Screen capture, recording, casting B - 2020

Continuing the screen recording from the previous chapter (, this time, we want to record only an active screen not the whole screen of a desktop running Ubuntu 13.1.

We'll use the FFmpeg's x11grab, a module for screen capture. This method known for giving the best results for capturing screen and is one of the the most flexible methods. It allows us to use a variety of inputs and output formats.

The following command will give us information about the currently active window:

$ xwininfo -id $(xprop -root | awk '/_NET_ACTIVE_WINDOW\(WINDOW\)/{print $NF}')

xwininfo can be used interactively. If we type in xwininfo on a terminal, it will prompt a request for our selection:

Please select the window about which you
     would like information by clicking the
     mouse in that window.

If we choose a window, it will give us lots of information about the selected window:

$ xwininfo

xwininfo: Please select the window about which you
          would like information by clicking the
          mouse in that window.

xwininfo: Window id: 0x402a18c "How to grab the desktop with FFmpeg "

  Absolute upper-left X:  94
  Absolute upper-left Y:  337
  Relative upper-left X:  0
  Relative upper-left Y:  0
  Width: 1272
  Height: 605
  Depth: 24
  Visual: 0x21
  Visual Class: TrueColor
  Border width: 0
  Class: InputOutput
  Colormap: 0x20 (installed)
  Bit Gravity State: NorthWestGravity
  Window Gravity State: NorthWestGravity
  Backing Store State: NotUseful
  Save Under State: no
  Map State: IsViewable
  Override Redirect State: no
  Corners:  +94+337  -0+337  -0--174  +94--174
  -geometry 1272x605-0+337

We want to extract two information: top-left coordinates and width/height.

However, we can still use the xwininfo interactively with a pipe. The following command will wait for us to specify the window we want:

$ xwininfo | grep -e Width -e Height -e Absolute

Once we select the window, it gives us only the information we want to extract:

$ xwininfo | grep -e Width -e Height -e Absolute
  Absolute upper-left X:  66
  Absolute upper-left Y:  52
  Width: 1262
  Height: 686

As shown in the example below, we can use the information from xwininfo, and make ffmpeg to get the right area to grab:

$ ffmpeg -video_size 798x400 -framerate 25 -f x11grab -i :0.0+492,210 -an xwininfo.mp4

Here is the code : scrcap.py:

# This code runs the following awk to get a window id for the currently active X11 window
# xwininfo -id $(xprop -root | awk '/_NET_ACTIVE_WINDOW\(WINDOW\)/{print $NF}')

# ffmpeg -video_size $resolution -framerate 25 -f x11grab 
#        -i :0.0+originX,originY -f alsa -ac 2 -i pulse capture.mp4

import subprocess

def capture():
	origin, resolution = getWindowGeometry()
	originScreen = ':0.0+' + origin[0] + ',' + origin[1]
 	cmd = ['ffmpeg','-video_size',resolution,'-framerate','25', 
 	       '-f','x11grab','-i',originScreen,'-f',
 	       'alsa','-ac','2','-i','pulse','capture.mp4']
 	cmd = map(lambda x: str(x), cmd)
	subprocess.call(cmd)

def getWindowGeometry():
	info = py_xwininfo()
	valid_info = []
	winDict = {}
	for item in info.split('\n'):
		if item.count(':') == 1:
			(key, value) = item.split(':')
			winDict[key.strip()] = value.strip()

	origin = []
	origin.append(winDict['Absolute upper-left X'])
	origin.append(winDict['Absolute upper-left Y'])

	width = winDict['Width'] 
	height = winDict['Height']
	if int(width) % 2 != 0:
		width = int(width) + 1
	if int(height) % 2 != 0:
		height = int(height) + 1
	resolution = str(width) + 'x' + str(height)

	return origin, resolution

def py_xwininfo():
	winId = getCurrentWinId()
	cmd = ['xwininfo','-id', winId]
	cmd = map(lambda x: str(x), cmd)
	p = subprocess.Popen(cmd, stdout = subprocess.PIPE)

	return p.communicate()[0]

def getCurrentWinId():
	cmd_1 = ['xprop', '-root']
	cmd_2 = ['awk', '/_NET_ACTIVE_WINDOW\(WINDOW\)/{print $NF}']
	p1 = subprocess.Popen(cmd_1, stdout = subprocess.PIPE)
	p2 = subprocess.Popen(cmd_2, stdin = p1.stdout, stdout=subprocess.PIPE)
	id = p2.communicate()[0]

	return id

if __name__ == '__main__':
	origin, resolution = getWindowGeometry()
	print 'resolution = %s' %resolution
	print 'origin = %s' %origin

	capture()

The problem of the code is it always gets the capture for the terminal where we run the python code. So, we need to give the users a room for putting their selection of the window to capture.

Python needs to be able to capture mouse event before it captures screen.

Coming - capture with Python & OpenCV...