need be improved

胡斌
Commit 3fa8d5719c0a8a076c6e28ffd77b5dab9fc4791d 3fa8d571 1 parent e84f94bd
merge_av/merge_av.cpp
--- a/merge_av/merge_av.cpp
查看文件 @3fa8d57
+++ b/merge_av/merge_av.cpp
查看文件 @3fa8d57
@@ -184,14 +184,14 @@ void merge_video_silence(fileinfo video, const char * aacfile, const char * dest
 	run_shell_cmd(buf);
 }
-void megre_audio_video(fileinfo audio, int nf, fileinfo video, const char * destfile)
+void merge_audio_video(fileinfo audio, int nf, fileinfo video, const char * destfile)
 {
 	char buf[2048];
 	sprintf(buf, "ffmpeg -y -i %d_%s -i %s %s %s %s", nf, audio.name.c_str(), video.name.c_str(), acodec_param, vcodec_param, destfile);
 	run_shell_cmd(buf);
 }
-void megre_audio_video(const char * audio, const char * video, const char * destfile)
+void merge_audio_video(const char * audio, const char * video, const char * destfile)
 {
 	char buf[2048];
 	sprintf(buf, "ffmpeg -y -i %s -i %s %s %s %s", audio, video, acodec_param, vcodec_param, destfile);
@@ -422,6 +422,8 @@ int process_files(const char * output_dest_file)
 	check_audio_duration();
 	get_duration_from_video_file();
+	//don't split video, for a video, using merged audios to mix with it
+	//for audio, mix with video or jpg
 	if (filesvideo.size()) {//has video files
 		if (filesaudio.size()){
@@ -446,16 +448,136 @@ int process_files(const char * output_dest_file)
 				for (; nv < filesvideo.size(); nv++) {
 					fileinfo video = filesvideo[nv];
-					if (video.start_time - audio_start > 0.1) {//video is behand audio too much
+					if (video.start_time < audio_start - 0.1)
+					{//video is much more ahead of audio,try padding silence first
+						if (video.end_time < audio_start + 0.1) {
+							sprintf(destfile, "%d_silence.aac", nf);
+							split_audio(silence_aac_file, 0, video.end_time - video.start_time, destfile);
+							tmp_files.push_back(destfile);
+							sprintf(destfile, "%d.ts", nf);
+							merge_audio_video(destfile, video.name.c_str(), destfile);
+							merged_files.push_back(destfile);
+							nf++;
+							continue; //for next video
+						}
+						else {
+							// combine a audio file for the video
+							double silence_audio_start = audio.end_time;//maybe need append silence 
+							double silence_audio_end = video.end_time;
+
+							bool need_append_silence = true;
+							bool to_next_video = false;
+
+							vector<std::string > merge_audio_files;
+							sprintf(destfile, "%d_0_silence.aac", nf);//a duration of silence
+							split_audio(silence_aac_file, 0, audio_start - video.start_time, destfile);
+							merge_audio_files.push_back(destfile);
+							tmp_files.push_back(destfile);
+
+							if (audio.end_time < video.end_time + 0.1 && audio.end_time > video.end_time - 0.1) {
+								merge_audio_files.push_back(audio.name); //whole audio file,just fit
+								audio_start = audio.end_time + 0.1;
+								need_append_silence = false;
+							}
+							else if (audio.end_time > video.end_time){ //split part of audio file
+								sprintf(destfile, "%d_%s", nf, audio.name.c_str());
+								merge_audio_files.push_back(destfile);
+								tmp_files.push_back(destfile);
+								split_audio(audio.name.c_str(), audio_start, video.end_time - audio_start, destfile);
+								audio_start = video.end_time;
+								need_append_silence = false;
+							}
+							else {
+								merge_audio_files.push_back(audio.name);
+
+								for (; i + 1 < filesaudio.size(); i++){//since video is not finished,try find next audio
+									audio = filesaudio[i + 1];
+									if (audio.start_time < video.end_time) {//next audio should split to fit the video
+										silence_audio_end = audio.start_time;
+
+										sprintf(destfile, "%d_%d_silence.aac", nf, i);
+										split_audio(silence_aac_file, 0, silence_audio_end - silence_audio_start, destfile);
+										merge_audio_files.push_back(destfile);
+										tmp_files.push_back(destfile);
+
+										if (audio.end_time > video.end_time - 0.1 && audio.end_time < video.end_time + 0.1) {//just match
+											merge_audio_files.push_back(audio.name);
+											need_append_silence = false;
+											audio_start = audio.end_time + 0.1;
+											i++;//this audio is used
+											break;
+										}
+										if (audio.end_time > video.end_time){
+											sprintf(destfile, "%d_%s", nf, audio.name.c_str());
+											merge_audio_files.push_back(destfile);
+											tmp_files.push_back(destfile);
+											split_audio(audio.name.c_str(), 0, video.end_time - audio.start_time, destfile);
+											need_append_silence = false;
+											//adjust timecode for the audio is part left
+											float cur_audio_start = video.end_time - audio.start_time;
+
+											audio_start = audio.start_time;
+
+											for (int j = i + 1; j < filesaudio.size(); j++){
+												filesaudio[j].start_time -= audio_start;
+												filesaudio[j].end_time -= audio_start;
+											}
+
+											for (int j = nv; j < filesvideo.size(); j++) {
+												filesvideo[j].start_time -= audio_start;
+												filesvideo[j].end_time -= audio_start;
+											}
+											i++;
+											audio = filesaudio[i];
+											audio_start = cur_audio_start;
+
+											to_next_video = true;
+
+											break;
+										}
+										merge_audio_files.push_back(audio.name);//whole audio should be appended
+										silence_audio_start = audio.end_time; //adjust the silence start
+									}
+									else {
+										break;//no need for next audio
+									}
+								}//end audio find for the video
+							}//end else
+
+							if (need_append_silence) {
+								sprintf(destfile, "%d_silence.aac", nf);
+								split_audio(silence_aac_file, 0, silence_audio_end - silence_audio_start, destfile);
+								merge_audio_files.push_back(destfile);
+								tmp_files.push_back(destfile);
+							}
+
+							sprintf(audio_file, "%d_merged.aac", nf);
+							merge_audio_file(merge_audio_files, audio_file);
+
+
+							sprintf(destfile, "%d.ts", nf);
+							merge_audio_video(audio_file, video.name.c_str(), destfile);
+							merged_files.push_back(destfile);
+							nf++;
+
+							if (!to_next_video){
+								nv++;//this video is processed
+								break;
+							}
+						}//end need combine
+					}//end video is ahead of audio
+					//-----VS-----
+					//AS----------
+					if (video.start_time - audio_start > 0.1) {//video is behind audio too much
 						sprintf(audio_file, "%d_%s", nf, audio.name.c_str());
 						if (video.start_time < audio.end_time - 0.1){
 							split_audio(audio.name.c_str(), audio_start, video.start_time - audio_start, audio_file);
 							audio_start = video.start_time;
-							tmp_files.push_back(audio_file);
 						}
 						else {
-							strcpy(audio_file, audio.name.c_str());
+							split_audio(audio.name.c_str(), audio_start, audio.end_time - audio_start, audio_file);
 						}
+						tmp_files.push_back(audio_file);
 						sprintf(pic_file, "%s.jpg", video.name.c_str());
 						get_video_first_frame_jpeg(video, pic_file);
@@ -466,11 +588,14 @@ int process_files(const char * output_dest_file)
 						merged_files.push_back(destfile);
 						nf++;
-						if (video.start_time > audio.end_time){//all audio file no video, to next audio
+						if (video.start_time >= audio.end_time - 0.1){//all audio file no video, to next audio
 							audio_start = audio.end_time + 0.1;//no audio left
 							break;
 						}
 					}
+					//----AS--------
+					//----VS--------
+					else if (audio_start - video.start_time < 0.1){
 						if (audio.end_time > video.end_time){ //this video finish, to next video
 							sprintf(destfile, "%d_%s", nf, audio.name.c_str());
@@ -479,7 +604,7 @@ int process_files(const char * output_dest_file)
 							audio_start = video.end_time;
 							sprintf(destfile, "%d.ts", nf);
-						megre_audio_video(audio, nf, video, destfile);
+							merge_audio_video(audio, nf, video, destfile);
 							merged_files.push_back(destfile);
 							nf++;
 						}
@@ -489,7 +614,7 @@ int process_files(const char * output_dest_file)
 							tmp_files.push_back(destfile);
 							sprintf(destfile, "%d.ts", nf);
-						megre_audio_video(audio, nf, video, destfile);
+							merge_audio_video(audio, nf, video, destfile);
 							merged_files.push_back(destfile);
 							audio_start = audio.end_time + 0.1;//no audio left
 							nf++;
@@ -503,10 +628,12 @@ int process_files(const char * output_dest_file)
 							merge_audio_files.push_back(destfile);
 							tmp_files.push_back(destfile);
+							audio_start = audio.end_time + 0.1;
+
 							double silence_audio_start = audio.end_time;
 							double silence_audio_end = video.end_time;
-						bool need_slilence = true;
+							bool need_silence = true;
 							bool to_next_video = false;
 							for (; i + 1 < filesaudio.size(); i++){//since video is not finished,try find next audio
 								audio = filesaudio[i + 1];
@@ -520,7 +647,8 @@ int process_files(const char * output_dest_file)
 									if (audio.end_time > video.end_time - 0.1 && audio.end_time < video.end_time + 0.1) {//just match
 										merge_audio_files.push_back(audio.name);
-									need_slilence = false;
+										need_silence = false;
+										audio_start = audio.end_time + 0.1;
 										i++;
 										break;
 									}
@@ -529,7 +657,7 @@ int process_files(const char * output_dest_file)
 										merge_audio_files.push_back(destfile);
 										tmp_files.push_back(destfile);
 										split_audio(audio.name.c_str(), 0, video.end_time - audio.start_time, destfile);
-									need_slilence = false;
+										need_silence = false;
 										//adjust timecode for the audio is part left
 										float cur_audio_start = video.end_time - audio.start_time;
@@ -554,13 +682,14 @@ int process_files(const char * output_dest_file)
 									}
 									merge_audio_files.push_back(audio.name);//whole audio should be appended
 									silence_audio_start = audio.end_time; //adjust the silence start
+									audio_start = audio.end_time + 0.1;
 								}
 								else {
 									break;//no need for next audio
 								}
 							}
-						if (need_slilence) {
+							if (need_silence) {
 								sprintf(destfile, "%d_silence.aac", nf);
 								split_audio(silence_aac_file, 0, silence_audio_end - silence_audio_start, destfile);
 								merge_audio_files.push_back(destfile);
@@ -572,7 +701,7 @@ int process_files(const char * output_dest_file)
 							sprintf(destfile, "%d.ts", nf);
-						megre_audio_video(audio_file, video.name.c_str(), destfile);
+							merge_audio_video(audio_file, video.name.c_str(), destfile);
 							merged_files.push_back(destfile);
 							nf++;
@@ -582,16 +711,24 @@ int process_files(const char * output_dest_file)
 							}
 						}
 					}
+				}
 				if (audio_start < audio.end_time){
-
 					sprintf(destfile, "%d_%s", nf, audio.name.c_str());
-
 					split_audio(audio.name.c_str(), audio_start, audio.end_time - audio_start, destfile);
 					tmp_files.push_back(destfile);
 					sprintf(destfile, "%d.ts", nf);
-					merge_audio_pic(audio, nf, blank_pic_file, destfile);
+					if (nv < filesvideo.size()) {
+						fileinfo video = filesvideo[nv];
+						sprintf(pic_file, "%s.jpg", video.name.c_str());
+						get_video_first_frame_jpeg(video, pic_file);
+						tmp_files.push_back(pic_file);
+					}
+					else {
+						strcpy(pic_file, blank_pic_file);
+					}
+					merge_audio_pic(audio, nf, pic_file, destfile);
 					merged_files.push_back(destfile);
 					nf++;
 				}