_DECODE_ERROR;
            srs_error("aac decode sequence header failed. ret=%d", ret);
            return ret;
        }
        
        int8_t audioObjectType = stream->read_1bytes();
        aac_sample_rate = stream->read_1bytes();
        
        aac_channels = (aac_sample_rate >> 3) & 0x0f;
        aac_sample_rate = ((audioObjectType << 1) & 0x0e) | ((aac_sample_rate >> 7) & 0x01);
        
        audioObjectType = (audioObjectType >> 3) & 0x1f;
        aac_object = (SrsAacObjectType)audioObjectType;
        
        got_sequence_header = true;
        
        return ret;
    }
    
    if (!got_sequence_header) {
        ret = ERROR_AAC_DECODE_ERROR;
        srs_error("aac no sequence header. ret=%d", ret);
        return ret;
    }
    
    // the left is the aac raw frame data.
    int16_t aac_raw_length = stream->size() - stream->pos();
    
    // write the ADTS header.
    // @see ISO_IEC_14496-3-AAC-2001.pdf, page 75,
    //      1.A.2.2 Audio_Data_Transport_Stream frame, ADTS
    // @see https://github.com/ossrs/srs/issues/212#issuecomment-64145885
    // byte_alignment()
    
    // adts_fixed_header:
    //      12bits syncword,
    //      16bits left.
    // adts_variable_header:
    //      28bits
    //      12+16+28=56bits
    // adts_error_check:
    //      16bits if protection_absent
    //      56+16=72bits
    // if protection_absent:
    //      require(7bytes)=56bits
    // else
    //      require(9bytes)=72bits