ffmpeg cheatsheet for glitching

ffmpeg cheatsheet for glitching

FFMPEG '-i', file, '-c:v', 'libopenjpeg', "jpeg2000\\" + name + ".jp2"\

Convert image to jpeg2000
	ffmpeg -i in.png -c:v libopenjpeg out.jp2
	Hex Edit out.jp2
	ffmpeg -i in.jp2 -c:v png out.png

General Edit
ffmpeg -i input.avi -c:v mpeg2video -g 999 -q:v 1 output.avi
*edit in avidemux/whatever*
ffmpeg -i input.avi -c:v h264 -preset medium -crf 24 final.mp4

Remove bFrames
	-x265-params b-frames=0"

GET FORMATS
ffmpeg.exe -codecs | grep DEV

Get list of Containers
 ffmpeg -formats | grep DE

Get Encoder options
	ffmpeg -h encoder=X
	X from Get Formats

sharpen abuse (loop dis)
	ffmpeg -i IN.jpg -q:v 31 -vf unsharp=luma_amount=0.2 input.jpg

Displacement Maps
 ( Remove crf and b:v parts to get normal quality)
	ffmpeg -i IN.mp4 -i IN.mp4 -i IN.mp4 -c:v hevc -crf 40 -b:v 20k -filter_complex "[0][1][2]displace=edge=wrap[middle];[middle]stereo3d=ar" -y res.mp4

Get Frames 
 	ffmpeg -i in.avi -q:v 0 imagename_%04d.jpg

convert image to video
	(single image into 10 sec video)
	ffmpeg -loop 1 -y -i sourceimage.jpg -t 10 -c:v codec-you-wanna-use video.avi

  (from Frames r == fps)
	ffmpeg -r 1/5 -i img%03d.png -c:v libx264 -vf fps=25 -pix_fmt yuv420p out.mp4

Video to RAW for sonifaction
	ffmpeg -ss 00:01:02 -t 5 -i IN.mp4 -f rawvideo -pix_fmt yuv420p result.yuv
	ss = start time
	t == seconds after ss to use

RAW to Video (After Sonfication)
	ffmpeg -f rawvideo -video_size 1280x720 -pix_fmt yuv420p -i result.yuv output.mp4


BELOW CREDIT oioiiooixiii:

STACK EFFECTS!!! can use ffmpeg but results will be different.
	ffplay -i IN.mp4 -vf \
   "
      scale=-2:720,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      spp=4:10,
      tblend=all_mode=average,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      spp=4:10,
      tblend=all_mode=average,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      spp=4:10,
      tblend=all_mode=average,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128,
      tblend=all_mode=difference128 
   "


# Isolate motion-vectors using 'difference128' blend filter
# - add brightness, contrast, and scaling, to taste
	ffplay \
	   -flags2 +export_mvs \
	   -i "video.mp4" \
	   -vf \
	   "
	      split[original],
	      codecview=mv=pf+bf+bb[vectors],
	      [vectors][original]blend=all_mode=difference128,
	      eq=contrast=7:brightness=-0.3,
	      scale=720:-2
	   "

# chromatic aberration in ffmpeg
# ffplay used for testing; filtergraph can be directly inserted into ffmpeg

ffplay\
   -i video.mkv\
   -vf\
      "split=3[r][g][b];\
      nullsrc=size=640x360[base1];\
      nullsrc=size=640x360[base2];\
      nullsrc=size=640x360[base3];\
      [r]lutrgb=g=0:b=0[red];\
      [g]lutrgb=r=0:b=0[green];\
      [b]lutrgb=r=0:g=0[blue];\
      [base1][red]overlay=x=10:shortest=1,format=rgb24[x];\
      [base2][green]overlay=x=0:shortest=1,format=rgb24[y];\
      [base3][blue]overlay=y=10:shortest=1,format=rgb24[z];\
      [x][y]blend=all_mode='addition'[xy];\
      [xy][z]blend=all_mode='addition'[xyz];\
      [xyz]crop=630:350:10:10,scale=640:360:out_color_matrix=bt709"

# Averaged Key-Frames
# To extract keyframes from a video using FFmpeg
ffmpeg -i video.mp4 -vf "select=eq(pict_type\,I)" -vsync vfr frame-%08d.png

# To merge images together with ImageMagick
convert *.png -evaluate-sequence mean output.png

#Pixel Array
# A trimmed down version, using only one instance of FFmpeg.
# The disadvantage with this is the lack of progress display.
ffmpeg -i input.vid -frames 1 -vf "scale=1:1,tile=640x220" output.png

# OR v1
# First instance of FFmpeg traverses the frames, the second concatenates them.
ffmpeg -y -i video.mkv -vf "scale=1:1" -c:v png -f image2pipe pipe:1 |\ 
ffmpeg -y -i pipe:0 -vf "tile=640x242" output.png 2> /dev/null

## ECHO!!!! by Rob Mac
## so to deconstruct, we have the normal input, then use seeking to offset each input of the same file each time: 
-i k1.mkv -ss 00:00:00.25 -i k1.mkv, and so on, i'm using offsets of 250ms (0.25 seconds) but you can use whatever
works for you, ffmpeg will try its best to match it. then we use [0][1]blend[a] and [1][2]blend[b] to make the two
echoes, nothing too unusual there. after that each needs a stage to blend back together, so we go [a][b]blend[p] to
mix down our echoes, then [0][p]blend[o] to mix back into main signal, and [o] to output. if we did more, say 4 instead 
of 2, we would do [0][1]blend[a], [1][2]blend[b], [2][3]blend[c], 
then [b][c]blend[mix1],[a][b]blend[mix2],[mix1][mix2]blend[echo],[0][echo]blend[o] instead. i probably wouldn't want to 
use this for more than a few echoes to be honest just because making the filter toplogy is a bit of a pain, but it encodes 
pretty quick and does some fun stuff in different blend modes..
ffmpeg -i k1.mkv -ss 00:00:00.25 -i k1.mkv -ss 00:00:00.50 -i k1.mkv -filter_complex "[0][1]blend=all_mode=lighten[a];[1][2]blend=all_mode=lighten[b];[a][b]blend=all_mode=lighten[p];[0][p]blend=all_mode=lighten[o]" -map "[o]" -c:v h264 -crf 15 kd.mp4