I have a small python program that works very well to capture short videos from webcams in linux (at least for laptops that have built-in webcams) using a sub-process with ffmpeg.
Now i'm trying to write the same program to capture webcams in windows, and i know i can't use the generic "/dev/video0" that works pretty well in linux, but i thought something like naming it "Integrated Camera" should be enough, but it fails.
Here's my linux code (that works):
import sys
from subprocess import call
from datetime import datetime
def record_webcam(seconds):
cam = '/dev/video0'
timestamp = datetime.now().strftime('%Y%m%d-%H%M%S')
filename = timestamp + 'something.mkv' #generated with more complexity in the actual code, but that isn't important
ffmpeg_cmd = 'ffmpeg -t {} -an -i {} -c:v libx264 -preset veryslow -crf 25 {}'.format(seconds, cam, filename).split()
p = call(ffmpeg_cmd)
return filename if p == 0 else False
if __name__ == '__main__':
record_webcam(sys.argv[1])
I have looked at the documentation for ffmpeg and tried to search for solution but so far i'm lost...
I know that "Integrated Camera"s are only available on some laptops and not others, and that it won't capture other cameras connected, but it's enough for my use case... but if you want a challenge I would also like to know how to apply it to any windows-pc with a camera regardless of what it's called.
Also, is it easier or more recommended to do what i'm trying here only with python tools, like OpenCV?
Thanks in advance! Edit: I answered my own question with a partial solution if anyone is interested based on a comment from @Mulvya, but if anyone can still explain to me the part about OpenCV I would still like to hear it...
Follow up question here: ffmpeg through python subprocess fails to find camera
I did it... based on what @Mulvya commented, i was able to list all cameras withing the code and extract a command from them, i now have a different issue about that, but i will ask another question for it. In the meanwhile, if anyone is interested in automatically selecting the first available camera on windows through python and ffmpeg, my solution is this:
import re
from subprocess import Popen, PIPE
list_cmd = 'ffmpeg -list_devices true -f dshow -i dummy'.split()
p = Popen(list_cmd, stderr=PIPE)
for line in iter(p.stderr.readline,''):
if flagcam:
cam = re.search('".*"',line.decode(encoding='UTF-8')).group(0)
cam = 'video=' + cam if cam else ''
flagcam = False
elif 'DirectShow video devices'.encode(encoding='UTF-8') in line:
flagcam = True
elif 'Immediate exit requested'.encode(encoding='UTF-8') in line:
break
the variable "cam" now holds the name of the cam as it is in DirectShow on windows
Follow up question here if anyone wants to help EDIT: also solved