vidstabdetect

Watch on youtube.com
doc

https://ffmpeg.org/ffmpeg-filters.html#vidstabdetect, https://ffmpeg.org/ffmpeg-filters.html#vidstabtransform

The original purpose of “vidstabdetect” is analysis of “stabilization/deshaking”, not visualization of the video data. Originally it is possible to do “stabilization/deshaking” by passing the analysis result of “vidstabdetect” to “vidstabtransform”. However, I think that using “vidstabdetect=show” as one of “visual effects” is also interesting.

show=1 (left side of uploaded video)

#! /bin/sh
ifn="Drifting with Cars.mp4"
ifnb="`basename \"${ifn}\" .mp4`"
pref="`basename $0 .sh`"

#
ffmpeg -y \
    -i "${ifn}" \
    -vf "vidstabdetect=show=1" \
    -an "${pref}_${ifnb}.mp4"

show=2 (right side of uploaded video)

#! /bin/sh
ifn="Drifting with Cars.mp4"
ifnb="`basename \"${ifn}\" .mp4`"
pref="`basename $0 .sh`"

#
ffmpeg -y \
    -i "${ifn}" \
    -vf "vidstabdetect=show=2" \
    -an "${pref}_${ifnb}.mp4"