• 在visual studio 2010中调用ffmpeg


    转自:http://blog.sina.com.cn/s/blog_4178f4bf01018wqh.html

     最近几天一直在折腾ffmpeg,在网上也查了许多资料,费了不少劲,现在在这里和大家分享一下。

    一、准备工作
        本来是想自己在windows下编译ffmpeg生成lib、dll等库文件的,但是折腾好久总是出错,于是果断放弃。幸好网上已经有了编译好的版本,可以拿过来直接用,网址为http://ffmpeg.zeranoe.com/builds/。我们需要的是32-bit Builds (Shared)和32-bit Builds (Dev),版本号要相对应。32-bit Builds (Shared)主要包含了所要用到的dll文件,32-bit Builds (Dev)主要包含了所要用的头文件和lib文件。其实这些lib并不是传统的静态库文件(真正的静态库文件是在lib目录下的*.a文件),他们是dll的导出文件。dll文件在32-bit Builds (Shared)的bin目录下,lib文件在32-bit Builds (Dev)的lib目录下,头文件在32-bit Builds (Dev)的include目录下。
        另外,C99中添加了几个新的头文件,Visual Studio中没有,所以需要你自己下载。并放至相应目录。对于VS2010来说通常是:C:Program Files (x86)Microsoft Visual Studio 10.0VCinclude。
    二、Visual Studio配置
        在准备好库文件之后,便可以进行Visual Studio简单的编译工作了,下面给出一份当前可以编译运行的代码,并以此为示例。
    // ffmpeg-example.cpp : Defines the entry point for the console application.
    //
    //#include "stdafx.h"
     
    #define inline _inline
    #ifndef INT64_C
    #define INT64_C(c) (c ## LL)
    #define UINT64_C(c) (c ## ULL)
    #endif
     
    #ifdef __cplusplus
    extern "C" {
    #endif
       
    #include <libavformat/avformat.h>
    #include <libavcodec/avcodec.h>
    #include <libswscale/swscale.h>
    #ifdef __cplusplus
    }
    #endif
     
    #include <stdio.h>
     
     
    static void SaveFrame(AVFrame *pFrame, int width, int height, int iFrame);
     
    int main (int argc, const char * argv[])
    {
        AVFormatContext *pFormatCtx;
        int             i, videoStream;
        AVCodecContext  *pCodecCtx;
        AVCodec         *pCodec;
        AVFrame         *pFrame; 
        AVFrame         *pFrameRGB;
        AVPacket        packet;
        int             frameFinished;
        int             numBytes;
        uint8_t         *buffer;
     
        // Register all formats and codecs
        av_register_all();
     
        // Open video file
        //if(avformat_open_input(&pFormatCtx, argv[1], NULL, NULL)!=0)
    if(avformat_open_input(NULL, argv[1], NULL, NULL)!=0)
            return -1; // Couldn't open file
     
        // Retrieve stream information
        if(av_find_stream_info(pFormatCtx)<0)
            return -1; // Couldn't find stream information
     
        // Dump information about file onto standard error
        av_dump_format(pFormatCtx, 0, argv[1], false);
     
        // Find the first video stream
        videoStream=-1;
        for(i=0; i<pFormatCtx->nb_streams; i++)
            if(pFormatCtx->streams[i]->codec->codec_type==AVMEDIA_TYPE_VIDEO)
            {
                videoStream=i;
                break;
            }
            if(videoStream==-1)
                return -1; // Didn't find a video stream
     
            // Get a pointer to the codec context for the video stream
            pCodecCtx=pFormatCtx->streams[videoStream]->codec;
     
            // Find the decoder for the video stream
            pCodec=avcodec_find_decoder(pCodecCtx->codec_id);
            if(pCodec==NULL)
                return -1; // Codec not found
     
            // Open codec
            if(avcodec_open(pCodecCtx, pCodec)<0)
                return -1; // Could not open codec
     
            // Hack to correct wrong frame rates that seem to be generated by some codecs
            if(pCodecCtx->time_base.num>1000 && pCodecCtx->time_base.den==1)
                pCodecCtx->time_base.den=1000;
     
            // Allocate video frame
            pFrame=avcodec_alloc_frame();
     
            // Allocate an AVFrame structure
            pFrameRGB=avcodec_alloc_frame();
            if(pFrameRGB==NULL)
                return -1;
     
            // Determine required buffer size and allocate buffer
            numBytes=avpicture_get_size(PIX_FMT_RGB24, pCodecCtx->width,
                pCodecCtx->height);
     
            //buffer=malloc(numBytes);
            buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));
     
            // Assign appropriate parts of buffer to image planes in pFrameRGB
            avpicture_fill((AVPicture *)pFrameRGB, buffer, PIX_FMT_RGB24,
                pCodecCtx->width, pCodecCtx->height);
     
            // Read frames and save first five frames to disk
            i=0;
            while(av_read_frame(pFormatCtx, &packet)>=0)
            {
                // Is this a packet from the video stream?
                if(packet.stream_index==videoStream)
                {
                    // Decode video frame
                    avcodec_decode_video2(pCodecCtx, pFrame, &frameFinished, &packet);
     
                    // Did we get a video frame?
                    if(frameFinished)
                    {
                        static struct SwsContext *img_convert_ctx;
     
    #if 0
                        // Older removed code
                        // Convert the image from its native format to RGB swscale
                        img_convert((AVPicture *)pFrameRGB, PIX_FMT_RGB24, 
                            (AVPicture*)pFrame, pCodecCtx->pix_fmt, pCodecCtx->width, 
                            pCodecCtx->height);
     
                        // function template, for reference
                        int sws_scale(struct SwsContext *context, uint8_t* src[], int srcStride[], int srcSliceY,
                            int srcSliceH, uint8_t* dst[], int dstStride[]);
    #endif
                        // Convert the image into YUV format that SDL uses
                        if(img_convert_ctx == NULL) {
                            int w = pCodecCtx->width;
                            int h = pCodecCtx->height;
     
                            img_convert_ctx = sws_getContext(w, h, 
                                pCodecCtx->pix_fmt, 
                                w, h, PIX_FMT_RGB24, SWS_BICUBIC,
                                NULL, NULL, NULL);
                            if(img_convert_ctx == NULL) {
                                fprintf(stderr, "Cannot initialize the conversion context! ");
                                exit(1);
                            }
                        }
                        int ret = sws_scale(img_convert_ctx, pFrame->data, pFrame->linesize, 0, 
                            pCodecCtx->height, pFrameRGB->data, pFrameRGB->linesize);
    #if 0 
                        // this use to be true, as of 1/2009, but apparently it is no longer true in 3/2009
                        if(ret) {
                            fprintf(stderr, "SWS_Scale failed [%d]! ", ret);
                            exit(-1);
                        }
    #endif
                        // Save the frame to disk
                        if(i++<=5)
                            SaveFrame(pFrameRGB, pCodecCtx->width, pCodecCtx->height, i);
                    }
                }
     
                // Free the packet that was allocated by av_read_frame
                av_free_packet(&packet);
            }
     
            // Free the RGB image
            //free(buffer);
            av_free(buffer);
            av_free(pFrameRGB);
     
            // Free the YUV frame
            av_free(pFrame);
     
            // Close the codec
            avcodec_close(pCodecCtx);
     
            // Close the video file
            av_close_input_file(pFormatCtx);
     
            return 0;
    }
     
    static void SaveFrame(AVFrame *pFrame, int width, int height, int iFrame)
    {
        FILE *pFile;
        char szFilename[32];
        int  y;
     
        // Open file
        sprintf(szFilename, "frame%d.ppm", iFrame);
        pFile=fopen(szFilename, "wb");
        if(pFile==NULL)
            return;
     
        // Write header
        fprintf(pFile, "P6 %d %d 255 ", width, height);
     
        // Write pixel data
        for(y=0; y<height; y++)
            fwrite(pFrame->data[0]+y*pFrame->linesize[0], 1, width*3, pFile);
     
        // Close file
        fclose(pFile);
    }
    首先在Visual Studio中新建工程,并将代码添加到工程的.cpp文件中。其次,就要配置工程的库文件和链接器。
    1. 设置ffmpeg头文件位置
    鼠标右键点击工程名,选择属性

    在visual <wbr>studio <wbr>2010中调用ffmpeg
    然后选择 配置属性 -> C/C++ -> 常规 -> 附加包含目录,添加目录为你下载的32-bit Builds (Dev)中的头文件目录。

    在visual <wbr>studio <wbr>2010中调用ffmpeg
     
    2. 设置ffmpeg的lib文件位置 
    鼠标右键点击工程名,选择属性, 然后选择 配置属性 -> 链接器 -> 常规 -> 附加库目录,添加目录为你下载的32-bit Builds (Dev)中的lib文件目录。 
    在visual <wbr>studio <wbr>2010中调用ffmpeg

    3. 设置ffmpeg的所引用的lib文件 
    鼠标右键点击工程名,选择属性, 然后选择 配置属性 -> 链接器 -> 输入 -> 附加依赖项,添加的文件为你下载的32-bit Builds (Dev)中的lib文件。 
    在visual <wbr>studio <wbr>2010中调用ffmpeg

    如果一切正常,这时你便可以编译成功。

    三、可能出现的问题
    1. 虽然编译通过,但是并不表示就可以运行,当你运行代码时会出现以下错误
    在visual <wbr>studio <wbr>2010中调用ffmpeg

    原因是,你虽然引用了LIB文件,但这并不是真正的静态库文件,而是对DLL的引用,所以当你调用ffmpeg库函数时,需要DLL文件在场。你可以用dumpbin(VS自带工具)来查看你生成的exe中引用了哪些DLL文件。你在命令行输入:
    >dumpbin D: est est.exe /imports
    然后根据显示的文件名,将下载的32-bit Builds (Shared)的bin文件夹下的dll文件拷贝到你新建的工程的源文件目录下。
    在visual <wbr>studio <wbr>2010中调用ffmpeg

    2. 上段代码中的变量argv[1],如果你没有指定参数,则会报错。所以如果不指定,则需要将其改为一个const char *的字符串变量,字符串内容为你想读入的视频的全路径,例如
    const char* filename = "D:\work\code\VideoConference\test\IMGP1816.AVI";
     
    上述都配置完毕就可以运行了,暂时先写这么多吧,后面我会继续总结的。
  • 相关阅读:
    MySQL7:性能优化
    MySQL6:视图
    MySQL5:触发器
    MySQL4:索引
    MySQL3:存储过程和函数
    MySQL1:MySQL函数汇总
    MySQL2:四种MySQL存储引擎
    Spring7:基于注解的Spring MVC(下篇)
    Spring6:基于注解的Spring MVC(上篇)
    Linux查看端口号
  • 原文地址:https://www.cnblogs.com/x_wukong/p/4840946.html
Copyright © 2020-2023  润新知