To answer your question, I would recommend using the LabVIEW display VIs as opposed to ActiveX objects if you are worried about the screen update rate from code to display. You can definitely benchmark the time required to run ActiveX code as compared to LabVIEW display VIs, but you cannot benchmark when the interface has actually updated. However, because ActiveX has an additional layer of communication, I would assume that this could potentially extend the time that this operation would take.
For your application as a whole, I also have other reservations about the timing requirements you have. Specifically, you might not be able to ensure that you have accurate correlation within 1 ms depending on the specifications of the machine and concurrent processes running. By this I mean that the screen update for any operation depends on how fast Windows can process the command to redraw the screen. If you have a slow machine, or a large amount of concurrent tasks for the processor, this timing will be more of an issue.
The ideal platform that I would suggest would be a Windows machine dedicated to the display of these images. You could then run the acquisition code on a real-time processor and use the Windows host as a time server. By doing this, you can then correlate the timestamps of a screen update with the DAQ timestamps. I hope this information is useful for you. Please post back if you have any further questions.
Mike D.
National Instruments
Applications Engineer
