Fastest way to take a screenshot with python on windows

What's the fastest way to take a screenshot on windows? PIL.ImageGrab is rather slow.. it takes between 4-5 seconds to take 30 screenshots of the same small window. Taking screenshots of the whole desktop is even slower.

You could use win32 APIs directly .

1. First give the focus to the App that you want to take screenshot of. link text

2. Win32 API can help with the screenshot:

import win32gui
import win32ui
import win32con

w = 1920 # set this
h = 1080 # set this
bmpfilenamename = "out.bmp" #set this

hwnd = win32gui.FindWindow(None, windowname)
wDC = win32gui.GetWindowDC(hwnd)
dcObj=win32ui.CreateDCFromHandle(wDC)
cDC=dcObj.CreateCompatibleDC()
dataBitMap = win32ui.CreateBitmap()
dataBitMap.CreateCompatibleBitmap(dcObj, w, h)
cDC.SelectObject(dataBitMap)
cDC.BitBlt((0,0),(w, h) , dcObj, (0,0), win32con.SRCCOPY)
dataBitMap.SaveBitmapFile(cDC, bmpfilenamename)

# Free Resources
dcObj.DeleteDC()
cDC.DeleteDC()
win32gui.ReleaseDC(hwnd, wDC)
win32gui.DeleteObject(dataBitMap.GetHandle())

Just found out how to do it with gtk. Seems fastest by far:

def image_grab_gtk(window):
    left, top, right, bot = get_rect(window)
    w = right - left
    h = bot - top

    s = gtk.gdk.Pixbuf(
        gtk.gdk.COLORSPACE_RGB, False, 8, w, h)

    s.get_from_drawable(
        gtk.gdk.get_default_root_window(),
        gtk.gdk.colormap_get_system(),
        left, top, 0, 0, w, h )

    final = Image.frombuffer(
        "RGB",
        (w, h),
        s.get_pixels(),
        "raw",
        "RGB",
        s.get_rowstride(), 1)
    return final

Without converting to a PIL Image, it's 8x faster than PIL on my test case. With converting, it's still ~2.7x faster.

You can use package mss:

Save screenshot to image file

import mss

with mss.mss() as sct:
    filename = sct.shot(output="output.png")

Get the numpy representation of screenshot

import mss
import numpy as np

with mss.mss() as sct:
    monitor = {"top": 160, "left": 160, "width": 160, "height": 135}
    img_array = np.array(sct.grab(monitor))
    # Do whatever you want...