LabVIEW
Creating screen recording app with labview
I tried all the examples made in labview,
but there is a problem with the save speed.
When I measured the while loop refresh rate it was 900 ms. I came to the conclusion that the user32.dll file was causing the slow operation.
No matter what I do, I cannot record the recording at normal speed. The video is recorded with very fast steps. Because the number of photos it takes is very low. Does anyone know the solution to this?
Why is the loop timing running so slow?
Any opinion given is worthy of respect
To get faster capture you have to use Windows GDI functions instead of Print Screen and Clipboard.
Most easiest way is to build small wrapper library, quick and dirty something like that:
LabVIEW code:
On my PC capturing running approx at 30 FPS:
Full project in attachment (LabVIEW 2022 + VDM + CVI 2020).
@Andrey_Dmitriev wrote:
To get faster capture you have to use Windows GDI functions instead of Print Screen and Clipboard.
Most easiest way is to build small wrapper library, quick and dirty something like that:
LabVIEW code:
On my PC capturing running approx at 30 FPS:
Full project in attachment (LabVIEW 2022 + VDM + CVI 2020).
I'm using labview 2018 32-bit, I can't open the file.
Thanks for your help. I think I understand the problem better now. The reason for the FPS drop is 2 screen usage. The example you shared is 30 FPS with one screen and 15 FPS recording with two screens. codec:Motion JPEG (NI Vision)
This is how it worked with the example I shared earlier.
But this FPS problem does not occur when I record two screens with a normal recorder application.(I'm recording with mp4 in lower quality. I think there is something about this.)
The CPU I am using is i7-7500U CPU 2.70 Ghz
What do you think is the solution to this problem?
Well, if you perform recording of two monitors with large desktop, then performance can drop.
The only enhancement point what you can do is to split capturing and encoding into two loops using producer/consumer design pattern, something like this:
Full code in the attachment (LV2018)
For further enhancement there are two obvious points - capturing and encoding. GDI+ is not the only possible way to capture, may be something DirectX-based will be faster (but I'm not very familiar with this). Encoder can be replaced with OpenCV (you will get more codecs available). Also using ffmpeg could be possible. If your actual recording tool is open source you can take a look inside how it works and which technology used behind, but integration into LabVIEW (if you prefer to have recorder frames in IMAQ images) could be painful.
