本文主要是介绍毕设系列之Libx264实时视频流(YUV 420P转H264视频编码篇),希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!
#PS:要转载请注明出处,本人版权所有
#PS:这个只是 《 我自己 》理解,如果和你的
#原则相冲突,请谅解,勿喷
开发环境:Ubuntu 16.04 LTS
本文的技术实现部分参考雷博士的这篇文章。http://blog.csdn.net/leixiaohua1020/article/details/42078645
1、现在网上关于H264的文章有很多,但是我个人认为最好的就是雷霄骅博士的x264部分的文章最详细。所以许多的细节部分,我推荐大家去雷博士的blog去看。本文只提及我们使用Libx264时候,我们要注意的问题。
2、 使用Libx264时候,我们需要关注的东西(下面用我的代码来说明假如我们要使用Libx264,那么我们需要注意的几个事情)。
//encoderx264_t * pX264Handle;//结构体是一个编码器实例句柄,要使用这个编码库,我们必须有一个这种变量,没有为啥。
//paramx264_param_t * pX264Param;//这个结构体就比较重要了,他是我们设置编码器参数的载体,我们必须具体的了解各种参数的意义。具体参数在下一节进行分析。
//input,output picx264_picture_t *pPic_In;//这就是YUV输入图像和输出图像的载体,这里面有一个pts参数需要注意,下面小节进行说明。x264_picture_t *pPic_Out;
//output h264 streamx264_nal_t * pNals;//这个也是比较重要的一个东西,他的作用是用来保存编码后,网络抽象层所保存的数据(NAL HEADER,NAL BODY),想具体了解,可以去看H264编码原理。
//user config callback//UESER_CONF_CALLBACK yX264_UserConfig;int (*yX264_UserConfig)(struct ymx264 * mvl);//私有,忽略
//pi_nal is the number of NAL units int pi_nal;//网络抽象单元个数
3、 编码器参数分析
//* cpuFlags mvl->pX264Param->i_threads = X264_SYNC_LOOKAHEAD_AUTO;//* 取空缓冲区继续使用不死锁的保证. //* 视频选项 mvl->pX264Param->i_width = FRAME_WIDTH; //* 要编码的图像宽度. mvl->pX264Param->i_height = FRAME_HEIGHT; //* 要编码的图像高度 mvl->pX264Param->i_frame_total = 0; //* 编码总帧数.不知道用0. /* Force an IDR keyframe at this interval */mvl->pX264Param->i_keyint_max = 10; //这个参数很重要,控制i帧的频率mvl->pX264Param->b_repeat_headers = 1; // 重复SPS/PPS 放到关键帧前面//做实时流播放,此参数必须ENABLE//* 流参数 //* how many b-frame between 2 references pictures */mvl->pX264Param->i_bframe = 5; //mvl->pX264Param->b_open_gop = 0; //* Keep some B-frames as references: 0=off, 1=strict hierarchical, 2=normal */mvl->pX264Param->i_bframe_pyramid = 0; //mvl->pX264Param->i_bframe_adaptive = X264_B_ADAPT_TRELLIS; //* Log参数,不需要打印编码信息时直接注释掉就行 //mvl->pX264Param->i_log_level = X264_LOG_DEBUG; //* 速率控制参数 //pX264Param->rc.i_bitrate = 1024 * 10;//* 码率(比特率,单位Kbps) ,重要//* muxing parameters 帧率控制,重要。mvl->pX264Param->i_fps_den = 1; //* 帧率分母mvl->pX264Param->i_fps_num = Y_STREAM_FPS;//* 帧率分子 mvl->pX264Param->i_timebase_den = mvl->pX264Param->i_fps_num; mvl->pX264Param->i_timebase_num = mvl->pX264Param->i_fps_den;
最后,我们需要注意一点:关于我们设置的帧率的问题,不一定是设置多少,播放的时候就是多少,只是一个参考值,编码器会尽量的把视频编码为这个帧率。
4、x264_picture_t * pPic_In->i_pts += 1; 此参数非常重要。如果不进行设置,视频流将不会正常播放。
/*
PTS:Presentation Time Stamp。PTS主要用于度量解码后的视频帧什么时候被显示出来
DTS:Decode Time Stamp。DTS主要是标识读入内存中的bit流在什么时候开始送入解码器中进行解码。
*/
5、关于颜色空间的问题,大家可以去百度YUV 420 ,YUV 422,YUV 444等这些原始图像的存储问题。具体来说,他们分为两类,一种是分组存储(例如:YYY*UUU*VVV*),一种是交叉存储(例如:YUYV)
6、此模块我的源代码
ym_x264.h
/*FileName:ym_x264.hVersion:1.0Description:Created On: 2017-3-19Modified date:Author:Sky
*/#ifndef _YM_X264_H
#define _YM_X264_H#ifdef __cplusplus
#if __cplusplus
extern "C"{
#endif
#endif /* __cplusplus */#include <stdint.h>
#include <stdio.h>#include <ym_x264_config.h>#include <x264.h>#include <stdlib.h>#define CLEAR_MEM(x) memset(&(x),0,sizeof(x))enum yX264Cmd{ DO_DEFAULT_PRESET = 0, DO_DEFAULT_USERCONF = 1, DO_PARAM_APPLY_PROFILE = 2, OPEN_ENCODER = 3, ENCODER_ENCODE = 4,};
enum yX264ColorSpace{Y_CSP_I444 = 0,Y_CSP_I422 = 1,Y_CSP_I420 = 2, Y_CSP_YUYV = 3,
};//typedef struct ymx264 yMX264;typedef struct ymx264{//encoderx264_t * pX264Handle;
//paramx264_param_t * pX264Param;
//input,output picx264_picture_t *pPic_In;x264_picture_t *pPic_Out;
//output h264 streamx264_nal_t * pNals;
//user config callback//UESER_CONF_CALLBACK yX264_UserConfig;int (*yX264_UserConfig)(struct ymx264 * mvl);
//pi_nal is the number of NAL units int pi_nal;long cur_pts;}yMX264; typedef int (*UESER_CONF_CALLBACK)(yMX264 * mvl);int yInitMX264(yMX264 * mvl);
int yDestroyMX264(yMX264 * mvl);
int yIoctlX264(enum yX264Cmd cmd,...);
//CSC = ColorSpaceCovert,FIP = Fill In_Pic
int yDoCSC_And_FIP(yMX264 * mvl,enum yX264ColorSpace t_csp, int cache_id);int yDo_Default_UserConf(yMX264 * mvl);#ifdef __cplusplus
#if __cplusplus
}
#endif
#endif /* __cplusplus */ #endif
ym_x264.c
/*FileName:ym_x264.cVersion:1.0Description:Created On: 2017-3-19Modified date:Author:Sky
*//*
x264_param_default():设置参数集结构体x264_param_t的缺省值。
x264_picture_alloc():为图像结构体x264_picture_t分配内存。
x264_encoder_open():打开编码器。
x264_encoder_encode():编码一帧图像。
x264_encoder_close():关闭编码器。
x264_picture_clean():释放x264_picture_alloc()申请的资源。存储数据的结构体如下所示。
x264_picture_t:存储压缩编码前的像素数据。
x264_nal_t:存储压缩编码后的码流数据。
*/
#include <ym_x264.h>uint8_t ImgCache[ImageCacheNum][FRAME_SIZE];
int yInitMX264(yMX264 * mvl){mvl->pX264Param = (x264_param_t *)malloc(sizeof(x264_param_t));// for (int i = 0; i < ImageCacheNum; i++){mvl->pPic_In = (x264_picture_t *)malloc(sizeof(x264_picture_t));mvl->pPic_Out = (x264_picture_t *)malloc(sizeof(x264_picture_t));mvl->pNals = NULL;mvl->pi_nal = 0;x264_picture_init(mvl->pPic_Out); x264_picture_alloc(mvl->pPic_In, FRAME_COLORSPACE, FRAME_WIDTH, FRAME_HEIGHT);// PTS FROM 0,AND AUTO INCRESE 1mvl->pPic_In->i_pts = 0;
// }mvl->cur_pts = 0;return 0;
}
int yDestroyMX264(yMX264 * mvl){// 清除图像区域 //for (int i = 0; i < ImageCacheNum; i++)x264_picture_clean(mvl->pPic_In); x264_encoder_close(mvl->pX264Handle); free(mvl->pX264Param);//for (int i = 0; i < ImageCacheNum; i++){free(mvl->pPic_In);free(mvl->pPic_Out);//}return 0;
}int yIoctlX264(enum yX264Cmd cmd,...){va_list arg;va_start(arg,cmd);yMX264 *mx264;mx264 = va_arg(arg,yMX264 *);va_end(arg);switch(cmd){case DO_DEFAULT_PRESET:{/* Get default params for preset/tuning *///x264_param_default(pParam); //this do default set for x264,but can not config some infoif( x264_param_default_preset( mx264->pX264Param, "veryfast", "zerolatency" ) < 0 ){printf("x264_param_default_preset failed!\n");return -1;}break;}case DO_DEFAULT_USERCONF:{//va_list arg;//va_start(arg,cmd);//mx264->yX264_UserConfig = va_arg(arg,UESER_CONF_CALLBACK);//mx264->yX264_UserConfig = NULL;//va_end(arg);if ( (*mx264->yX264_UserConfig)(mx264) < 0){printf("Do user conf callback failed.\n");return -1;}break;}case DO_PARAM_APPLY_PROFILE:{//x264_profile_names[0] = baseline , to set stream-qualityif (x264_param_apply_profile(mx264->pX264Param, x264_profile_names[0]) < 0 ){printf("x264_param_apply_profile failed.\n");return -1;} break;}case OPEN_ENCODER:{//open encoderif( (mx264->pX264Handle = x264_encoder_open(mx264->pX264Param)) == NULL){printf("x264_encoder_open failed.\n");return -1; }break;}case ENCODER_ENCODE:{
/*x264_encoder_encode:* encode one picture.* *pi_nal is the number of NAL units outputted in pp_nal.* returns the number of bytes in the returned NALs.* returns negative on error and zero if no NAL units returned.* the payloads of all output NALs are guaranteed to be sequential in memory. int x264_encoder_encode( x264_t *, x264_nal_t **pp_nal, int *pi_nal, x264_picture_t *pic_in, x264_picture_t *pic_out );
*/if ( x264_encoder_encode(mx264->pX264Handle, &mx264->pNals, &mx264->pi_nal, mx264->pPic_In, mx264->pPic_Out) < 0){printf("x264_encoder_encode failed.\n");return -1; }
/*
PTS:Presentation Time Stamp。PTS主要用于度量解码后的视频帧什么时候被显示出来
DTS:Decode Time Stamp。DTS主要是标识读入内存中的bit流在什么时候开始送入解码器中进行解码。
*///Presentation Time Stamp。PTS主要用于度量解码后的视频帧什么时候被显示出来//MUST DO THIS ,IT DECIDE ,主要用于度量解码后的视频帧什么时候被显示出来//mx264->pPic_In->i_pts += 1;mx264->pPic_In->i_pts += 1; printf("pts = %d\n",mx264->pPic_In->i_pts);break; }default :{printf("No this cmd to analyse\n");return -1;break;}}return 0;
}int yDoCSC_And_FIP(yMX264 * mvl,enum yX264ColorSpace t_csp, int cache_id){switch(t_csp){ case Y_CSP_I444:{
/*read(fd_in,pPic_In->img.plane[0],FRAME_SIZE); //Y read(fd_in,pPic_In->img.plane[1],FRAME_SIZE); //U read(fd_in,pPic_In->img.plane[2],FRAME_SIZE); //V
*/break;} case Y_CSP_I420:{
/*#ifndef ENABLE_YUYVTOI420 read(fd_in,pPic_In->img.plane[0],FRAME_SIZE); //Y read(fd_in,pPic_In->img.plane[1],FRAME_SIZE/4); //U read(fd_in,pPic_In->img.plane[2],FRAME_SIZE/4); //V #else//YUYV to I420read(fd_in,Cache,FRAME_SIZE*2); //read one frame to cache //must set to 0int id_u = 0, id_v = 0 , id_y = 0;for (int i = 0; i < FRAME_SIZE*2 ;i+=4){ pPic_In->img.plane[0][id_y] = Cache[i];//get Yid_y++;pPic_In->img.plane[0][id_y] = Cache[i+2];//get Yid_y++;if ( ((int)((i)/1280)%2) == 0 ){pPic_In->img.plane[1][id_u] = Cache[i+1];//get UpPic_In->img.plane[2][id_v] = Cache[i+3];//get Vid_u++;id_v++;}}#endif
*/break;} case Y_CSP_YUYV:{// firstly,Do YUYV to I420 ,then, Fill in_picint id_u = 0, id_v = 0 , id_y = 0;for (int i = 0; i < FRAME_SIZE ;i+=4){ mvl->pPic_In->img.plane[0][id_y] = ImgCache[cache_id][i];//get Yid_y++;mvl->pPic_In->img.plane[0][id_y] = ImgCache[cache_id][i+2];//get Yid_y++;if ( ((int)((i)/1280)%2) == 0 ){mvl->pPic_In->img.plane[1][id_u] = ImgCache[cache_id][i+1];//get Umvl->pPic_In->img.plane[2][id_v] = ImgCache[cache_id][i+3];//get Vid_u++;id_v++;}}break;}case Y_CSP_I422:{
/*read(fd_in,pPic_In->img.plane[0],FRAME_SIZE); //Y read(fd_in,pPic_In->img.plane[1],FRAME_SIZE/2); //U read(fd_in,pPic_In->img.plane[2],FRAME_SIZE/2); //V
*/break;} default:{ printf("Colorspace Not Support.\n"); return -1;} }}int yDo_Default_UserConf(yMX264 * mvl){//* cpuFlags mvl->pX264Param->i_threads = X264_SYNC_LOOKAHEAD_AUTO;//* 取空缓冲区继续使用不死锁的保证. //* 视频选项 mvl->pX264Param->i_width = FRAME_WIDTH; //* 要编码的图像宽度. mvl->pX264Param->i_height = FRAME_HEIGHT; //* 要编码的图像高度 mvl->pX264Param->i_frame_total = 0; //* 编码总帧数.不知道用0. /* Force an IDR keyframe at this interval */mvl->pX264Param->i_keyint_max = 10; mvl->pX264Param->b_repeat_headers = 1; // 重复SPS/PPS 放到关键帧前面//* 流参数 //* how many b-frame between 2 references pictures */mvl->pX264Param->i_bframe = 5; //mvl->pX264Param->b_open_gop = 0; //* Keep some B-frames as references: 0=off, 1=strict hierarchical, 2=normal */mvl->pX264Param->i_bframe_pyramid = 0; //mvl->pX264Param->i_bframe_adaptive = X264_B_ADAPT_TRELLIS; //* Log参数,不需要打印编码信息时直接注释掉就行 //mvl->pX264Param->i_log_level = X264_LOG_DEBUG; //* 速率控制参数 //pX264Param->rc.i_bitrate = 1024 * 10;//* 码率(比特率,单位Kbps) //* muxing parameters mvl->pX264Param->i_fps_den = 1; //* 帧率分母mvl->pX264Param->i_fps_num = Y_STREAM_FPS;//* 帧率分子 mvl->pX264Param->i_timebase_den = mvl->pX264Param->i_fps_num; mvl->pX264Param->i_timebase_num = mvl->pX264Param->i_fps_den; return 0;
}
#PS:请尊重原创,不喜勿喷
#PS:要转载请注明出处,本人版权所有.
有问题请留言,看到后我会第一时间回复
这篇关于毕设系列之Libx264实时视频流(YUV 420P转H264视频编码篇)的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!