FFmpeg源码地址:https://github.com/FFmpeg/FFmpeg
FFmpeg可执行文件地址:https://ffmpeg.org/download.html
◆ 用于音视频转码, 也可以从url/现场音频/视频源抓取输入源等。笔者从网上摘抄了一部分ffmpeg常用参数如下(尤其在开发过程中,由于ffmpeg版本不同,ffmpeg参数也有少量出入,建议在命令行窗口输入“ffmpeg -h”查看本机部署的ffmpeg支持的参数):
ffmpeg 范例
//从视频第3秒开始剪切,持续4秒,并保存文件
ffmpeg -ss 00:00:03 -t 00:00:04 -i test.mpg -vcodec copy -acodec copy test_cut.mpg
◆ 一个非常简单和可移植的媒体播放器,使用FFmpeg库和SDL库。ffplay参数如下:
ffplay.exe范例:
//播放视频
ffplay -i test.mpg
◆ 查看多媒体文件的信息。ffprobe参数如下
◆ ffprobe.exe范例:
//查看视频文件的音频流和视频流信息
ffprobe test.mpg
-
- /*
- author:八小时码字员
- file:vPlayer_sdl2.cpp
- */
- #include "vPlayer_sdl2.h"
- #include "output_log.h"
-
- #include <stdio.h>
- #include <iostream>
- using namespace std;
-
- #define __STDC_CONSTANT_MACROS
- extern "C"
- {
- #include <libavcodec/avcodec.h>
- #include <libavformat/avformat.h>
- #include <libswscale/swscale.h>
- #include <libavutil/imgutils.h>
- #include <SDL.h>
- }
-
- static int g_frame_rate = 1;
- static int g_sfp_refresh_thread_exit = 0;
- static int g_sfp_refresh_thread_pause = 0;
-
- #define SFM_REFRESH_EVENT (SDL_USEREVENT+1)
- #define SFM_BREAK_EVENT (SDL_USEREVENT+2)
-
- typedef struct FFmpeg_V_Param_T
- {
- AVFormatContext *pFormatCtx;
- AVCodecContext *pCodecCtx;
- SwsContext *pSwsCtx;
- int video_index;
- }FFmpeg_V_Param;
-
- typedef struct SDL_Param_T
- {
- SDL_Window *p_sdl_window;
- SDL_Renderer *p_sdl_renderer;
- SDL_Texture *p_sdl_texture;
- SDL_Rect sdl_rect;
- SDL_Thread *p_sdl_thread;
- }SDL_Param;
-
- /*
- return value:zero(success) non-zero(failure)
- */
- int init_ffmpeg(FFmpeg_V_Param* p_ffmpeg_param, char* filePath)
- {
- //init FFmpeg_V_Param
- p_ffmpeg_param->pFormatCtx = avformat_alloc_context();
- const AVCodec *pCodec = NULL;
-
- //do global initialization of network libraries
- avformat_network_init();
-
- //open input stream
- if (avformat_open_input(&(p_ffmpeg_param->pFormatCtx), filePath, NULL, NULL) != 0)
- {
- output_log(LOG_ERROR, "avformat_open_input error");
- return -1;
- }
-
- //find stream info
- if (avformat_find_stream_info(p_ffmpeg_param->pFormatCtx, NULL) < 0)
- {
- output_log(LOG_ERROR, "avformat_find_stream_info error");
- return -1;
- }
-
- //get video pCodecParms, codec and frame rate
- for (int i = 0; i < p_ffmpeg_param->pFormatCtx->nb_streams; i++)
- {
- AVStream *pStream = p_ffmpeg_param->pFormatCtx->streams[i];
- if (pStream->codecpar->codec_type == AVMEDIA_TYPE_VIDEO)
- {
- pCodec = avcodec_find_decoder(pStream->codecpar->codec_id);
- p_ffmpeg_param->pCodecCtx = avcodec_alloc_context3(pCodec);
- avcodec_parameters_to_context(p_ffmpeg_param->pCodecCtx, pStream->codecpar);
- g_frame_rate = pStream->avg_frame_rate.num / pStream->avg_frame_rate.den;
- p_ffmpeg_param->video_index = i;
- }
- }
- if (!p_ffmpeg_param->pCodecCtx)
- {
- output_log(LOG_ERROR, "could not find video codecCtx");
- return -1;
- }
-
- //open codec
- if (avcodec_open2(p_ffmpeg_param->pCodecCtx, pCodec, NULL))
- {
- output_log(LOG_ERROR, "avcodec_open2 error");
- return -1;
- }
-
- //get scale pixelformat context
- p_ffmpeg_param->pSwsCtx = sws_getContext(p_ffmpeg_param->pCodecCtx->width,
- p_ffmpeg_param->pCodecCtx->height, p_ffmpeg_param->pCodecCtx->pix_fmt,
- p_ffmpeg_param->pCodecCtx->width, p_ffmpeg_param->pCodecCtx->height,
- AV_PIX_FMT_YUV420P, SWS_BICUBIC, NULL, NULL, NULL);
-
- av_dump_format(p_ffmpeg_param->pFormatCtx, p_ffmpeg_param->video_index, filePath, 0);
-
- return 0;
- }
-
- /*
- return value:zero(success) non-zero(failure)
- */
- int release_ffmpeg(FFmpeg_V_Param* p_ffmpeg_param)
- {
- if (!p_ffmpeg_param)
- return -1;
-
- //realse scale pixelformat context
- if (p_ffmpeg_param->pSwsCtx)
- sws_freeContext(p_ffmpeg_param->pSwsCtx);
-
- //close codec
- if (p_ffmpeg_param->pCodecCtx)
- avcodec_close(p_ffmpeg_param->pCodecCtx);
-
- //close input stream
- if (p_ffmpeg_param->pFormatCtx)
- avformat_close_input(&(p_ffmpeg_param->pFormatCtx));
-
- //free AVCodecContext
- if (p_ffmpeg_param->pCodecCtx)
- avcodec_free_context(&(p_ffmpeg_param->pCodecCtx));
-
- //free AVFormatContext
- if (p_ffmpeg_param->pFormatCtx)
- avformat_free_context(p_ffmpeg_param->pFormatCtx);
-
- //free FFmpeg_V_Param
- delete p_ffmpeg_param;
- p_ffmpeg_param = NULL;
-
- return 0;
- }
-
- int sfp_refresh_thread(void* opaque)
- {
- g_sfp_refresh_thread_exit = 0;
- g_sfp_refresh_thread_pause = 0;
- while (!g_sfp_refresh_thread_exit)
- {
- if (!g_sfp_refresh_thread_pause)
- {
- SDL_Event sdl_event;
- sdl_event.type = SFM_REFRESH_EVENT;
- SDL_PushEvent(&sdl_event);
- }
- SDL_Delay(1000 / g_frame_rate);
- }
- g_sfp_refresh_thread_exit = 0;
- g_sfp_refresh_thread_pause = 0;
- SDL_Event sdl_event;
- sdl_event.type = SFM_BREAK_EVENT;
- SDL_PushEvent(&sdl_event);
- return 0;
- }
-
- int init_sdl2(SDL_Param_T *p_sdl_param, int screen_w, int screen_h)
- {
- if (SDL_Init(SDL_INIT_AUDIO | SDL_INIT_VIDEO | SDL_INIT_TIMER))
- {
- output_log(LOG_ERROR, "SDL_Init error");
- return -1;
- }
- p_sdl_param->p_sdl_window = SDL_CreateWindow("vPlayer_sdl", SDL_WINDOWPOS_UNDEFINED,
- SDL_WINDOWPOS_UNDEFINED, screen_w, screen_h, SDL_WINDOW_OPENGL);
- if (!p_sdl_param->p_sdl_window)
- {
- output_log(LOG_ERROR, "SDL_CreateWindow error");
- return -1;
- }
- p_sdl_param->p_sdl_renderer = SDL_CreateRenderer(p_sdl_param->p_sdl_window, -1, 0);
- p_sdl_param->p_sdl_texture = SDL_CreateTexture(p_sdl_param->p_sdl_renderer, SDL_PIXELFORMAT_IYUV,
- SDL_TEXTUREACCESS_STREAMING, screen_w, screen_h);
- p_sdl_param->sdl_rect.x = 0;
- p_sdl_param->sdl_rect.y = 0;
- p_sdl_param->sdl_rect.w = screen_w;
- p_sdl_param->sdl_rect.h = screen_h;
- p_sdl_param->p_sdl_thread = SDL_CreateThread(sfp_refresh_thread, NULL, NULL);
-
- return 0;
- }
-
- int release_sdl2(SDL_Param_T *p_sdl_param)
- {
- SDL_DestroyTexture(p_sdl_param->p_sdl_texture);
- SDL_DestroyRenderer(p_sdl_param->p_sdl_renderer);
- SDL_DestroyWindow(p_sdl_param->p_sdl_window);
- SDL_Quit();
- return 0;
- }
-
- int vPlayer_sdl2(char* filePath)
- {
- //ffmpeg param
- FFmpeg_V_Param *p_ffmpeg_param = NULL;
- AVPacket *packet = NULL;
- AVFrame *pFrame = NULL, *pFrameYUV = NULL;
- int out_buffer_size = 0;
- unsigned char* out_buffer = 0;
-
- //sdl param
- SDL_Param_T *p_sdl_param = NULL;
- SDL_Event sdl_event;
-
- int ret = 0;
-
- //init ffmpeg
- p_ffmpeg_param = new FFmpeg_V_Param();
- memset(p_ffmpeg_param, 0, sizeof(FFmpeg_V_Param));
- if (init_ffmpeg(p_ffmpeg_param, filePath))
- {
- ret = -1;
- goto end;
- }
- packet = av_packet_alloc();
- pFrame = av_frame_alloc();
- pFrameYUV = av_frame_alloc();
- out_buffer_size = av_image_get_buffer_size(AV_PIX_FMT_YUV420P,
- p_ffmpeg_param->pCodecCtx->width, p_ffmpeg_param->pCodecCtx->height, 1);
- out_buffer = (unsigned char*)av_malloc(out_buffer_size);
- av_image_fill_arrays(pFrameYUV->data, pFrameYUV->linesize, out_buffer,
- p_ffmpeg_param->pCodecCtx->pix_fmt,
- p_ffmpeg_param->pCodecCtx->width, p_ffmpeg_param->pCodecCtx->height, 1);
- //init sdl2
- p_sdl_param = new SDL_Param_T();
- memset(p_sdl_param, 0, sizeof(SDL_Param_T));
- if (init_sdl2(p_sdl_param, p_ffmpeg_param->pCodecCtx->width, p_ffmpeg_param->pCodecCtx->height))
- {
- ret = -1;
- goto end;
- }
-
- //demuxing and show
- while (true)
- {
- int temp_ret = 0;
- SDL_WaitEvent(&sdl_event);
- if (sdl_event.type == SFM_REFRESH_EVENT)
- {
- while (true)
- {
- if (av_read_frame(p_ffmpeg_param->pFormatCtx, packet) < 0)
- {
- g_sfp_refresh_thread_exit = 1;
- break;
- }
- if (packet->stream_index == p_ffmpeg_param->video_index)
- {
- break;
- }
- }
- if (avcodec_send_packet(p_ffmpeg_param->pCodecCtx, packet))
- g_sfp_refresh_thread_exit = 1;
-
- do
- {
- temp_ret = avcodec_receive_frame(p_ffmpeg_param->pCodecCtx, pFrame);
- if (temp_ret == AVERROR_EOF)
- {
- g_sfp_refresh_thread_exit = 1;
- break;
- }
- if (temp_ret == 0)
- {
- sws_scale(p_ffmpeg_param->pSwsCtx, (const unsigned char* const*)pFrame->data,
- pFrame->linesize, 0, p_ffmpeg_param->pCodecCtx->height, pFrameYUV->data,
- pFrameYUV->linesize);
-
- SDL_UpdateTexture(p_sdl_param->p_sdl_texture, &(p_sdl_param->sdl_rect),
- pFrameYUV->data[0], pFrameYUV->linesize[0]);
-
- SDL_RenderClear(p_sdl_param->p_sdl_renderer);
- SDL_RenderCopy(p_sdl_param->p_sdl_renderer, p_sdl_param->p_sdl_texture,
- NULL, &(p_sdl_param->sdl_rect));
- SDL_RenderPresent(p_sdl_param->p_sdl_renderer);
- }
- } while (temp_ret != AVERROR(EAGAIN));
-
- //av_packet_unref(packet);
- }
- else if (sdl_event.type == SFM_BREAK_EVENT)
- {
- break;
- }
- else if (sdl_event.type == SDL_KEYDOWN)
- {
- if (sdl_event.key.keysym.sym == SDLK_SPACE)
- g_sfp_refresh_thread_pause = !g_sfp_refresh_thread_pause;
- if (sdl_event.key.keysym.sym == SDLK_q)
- g_sfp_refresh_thread_exit = 1;
- }
- else if (sdl_event.type == SDL_QUIT)
- {
- g_sfp_refresh_thread_exit = 1;
- }
- }
-
- end:
- release_ffmpeg(p_ffmpeg_param);
- av_packet_free(&packet);
- av_frame_free(&pFrame);
- av_frame_free(&pFrameYUV);
- release_sdl2(p_sdl_param);
- return ret;
- }
- /*
- author:八小时码字员
- file:output_log.cpp
- */
- #include
- #include
-
- #define MAX_BUF_LEN 1024
-
- int g_log_debug_flag = 1;
- int g_log_info_flag = 1;
- int g_log_warnning_flag = 1;
- int g_log_error_flag = 1;
-
- enum LOG_LEVEL
- {
- LOG_DEBUG,
- LOG_INFO,
- LOG_WARNING,
- LOG_ERROR
- };
-
- void set_log_flag(int log_debug_flag, int log_info_flag, int log_warnning_flag,
- int log_error_flag)
- {
- g_log_debug_flag = log_debug_flag;
- g_log_info_flag = log_info_flag;
- g_log_warnning_flag = log_warnning_flag;
- g_log_error_flag = log_error_flag;
- }
-
- void output_log(LOG_LEVEL log_level, const char* fmt, ...)
- {
- va_list args;
- va_start(args, fmt);
-
- char buf[MAX_BUF_LEN] = { 0 };
- vsnprintf(buf, MAX_BUF_LEN - 1, fmt, args);
-
- switch (log_level)
- {
- case LOG_DEBUG:
- if (g_log_debug_flag)
- printf("[Log-Debug]:%s\n", buf);
- break;
- case LOG_INFO:
- if (g_log_info_flag)
- printf("[Log-Info]:%s\n", buf);
- break;
- case LOG_WARNING:
- if (g_log_warnning_flag)
- printf("[Log-Warnning]:%s\n", buf);
- break;
- case LOG_ERROR:
- if (g_log_error_flag)
- printf("[Log-Error]:%s\n", buf);
- break;
- default:
- break;
- }
-
- va_end(args);
- }
转码比较好理解,就是将解码和编码结合起来,过程为:解封装->解码->编码->封装。逻辑如下:
◆ 解封装:将音视频文件的封装格式去掉,获取视频流(H.264)和音频流(AAC)
◆ 解码:将视频流解码成原始图像数据(YUV),将音频流解码成原始音频数据(PCM)
◆ 编码:将原始图像(YUV)进行编码(MPG2),将音频流进行编码(MP3)
◆ 封装:将视频流和音频流封装成视频文件
FFmpeg结构体主要分为三个层次:协议层(AVIOContext)、封装层(AVInputFormat)、解码层(AVStream)。具体如下:
关键结构体主要有以下八个:
描述媒体文件或媒体流的构成和基本信息,贯穿ffmpeg使用整个流程
- AVInputFormat *iformat、AVOutputFormat *oformat:输入或者输出流的格式(只能存在一个)
- AVIOContext *pb:管理输入输出数据
- unsigned int nb_streams:音视频流的个数
- AVStream **streams:音视频流
- char *url:文件名
- int64_t duration:时长
- int bit_rate:比特率(单位bite/s)
- AVDictionary *metadata:元数据(查看元数据:ffprobe filename)
文件的封装格式
- char* name:封装格式的名字
- char* long_name:封装格式的长名字
- char* extensions:文件扩展名
文件(协议)操作的顶层对象
- unsigned char *buffer:缓冲开始位置
- int buffer_size:缓冲区大小(默认32768)
- unsigned char *buf_ptr:当前指针读取到的位置
- unsigned char *buf_end:缓存结束的位置
- void *opaque:URLContext结构体
- (*read_packet)(...):读取音视频数据的函数指针
- (*write_packet)(...):写入音视频数据的函数指针
- (*read_pause)(...):网络流媒体协议的暂停或恢复播放函数指针
每种协议,有一个协议操作对象和一个关联的协议对象
- char* name:协议名称
- const struct URLProtocol *prot:协议操作对象(ff_file_protocol、ff_librtmp_protocol...)
- void *priv_data:协议对象(FileContext、LibRTMPContext)
协议操作对象
存储音频流或视频流的结构体
- int index:音频流或视频流的索引
- AVRational time_base:计算pts或dts是使用的时间戳基本单位(显示时间:pt = av_q2d(video_stream->time_base) * frame->pts)
- int64_t duration:该视频/音频流长度
- AVRational avg_frame_rate:平均帧率(对于视频来说,frame_rate=avg_frame_rate.num / avg_frame_rate.den)
- AVCodecParameters *codecpar:解码器参数
◆新的 ffmpeg 中 AVStream.codecpar(struct AVCodecParameter) 代替 AVStream.codec(struct AVCodecContext):AVCodecParameter 是由 AVCodecContext 分离出来的,AVCodecParameter中没有函数
◆AVCodecContext 结构体仍然是编解码时不可或缺的结构体:avcodec_send_packet 和 avcodec_receive_frame 使用 AVCodecContext
◆ AVCodecContext 和 AVCodec 的获取方法
- //=============== new version code ===============
- char filePath[] = "test.mp4";
- AVFormatContext *pFormatCtx;
- AVCodecContext *pCodecCtx;
- AVCodec *pCodec;
- pFormatCtx = avformat_alloc_context();
-
- av_register_all();
- avformat_network_init();
- avformat_open_input(&pFormatCtx, filePath, NULL, NULL);
- avformat_find_stream_info(pFormatCtx, NULL);
- for (int i = 0; i < pFormatCtx->nb_streams; i++)
- {
- AVStream *pStream = pFormatCtx->streams[i];
- pCodec = avcodec_find_decoder(pStream->codecpar->codec_id);
- pCodecCtx = avcodec_alloc_context3(pCodec);
- avcodec_parameters_to_context(pCodecCtx, pStream->codecpar);
- }
- //============ old version code ============
- char filePath[] = "test.mp4";
- AVFormatContext *pFormatCtx;
- AVCodecContext *pCodecCtx;
- AVCodec *pCodec;
- pFormatCtx = avformat_alloc_context();
-
- av_register_all();
- avformat_network_init();
- avformat_open_input(&pFormatCtx, filePath, NULL, NULL);
- avformat_find_stream_info(pFormatCtx, NULL);
- for (int i = 0; i < pFormatCtx->nb_streams; i++)
- {
- AVStream *pStream = pFormatCtx->streams[i];
- pCodecCtx = pStream->codec;
- pCodec = avcodec_find_decoder(pCodecCtx->codec_id);
- }
◆ 关键参数:可参考 avcodec_parameters_to_context 源码
- num AVMediaType codec_type:编解码器的类型(视频,音频...)
- enum AVCodecID codec_id:标示特定的编码器
- AVCodecContext:struct AVCodec *codec:采用的解码器AVCodec(H.264,MPEG2...)
- int bit_rate:平均比特率
- uint8_t *extradata; int extradata_size:针对特定编码器包含的附加信息(例如对于H.264解码器来说,存储SPS,PPS等)
- AVCodecContext:enum AVPixelFormat pix_fmt:像素格式(视频)
- int width, height:宽和高(视频)
- AVCodecContext:enum AVSampleFormat sample_fmt:采样格式(音频)
- int sample_rate:采样率(音频)
- int channels:声道数(音频)
- uint64_t channel_layout:声道格式
- AVCodecParameters:int format:像素格式(视频)/采样格式(音频
编解码器结构体
◆ 每一个解码器对应一个AVCodec结构体,比如:AVCodec ff_h264_decoder,AVCodec ff_jpeg2000_decoder
◆ 关键成员变量如下:
- const char *name:编解码器短名字(形如:"h264")
- const char *long_name:编解码器全称(形如:"H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10")
- enum AVMediaType type:媒体类型:视频、音频或字母
- enum AVCodecID id:标示特定的编码器
- const AVRational *supported_framerates:支持的帧率(仅视频)
- const enum AVPixelFormat *pix_fmts:支持的像素格式(仅视频)
- const int *supported_samplerates:支持的采样率(仅音频)
- const enum AVSampleFormat *sample_fmts:支持的采样格式(仅音频)
- const uint64_t *channel_layouts:支持的声道数(仅音频
存储解码前数据的结构体
◆ 关键成员变量
- AVBufferRef *buf:管理data指向的数据
- uint8_t *data:压缩编码的数据
- int size:data的大小
- int64_t pts:显示时间戳
- int64_t dts:解码时间戳
- int stream_index:标识该AVPacket所属的视频/音频流
◆ AVPacket的内存管理:AVPacket本身并不包含压缩的数据,通过data指针引用数据的缓存空间
◆ 多个AVPacket共享同一个数据缓存(AVBufferRef、AVBuffer)
◆ AVPacket拥有独立的数据缓存
- av_read_frame(pFormatCtx, packet) // 读取Packet
- av_packet_ref(dst_pkt,packet) // dst_pkt 和 packet 共享同一个数据缓存空间,引用计数+1
- av_packet_unref(dst_pkt); //释放 pkt_pkt 引用的数据缓存空间,引用计数-1
◆ AVBuffer 关键成员变量
- uint8_t *data:压缩编码的数据
- size_t size:数据长度
- atomic_uint refcount:引用计数,如果引用计数为0,则释放数据缓存空间
存储解码后数据的结构体
- uint8_t *data[AV_NUM_DATA_POINTERS]:解码后原始数据(对视频来说是YUV,RGB,对音频来说是PCM)
- int linesize[AV_NUM_DATA_POINTERS]:data中“一行”数据的大小。注意:未必等于图像的宽,一般大于图像的宽。
- int width, height:视频帧宽和高(1920x1080,1280x720...)
- int nb_samples:音频的一个AVFrame中可能包含多个音频帧,在此标记包含了几个
- int format:解码后原始数据类型(YUV420,YUV422,RGB24...)
- int key_frame:是否是关键帧
- enum AVPictureType pict_type:帧类型(I,B,P...)
- AVRational sample_aspect_ratio:图像宽高比(16:9,4:3...)
- int64_t pts:显示时间戳
- int coded_picture_number:编码帧序号
- int display_picture_number:显示帧序号