Screencast script based on FFMPEG

Question

I am trying to write my own screencast script based on FFmpeg. I want to be able to draw a rectangle on the screen and force FFmpeg to make a screencast of this area only, for that I would need the coordinates of x,y of the starting point, and the resolution of the selected area. I can draw a rectangular area with scrot -s and I can also get the coordinates with xdotool getmouselocation but I don't know how I can save the location of the scrot start location.

I am imagining my script will be something like this:

1. scrot -s - draw a rectangular area on the screen
2. save start_x, start_y locations in a variable
3. ffmpeg -i scrot_screenshot - to get resolution of the selected area
4. ffmpeg -video_size $area_resolution -f x11grab -i :0.0+start_x,start_y screen_capture.mp4 - to create the screencast

So I need a solution for 2 only, I can do the rest of the script on my own.

I can, of course, use a different than scrot command for drawing the rectangular area of the screen, but so far, I didn't find a suitable replacement.

The final goal is to create such a script and I will for sure publish it on my Github account with the respective documentation on how to use it.

Answer 1

Try this;

#!/bin/bash

echo click T.L.
echo -ne "\e[?1000h"
while read -srn 6; do
    eval "$(xdotool getmouselocation --shell)"
    TL=$X,$Y
    echo -ne "\e[?1000l"
    break
done
sleep 0.5
echo click B.R.
echo -ne "\e[?1000h"
while read -srn 6; do
    eval "$(xdotool getmouselocation --shell)"
    BR=$X,$Y
    echo -ne "\e[?1000l"
    break
done
echo "$TL.$BR"

with a transparent term over the area you want to record.

ref; https://stackoverflow.com/a/5970472/1133275