There is no way to play the AAC data of the naked stream, because the device cannot find the relevant information of the AAC format; it needs to add the ADTS header to be able to play, and each frame of audio package has an ADTS header, and the ADTS header has 7 characters if there is no crc check Section length, if there is crc check, it is 9 bytes long.
ADTS field protocol:
| field | length | describe |
| synword | 12bit | Fixed 0xFFF, used for synchronization, the beginning of a frame |
| id | 1bit | MPEG identifier, 0: MPEG-4, 1: MPEG-2 |
| layer | 2bit |
Usually 00 |
| protection_absent | 1bit | CRC verification flag, 0: with CRC verification, 1: without CRC verification |
| profile | 2bit | AAC level, in AVStream of ffmpeg: streams[audio]->codecpar->profile audio: frame index |
| sampling_frequency_index | 4 bits | Sampling rate subscript, the sampling rate corresponding to the subscript is as follows: 0: 96000 Hz streams[audio]->codecpar->sample_rate audio: frame index |
| private_bit | 1bit | Private bit, 0 when encoding, ignored when decoding |
| channel_configuration | 3bit | 声道数。 front - left: left channel front - right：右声道 back - left：后置左 back - right：后置右 side - left：侧置左 side - right：侧置右 LFE - channel：低频声道 音频通道也在ffmpeg的AVStream中： streams[audio]->codecpar->channels audio:帧索引 |
| orininal_copy | 1bit | 编码是设置为0，解码时忽略 |
| home | 1bit | 编码时设置为0，解码时忽略 |
| copyrigth_identification_bit | 1bit | 编码时设置为0，解码时忽略 |
| copyrigth_identification_stat | 1bit | 编码时设置为0，解码时忽略 |
| aac_frame_length | 13bit | 一个ADTS帧的⻓度，包括ADTS头和AAC原始流。 |
| adts_bufferfullness | 11bit | 缓冲区充满度，0x7FF说明是码率可变的码流，不需要此字段。CBR可能需要此字段，不同编码器使用情况不同。具体查看附录。 |
| number_of_raw_data_blocks_in_frame | 2bit | 表示ADTS帧中有number_of_raw_data_blocks_in_frame + 1个AAC原始帧，为0表示说ADTS帧中只有一个AAC数据. |
| crc | 16bit | protection_absent为0就有该字段，否则没有该字段 |
ffmpeg解复用时，MP4，FLV格式的包解出的音频流是纯AAC流，不带ADTS头数据，需要人工添加ADTS头。
添加头的代码实现接口：
int adts_header(char *const p_adts_header, const int data_length,
const int profile, const int samplerate,
const int channels) {
int sampling_frequency_index = 3; // 默认使用48000hz
int adtsLen = data_length + 7;
// 匹配采样率
int frequencies_size = sizeof(sampling_frequencies) / sizeof(sampling_frequencies[0]);
int i = 0;
for (i = 0; i < frequencies_size; i++) {
if (sampling_frequencies[i] == samplerate) {
sampling_frequency_index = i;
break;
}
}
if (i >= frequencies_size) {
std::cout << "没有找到支持的采样率" << std::endl;
return -1;
}
p_adts_header[0] = 0xff; //前12bit固定0xfff 高8bits
p_adts_header[1] = 0xf0; //前12bit的低四位 低4bits
p_adts_header[1] |= (0 << 3); //0：MPEG-4, 1：MPEG-2 1bit
p_adts_header[1] |= (0 << 1); //一般为0 2bits
p_adts_header[1] |= 1; //1：没有crc校验字段 1bit
p_adts_header[2] = (profile) << 6; //aac级别，可以使用ffmpeg获取 2bits
p_adts_header[2] |=
(sampling_frequency_index & 0x0f) << 2; //可以使用ffgmpeg从包中获得 4bits
p_adts_header[2] |= (0 << 1); //私有位 编码时为0 1bit
p_adts_header[2] |= (channels & 0x04) >> 2; //3bit的声道设置的最高位 高1bit
p_adts_header[3] = (channels & 0x03) << 6; //3bit的声道设置的最低两位 低2bits
p_adts_header[3] |= (0 << 5); //编码设置为0 1bit
p_adts_header[3] |= (0 << 4); //编码设置为0 1bit
p_adts_header[3] |= (0 << 3); //编码设置为0 1bit
p_adts_header[3] |= (0 << 2); //编码设置为0 1bit
p_adts_header[3] |= ((adtsLen & 0x1800) >> 11); //帧长度包括ADTS头长度 高2bits
p_adts_header[4] = (uint8_t) ((adtsLen & 0x7f8) >> 3); //帧长度包括ADTS头长度 中间8bits
p_adts_header[5] = (uint8_t) ((adtsLen & 0x7) << 5); //帧长度包括ADTS头长度 低3bits
p_adts_header[5] |= 0x1f; //可变码率vbr:0x7ff 高5bits
p_adts_header[6] = 0xfc; //11111100 //buffer fullness:0x7ff 低6bits
return 0;
}代码中的sampling_frequencies[]数组是一个sampling_frequency_index字段的采样率对照表 。