准备
ffmpeg 4.4
一个MP4或flv格式的视频文件
分离流程
大致分为以下几个简单步骤：
1.使用avformat_open_input 函数打开文件并初始化结构AVFormatContext
2.查找是否存在音频和视频信息
3.构建一个h264_mp4toannexb比特流的过滤器，用来给视频avpaket包添加头信息
4.打开2个输出文件（音频， 视频）
5.循环读取视频文件，并将音视频分别写入文件
注意：音频需要手动添加头信息，没有提供aac的adts自动添加的过滤器
源码
#include <stdio.h>
extern "C"
{
#include <libavformat/avformat.h>
}
/* 打印编码器支持该采样率并查找指定采样率下标 */
static int find_sample_rate_index(const AVCodec* codec, int sample_rate)
{
const int* p = codec->supported_samplerates;
int sample_rate_index = -1; //支持的分辨率下标
int count = 0;
while (*p != 0) {// 0作为退出条件，比如libfdk-aacenc.c的aac_sample_rates
printf("%s 支持采样率: %dhz 对应下标：%d\n", codec->name, *p, count);
if (*p == sample_rate)
sample_rate_index = count;
p++;
count++;
}
return sample_rate_index;
}
/// <summary>
/// 给aac音频数据添加adts头
/// </summary>
/// <param name="header">adts数组</param>
/// <param name="sample_rate">采样率</param>
/// <param name="channals">通道数</param>
/// <param name="prfile">音频编码器配置文件（FF_PROFILE_AAC_LOW 定义在 avcodec.h）</param>
/// <param name="len">音频包长度</param>
void addHeader(char header[], int sample_rate, int channals, int prfile, int len)
{
uint8_t sampleIndex = 0;
switch (sample_rate) {
case 96000: sampleIndex = 0; break;
case 88200: sampleIndex = 1; break;
case 64000: sampleIndex = 2; break;
case 48000: sampleIndex = 3; break;
case 44100: sampleIndex = 4; break;
case 32000: sampleIndex = 5; break;
case 24000: sampleIndex = 6; break;
case 22050: sampleIndex = 7; break;
case 16000: sampleIndex = 8; break;
case 12000: sampleIndex = 9; break;
case 11025: sampleIndex = 10; break;
case 8000: sampleIndex = 11; break;
case 7350: sampleIndex = 12; break;
default: sampleIndex = 4; break;
}
uint8_t audioType = 2; //AAC LC
uint8_t channelConfig = 2; //双通道
len += 7;
//0,1是固定的
header[0] = (uint8_t)0xff; //syncword:0xfff 高8bits
header[1] = (uint8_t)0xf0; //syncword:0xfff 低4bits
header[1] |= (0 << 3); //MPEG Version:0 for MPEG-4,1 for MPEG-2 1bit
header[1] |= (0 << 1); //Layer:0 2bits
header[1] |= 1; //protection absent:1 1bit
//根据aac类型,采样率,通道数来配置
header[2] = (audioType - 1) << 6; //profile:audio_object_type - 1 2bits
header[2] |= (sampleIndex & 0x0f) << 2; //sampling frequency index:sampling_frequency_index 4bits
header[2] |= (0 << 1); //private bit:0 1bit
header[2] |= (channelConfig & 0x04) >> 2; //channel configuration:channel_config 高1bit
//根据通道数+数据长度来配置
header[3] = (channelConfig & 0x03) << 6; //channel configuration:channel_config 低2bits
header[3] |= (0 << 5); //original：0 1bit
header[3] |= (0 << 4); //home：0 1bit
header[3] |= (0 << 3); //copyright id bit：0 1bit
header[3] |= (0 << 2); //copyright id start：0 1bit
header[3] |= ((len & 0x1800) >> 11); //frame length：value 高2bits
//根据数据长度来配置
header[4] = (uint8_t)((len & 0x7f8) >> 3); //frame length:value 中间8bits
header[5] = (uint8_t)((len & 0x7) << 5); //frame length:value 低3bits
header[5] |= (uint8_t)0x1f; //buffer fullness:0x7ff 高5bits
header[6] = (uint8_t)0xfc;
}
int main() {
AVFormatContext* ifmt_ctx = NULL;
AVPacket pkt;
int ret, i;
int videoindex = -1, audioindex = -1;
const char* in_filename = "D:/测试工程/sound/beautlWorld.mp4";
const char* out_filename_v = "D:/测试工程/sound/ffmpeg_demo.h264";
const char* out_filename_a = "D:/测试工程/sound/ffmpeg_demo.aac";
if ((ret = avformat_open_input(&ifmt_ctx, in_filename, 0, 0)) < 0) {
printf("Could not open input file.");
return -1;
}
if ((ret = avformat_find_stream_info(ifmt_ctx, 0)) < 0) {
printf("Failed to retrieve input stream information");
return -1;
}
videoindex = -1;
for (i = 0; i < ifmt_ctx->nb_streams; i++) { //nb_streams：视音频流的个数
if (ifmt_ctx->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_VIDEO)
videoindex = i;
else if (ifmt_ctx->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_AUDIO)
audioindex = i;
}
printf("\nInput Video===========================\n");
av_dump_format(ifmt_ctx, 0, in_filename, 0); // 打印信息
printf("\n======================================\n");
FILE* fp_audio = fopen(out_filename_a, "wb+");
FILE* fp_video = fopen(out_filename_v, "wb+");
AVBSFContext* bsf_ctx = NULL;
const AVBitStreamFilter* pfilter = av_bsf_get_by_name("h264_mp4toannexb");
if (pfilter == NULL) {
printf("Get bsf failed!\n");
}
if ((ret = av_bsf_alloc(pfilter, &bsf_ctx)) != 0) {
printf("Alloc bsf failed!\n");
}
ret = avcodec_parameters_copy(bsf_ctx->par_in, ifmt_ctx->streams[videoindex]->codecpar);
if (ret < 0) {
printf("Set Codec failed!\n");
}
ret = av_bsf_init(bsf_ctx);
if (ret < 0) {
printf("Init bsf failed!\n");
}
//这里遍历音频编码器打印支持的采样率，并找到当前音频采样率所在的下表，用于后面添加adts头
//本程序并没有使用，只是测试，如果为了程序健壮性可以采用此方式
AVCodec* codec = nullptr;
codec = avcodec_find_encoder(ifmt_ctx->streams[audioindex]->codecpar->codec_id);
int sample_rate_index = find_sample_rate_index(codec, ifmt_ctx->streams[audioindex]->codecpar->sample_rate);
printf("分辨率数组下表：%d\n", sample_rate_index);
while (av_read_frame(ifmt_ctx, &pkt) >= 0) {
if (pkt.stream_index == videoindex) {
av_bsf_send_packet(bsf_ctx, &pkt);
while (true)
{
ret = av_bsf_receive_packet(bsf_ctx, &pkt);
if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF)
break;
else if (ret < 0) {
printf("Receive Pkt failed!\n");
break;
}
printf("Write Video Packet. size:%d\tpts:%lld\n", pkt.size, pkt.pts);
fwrite(pkt.data, 1, pkt.size, fp_video);
}
}
else if (pkt.stream_index == audioindex) {
printf("Write Audio Packet. size:%d\tpts:%lld\n", pkt.size, pkt.pts);
char adts[7] = { 0 };
addHeader(adts, ifmt_ctx->streams[audioindex]->codecpar->sample_rate,
ifmt_ctx->streams[audioindex]->codecpar->channels,
ifmt_ctx->streams[audioindex]->codecpar->profile,
pkt.size);
fwrite(adts, 1, 7, fp_audio);
fwrite(pkt.data, 1, pkt.size, fp_audio);
}
av_packet_unref(&pkt);
}
av_bsf_free(&bsf_ctx);
fclose(fp_video);
fclose(fp_audio);
avformat_close_input(&ifmt_ctx);
return 0;
if (ifmt_ctx)
avformat_close_input(&ifmt_ctx);
if (fp_audio)
fclose(fp_audio);
if (fp_video)
fclose(fp_video);
if (bsf_ctx)
av_bsf_free(&bsf_ctx);
return -1;
}
小记
1.av_read_frame 就是读取文件并返回下一帧
2.视频的头信息不在avpacket中需要使用bsf过滤器来添加
3.aac音频头信息需要手动添加
扫一扫
1754
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
