-
Notifications
You must be signed in to change notification settings - Fork 9
/
ffmpeg.sh
113 lines (86 loc) · 6.97 KB
/
ffmpeg.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
ffmpeg -i long_sample.mp4 -vf "transpose=1" -c:a copy out.mov
# 逆时针旋转270度
ffmpeg -i long_sample.mp4 -map_metadata 0 -metadata:s:v rotate="90" -codec copy output.mp4
ffmpeg -i long_sample.mp4 -map_metadata 0 -metadata:s:v rotate="270" -codec copy output270.mp4
# 从 mov 转成 mp4
ffmpeg -i 2018Demo-HD-1080p.mov -vcodec copy -acodec copy 2018Demo-HD-1080p.mp4
# 从 mov 转成 mp4，并逆时针旋转270度
ffmpeg -i 2018Demo-720p.mov -map_metadata 0 -metadata:s:v rotate="270" -codec copy 2018Demo-720p-rotate270.mp4
# 将视频重新编码
ffmpeg -i ./qshell/home.mov -s hd720 -b:v 384k -vcodec libx264 -r 23.976 -c:a aac -ac 2 -ar 44100 -ab 64k -profile:v baseline -pix_fmt yuv420p -crf 22 -deinterlace 2.mp4
# 无声版
ffmpeg -i ./qshell/home.mov -s hd720 -b:v 256k -vcodec libx264 -r 24 -an -profile:v baseline -pix_fmt yuv420p -crf 24 -deinterlace 2.no.audio.mp4
ffmpeg -i ./qshell/home.mov -s hd720 -b:v 256k -c:v libx264 -preset veryslow -profile:v baseline -pix_fmt yuv420p -crf 30 -deinterlace -an $(date +'%H%M%S').mp4
# -b:v 384k 视频比特率为 384k
# -crf Constant Rate Factor 取值范围0~51，其中0为无损模式，数值越大，画质越差，生成的文件却越小。从主观上讲，18~28是一个合理的范围，18被认为是视觉无损的
ffmpeg -i MainVideo.avi -f gdigrab -framerate 25 -video_size 300x200 -i title="MyWindow" -filter_complex "[0]setpts=PTS-STARTPTS[b];[b][1:v]overlay=(main_w-overlay_w):main_h-overlay_h[v]" -map "[v]" -c:v libx264 -r 25 out.mp4
ffmpeg -i 1.mp4 -f gdigrab -framerate 25 -video_size 300x200 -i title="MyWindow"
-filter_complex
"[1]split[m][a];
[a]format=yuv444p,geq='if(gt(lum(X,Y),0),255,0)',hue=s=0[al];
[m][al]alphamerge[ovr];
[0][ovr]overlay=(main_w-overlay_w):main_h-overlay_h[v]"
-map "[v]" -c:v libx264 -r 25 out.mp4
# 原始命令 覆盖一张图片，只覆盖0-20秒之间的视频
ffmpeg -i input.mp4 -i image.png \
-filter_complex "[0:v][1:v] overlay=25:25:enable='between(t,0,20)'" \
-pix_fmt yuv420p -c:a copy \
output.mp4
# 切出小段视频 -t指定秒数
ffmpeg -t 12 -i sxc_146952623220.mp4 12s$(date +%m%d%H%M).mp4
# 处理未处理妥当的视频，只覆盖右上角
ffmpeg -hide_banner -loglevel panic -y -i 5s05051105.mp4 -i right_top.png -filter_complex "[1]scale=335:255[s1];[0][s1]overlay=946:-1[v2]" -map "[v2]" -map 0:a -pix_fmt yuv420p -movflags +faststart output$(date +%m%d%H%M).mp4
# When encoding is finished -movflags +faststart will relocate some data to the beginning of the file. This is useful, for example, if you are outputting to MP4 and your viewers will watch via progressive download such as from a browser.
# https://superuser.com/questions/848605/ffmpeg-video-from-single-image-with-multiple-video-overlays
# 在指定位置上覆盖一张图片
# crop然后scale，需要传递参数setsar=1
ffmpeg -hide_banner -y -i input/070.mp4 -i left_top.png -i right_top.png -i right_bottom.png -i logo_middle.png -filter_complex '[0:v]crop=1424:720:0:40,scale=1280:720,setsar=1[bg];[bg][1]overlay=0:-4[v1];[v1][3]overlay=main_w-overlay_w+2:main_h-overlay_h+2[v3];[v3][4]overlay=(main_w-overlay_w)/2:(main_h-overlay_h)/2[v4]' -map '[v4]' -map 0:a -pix_fmt yuv420p -movflags +faststart processing/070_$(date +%m%d%H%M).mp4
# crop filter 截取视频
# https://video.stackexchange.com/questions/4563/how-can-i-crop-a-video-with-ffmpeg
# 图片缩放
ffmpeg -i logo_middle_95.png -filter_complex "[0]scale=382:300[s0]" -map "[s0]" logo_middle_382x300.png
# 截图
ffmpeg -ss 00:00:4 -i sxc_150945796413_1352.mp4 -vframes 1 -filter:v "scale=1280:720" -q:v 2 -pix_fmt yuvj420p output$(date +%m%d%H%M).jpg
ffmpeg -ss 10 -i sxc_147506207636.mp4 -vframes 1 -filter:v "scale=1280:720" -q:v 2 -pix_fmt yuvj420p output$(date +%m%d%H%M).jpg
ffmpeg -ss 10 -i sxc_150167870155.mp4 -vframes 1 -filter_complex "[0]crop=1424:720:0:40,scale=1280:720,setsar=1" -q:v 2 -pix_fmt yuvj420p 155.jpg
ffmpeg -ss 10 -i sxc_146829584623.mp4 -vframes 1 -filter:v "scale=1280:720" -q:v 2 -pix_fmt yuvj420p output$(date +%m%d%H%M).jpg
ffmpeg -ss 10 -i sxc_147406889950.mp4 -vframes 1 -filter:v "scale=1280:720" -q:v 2 -pix_fmt yuvj420p output$(date +%m%d%H%M).jpg
ffmpeg -ss 10 -i sxc_148236441731.mp4 -vframes 1 -filter:v "scale=1280:720" -q:v 2 -pix_fmt yuvj420p output$(date +%m%d%H%M).jpg
# 从视频中截取出一张图
# https://stackoverflow.com/questions/27568254/how-to-extract-1-screenshot-for-a-video-with-ffmpeg-at-a-given-time
# 定位覆盖图片似乎可以用这种方法，看起来非常不错
ffmpeg -i input.mp4 -i logo.png -filter_complex "[0:v]scale=512:-1[bg];[bg][1:v]overlay=(main_w-overlay_w)/2:(main_h-overlay_h)/2" output
# https://stackoverflow.com/questions/10934420/ffmpeg-how-to-scale-a-video-then-apply-a-watermark
# -map 的使用：https://trac.ffmpeg.org/wiki/Map
# 颜色 383838(262626) 变成353535(232323)
ffmpeg -loop 1 -t 24 -i "image.jpg" -filter_complex "color=000000:s=640x360[bg];[bg][0]overlay=shortest=1:y='min(0,-(t)*26)'" -qscale 1 -y out.mpg
ffmpeg -i moonmen.mp4 -i transparent_overlay.mp4 -filter_complex
"[1]format=rgba,geq=r='r(X,Y)':a='if(between(r(X,Y),77,87)*between(g(X,Y),39,49)*between(b(X,Y),6,16),0,255)'[ovr];
[0][ovr]overlay"
output.mp4
ffmpeg -i 1.mp4 -i 2.mp4 -filter_complex
"[1]split[m][a];
[a]geq='if(gt(lum(X,Y),16),255,0)',hue=s=0[al];
[m][al]alphamerge[ovr];
[0][ovr]overlay"
output.mp4
ffmpeg -i input.avi
-vf
"yadif,format=rgb24,
lutrgb=r='if(eq(val,0),3,val)':g='if(eq(val,0),3,val)':b='if(eq(val,0),3,val)'"
image%d.png
ffmpeg -i input.mp4 -vf "yadif,format=rgb24,lutrgb=r='if(eq(val,38),35,val)':g='if(eq(val,38),35,val)':b='if(eq(val,38),35,val)'" -pix_fmt yuv420p -c:a copy output$(date +%H%M%S).mp4
ffmpeg -i input.mp4 -vf "format=rgb24,lutrgb=r='if(eq(val,38),255,val)':g='if(eq(val,38),255,val)':b='if(eq(val,38),255,val)'" -pix_fmt yuv420p -c:a copy output$(date +%H%M%S).mp4
ffmpeg -i input.mp4 -vf "format=rgb24,lutrgb=r='if(lt(val,39),255,val)':g='if(lt(val,39),255,val)':b='if(lt(val,39),255,val)'" -pix_fmt yuv420p -c:a copy output$(date +%H%M%S).mp4
# 探测尺寸
ffprobe -hide_banner -v quiet -show_streams input.mp4 | grep -E '^(width|height)'
# 利用shell获取尺寸
size=$(ffprobe -hide_banner -v quiet -show_streams sxc_151444152211.mp4 | grep -E '^(width|height)')
width=$(echo $size | awk '{print $1}' | cut -d= -f2)
height=$(echo $size | awk '{print $2}' | cut -d= -f2)
# 检测视频的实际位置，截除黑边
# cropdetect=24:16:0 第一个参数24表示所有颜色值小于24的，认为是黑色，用来定位黑边；第二个参数16，表示宽和高都要是16的倍数，如果不希望视频被裁剪，可以用2；第三个参数传0就行了，具体还不太理解
ffmpeg -hide_banner -y -i sxc_150203400070_black.mp4 -t 1 -vf cropdetect=24:16:0 -f null - 2>&1 | awk '/crop/ { print $NF }' | tail -1
ffmpeg -hide_banner -y -i filter1368.mp4 -t 1 -vf cropdetect=24:2:0 -f null - 2>&1 | awk '/crop/ { print $NF }' | tail -1
# cropdetect https://stackoverflow.com/questions/17265381/ffmpeg-get-value-from-cropdetect
# http://www.ffmpeg.org/ffmpeg-filters.html#cropdetect