Desktop Duplication API doesnt track mouse movements(and its correct), we can add an option to capture mouse and draw it

Hi,

I've been trying to compile the package from source but getting error regarding: C:\Microsoft.Cpp.Default.props. I've installed CMAKE and Visual Studio 14 2015 x64 build tools. After moving to directory, i'm simply running the cmake_build.cmd file but no success. The command i ran:

$ ./cmake_build.cmd

Error:

Output:

[2020-10-20 22:03:13.238] [recordLogger] [info] Created d3d11 device from 0
[2020-10-20 22:03:13.238] [recordLogger] [info] found 3 displays
[2020-10-20 22:03:13.238] [recordLogger] [error] desktopdiplication is unsupported on current desktop mode

I have 3 monitors, but I also tried it with one, and it still didn't work.

I have recorded a video with start_video_recording and handed in different values for fps.
However, this did not have any effect, the frame rate of the video was always variable, around 20, although the parameter had a value of e.g. 3 fps.

Traceback (most recent call last):
File "E:\python prog\screen_recorder.py", line 85, in
main ()
File "E:\python prog\screen_recorder.py", line 67, in main
pid = int (sys.argv[1]) # pid == 0 means capture full screen
IndexError: list index out of range

win10 x64
monitor: 1366x768

Traceback (most recent call last):
  File "E:\ChatBotAI\AiCharli.py", line 28, in <module>
    main ()
  File "E:\ChatBotAI\AiCharli.py", line 10, in main
    pid = int (sys.argv[1]) # pid == 0 means capture full screen
IndexError: list index out of range

Process returned 1 (0x1)        execution time : 0.299 s

Record sound for video

I was testing on a virtual machine without a GPU and to my surprise the processor consumption of it is too high and slowing down the other processes. Is there any way to make it consume less power?

Use https://docs.microsoft.com/en-us/windows/win32/api/d3d11/nf-d3d11-id3d11devicecontext-copysubresourceregion to get subregion

Currently we determine process window position on the desktop in the beginning and use these coordinates. Probably we can do it in real time and handle the case if process window was moved

Hi,

When trying to record a vertical display, the output video file is completely black.

I also do not get any errors or message out of the ordinary in the log.

Do you know what causes this issues and how I could solve it?

Traceback (most recent call last):
File "d:\Documents\Programming\PROJECTS\scr_cap.py", line 28, in
main ()
File "d:\Documents\Programming\PROJECTS\scr_cap.py", line 10, in main
pid = int (sys.argv[1])
IndexError: list index out of range

I had guessed that the order of desktop_num values is the same as the Identify order in the 'Display settings', or the get_monitors() order in the package 'screeninfo'.
However I only found that the primary display is always 0, and I don't know in what order the rest are numbered.
Could you tell me if there is a callable function from ScreenRecorder.dll that can get desktop_num and other key values corresponding to each displays or desktops?

Hello , i'm getting following error when i try to record something ,please help me with the following issue:
[2020-08-25 17:09:11.433] [recordLogger] [error] Could not init InitializeSinkWriter
[2020-08-25 17:09:11.435] [recordLogger] [error] MFEncoder is not null
[2020-08-25 17:09:11.438] [recordLogger] [error] Could not init InitializeSinkWriter
[2020-08-25 17:09:11.441] [recordLogger] [error] MFEncoder is not null
[2020-08-25 17:09:11.444] [recordLogger] [error] Could not init InitializeSinkWriter
[2020-08-25 17:09:11.450] [recordLogger] [error] MFEncoder is not null

Is there any method to select which monitor to capture the screen from? because using only PID produces some limitations and setting its value to 0 to capture the main monitor is also not so functional when having more than one monitor.

HI, first of really great for pulling this off. Its really helpful to me. I've just found an issue using the library when doing the final compile in a program. The library doesn't get detected by PyInstaller or other compilation tools. The imports do get detected and works fine, however the ScreenRecorder.dll file doesn't get detected and is not placed in the compiled directory. I tried the following commands:

$ pyinstaller file.py
$ pyinstaller --hidden-import screen_recorder_sdk file.py

At the end, i'd to go to the directory where the library is installed and manually copy the whole and place in compilation directory.

When compiling with Visual Studio 14 2015, the cmake command throws the following error:

C:\Users\sheik\Documents\Projects\screen_recorder_sdk\build>cmake --build . --config Release
Microsoft (R) Build Engine version 14.0.25420.1
Copyright (C) Microsoft Corporation. All rights reserved.

  Checking Build System
  Building Custom Rule C:/Users/sheik/Documents/Projects/screen_recorder_sdk/CMakeLists.txt
  ScreenRecorder.cpp
C:\Users\sheik\Documents\Projects\screen_recorder_sdk\src\ScreenRecorder\inc\RecorderDDA.h(6): fatal error C1083: Canno
t open include file: 'd3d11_4.h': No such file or directory [C:\Users\sheik\Documents\Projects\screen_recorder_sdk\buil
d\ScreenRecorder.vcxproj]
  Recorder.cpp
  RecorderDDA.cpp
C:\Users\sheik\Documents\Projects\screen_recorder_sdk\src\ScreenRecorder\inc\RecorderDDA.h(6): fatal error C1083: Canno
t open include file: 'd3d11_4.h': No such file or directory [C:\Users\sheik\Documents\Projects\screen_recorder_sdk\buil
d\ScreenRecorder.vcxproj]
  MFEncoder.cpp
  HandleD3DWindow.cpp
  Generating Code...

I've Direct X installed. And the compilation works fine when trying from Visual Studio 17 2022 but throws this error from 14 2015 version.

I have numpy preinstalled in my Windows but when trying to install screen recorder sdk it fails to build numpy. How can I solve this?

Add an option to capture webcam and join it with recorded video/screenshots

Hi, first of all good work creating this, it's easy to use and does the job.
I started using the lib for my nightly automated regression tests in order to capture video records of the tests that are failing, to help with debug.
Since then I'm seeing some errors with Insufficient memory thrown by opencv.
The thing is that the error comes when I try to do some image recognition using a different library that uses opencv.
Yesterday I deactivated screen recording during the tests and the errors are gone.
Maybe there's a memory leak here that prevents opencv to allocate memory for other tasks.

Here is the error I get:
File "C:\GitLab-Runner\builds\hbav6Snu\0\COP_QA\COP\d2c_regression\libs\image_utils.py", line 66, in locate
return pyautogui.locateOnScreen(ELEMENTS_PATH + element_image, minSearchTime=3, confidence=confidence)
File "c:\work_dir\cop_regression\lib\site-packages\pyscreeze_init_.py", line 360, in locateOnScreen
retVal = locate(image, screenshotIm, **kwargs)
File "c:\work_dir\cop_regression\lib\site-packages\pyscreeze_init_.py", line 340, in locate
points = tuple(locateAll(needleImage, haystackImage, **kwargs))
File "c:\work_dir\cop_regression\lib\site-packages\pyscreeze_init_.py", line 219, in _locateAll_opencv
result = cv2.matchTemplate(haystackImage, needleImage, cv2.TM_CCOEFF_NORMED)
cv2.error: OpenCV(4.1.0) C:\projects\opencv-python\opencv\modules\core\src\alloc.cpp:55: error: (-4:Insufficient memory) Failed to allocate 49838424 bytes in function 'cv::OutOfMemoryError'

Also here's the simplified code that I'm using to capture screen records, maybe I'm not using this correctly.

@pytest.fixture(scope="function", autouse=True)
def _capture_screen_record(self):
    screen_recorder.disable_log()
    params = screen_recorder.RecorderParams()
    screen_recorder.init_resources(params)
    screen_recorder.start_video_recording("temp_screen_record.mp4", 20, 4000000, True)
    yield
    screen_recorder.stop_video_recording()
    screen_recorder.free_resources()

Any input is appreciated, thanks

Hi,
I got pid of browser instance which i passed but still getting below error. help me pls.
[2020-07-17 11:28:57.177] [recordLogger] [info] Created d3d11 device from 0
[2020-07-17 11:28:57.178] [recordLogger] [info] found 1 displays
[2020-07-17 11:28:57.182] [recordLogger] [error] Unable to find window for pid 21708
[2020-07-17 11:28:57.182] [recordLogger] [error] Unable to get process window

Part of Code -

screen_recorder.enable_dev_log()
params = screen_recorder.RecorderParams()
screen_recorder.init_resources(params)

Output (as soon as i run the code) -

[2020-11-09 16:22:46.415] [recordLogger] [info] Created d3d11 device from 0
[2020-11-09 16:22:46.415] [recordLogger] [info] found 1 displays
[2020-11-09 16:22:46.418] [recordLogger] [error] invalid arg to create desktopdiplication
[2020-11-09 16:22:46.418] [recordLogger] [trace] stop video recording is called
[2020-11-09 16:22:48.433] [recordLogger] [trace] wait res for stop video is false