NI-IMAQdx - My recorded videos are "accelerated"

[mthheitor]

Hello

I building a modular VI for a friction welding testing machine.

The communication between the VIs is done with queues and notifications. Part of my code is attached below.

One of the "sensors" of the machine is a USB-Câmera (Logitech HD Pro Webcam C920)

But when I record videos, they are "accelerated". Example: If I record 10 sec video, when I play it in VLC, the video have 8 sec (approximately)

I'm using the FF Video Codec 1 (FFV1) (NI Vision) to write a .avi file, here's the subVI that start the video and the VI that record:

subVI writeVideo2a.visubVI write Video2b.vi

These subVIs needs to be separated from the VI that is grabbing the camera image, because several VIs can record video.

How can I write a good compressed video without being a "accelerated one"?

[mthheitor]

I've made a simplified version of my code, to testing purposes.VI that records a video from a USB CameraBut still get the same problem, the timing of the videos are all wrong...

If I record 10 sec of video, the video file plays the same within 6-7 sec 

Have this ever happened with you? What can I do to solve this problem?

[udka]

Can you try with delay(1/30) in writing loop exactly with the fps(30) you are writing to the avi file?

[mthheitor]

@udka

I did what you suggested, but did not worked. The video continues being "accelerated". The delay did nothing 😞

What I've done to partially resolve the problem was change the codec from "FFV1" to "Motion JPEG (NI Vision)".

Now the videos are OK, but I think if other users try to use FFV1 codec they will be frustrated like I am now.

Here's my new code:

writeVideo

Something that took me a while to find out was how to change de resolution of the camera:

You have to find the VideoMode enum code with the "IMAQdx Enumerate Video Modes.vi", them create a property node "VideoMode" with the correct number.IMAQdx Enumerate Video Modes.vi

Using NI Vision directly on LabVIEW should be easier, maybe NI can work in this module.

Because there's a lot of questions on "Machine Vision" without solution