gbugnard
October 11, 2007, 21:19:33
Hi
Is it possible to get the time delay between the start of a streaming with the method "AviStartCapture" and the real beginning of the capture?
I am using a DFG/1394-1e with LabView and obviously, there is a time shift of about 80 to 200 ms between the start of the capture in LabView and the real capture start.
To control the sync, the camera films a portion of LabView's front panel with a digital display of the actual time in LabView. The image in the movie always starts between 80 or 200 ms (the sample frequency is 25 Hz). After the test, I extract every frame out of the AVI with QuickTime Pro and name the image file with the frame number. With this method, one can compute the theoretical time of the frame:
Time = (i-1) * 40ms/frames + offset
with i the frame number. With this formula, the theoritical time always corresponds to the time on the image for all the tests I ran. The only problem I have, the offset is always different. Is it possible to compute it directly during the experiment?
As example the frame 1 and 167 of one of the test. The first frame starts at 121ms. The theorital time for the frame 167 is then: (167-1)*40ms + 121ms = 6761ms which is quite the same as on the picture.
Thanks!
Is it possible to get the time delay between the start of a streaming with the method "AviStartCapture" and the real beginning of the capture?
I am using a DFG/1394-1e with LabView and obviously, there is a time shift of about 80 to 200 ms between the start of the capture in LabView and the real capture start.
To control the sync, the camera films a portion of LabView's front panel with a digital display of the actual time in LabView. The image in the movie always starts between 80 or 200 ms (the sample frequency is 25 Hz). After the test, I extract every frame out of the AVI with QuickTime Pro and name the image file with the frame number. With this method, one can compute the theoretical time of the frame:
Time = (i-1) * 40ms/frames + offset
with i the frame number. With this formula, the theoritical time always corresponds to the time on the image for all the tests I ran. The only problem I have, the offset is always different. Is it possible to compute it directly during the experiment?
As example the frame 1 and 167 of one of the test. The first frame starts at 121ms. The theorital time for the frame 167 is then: (167-1)*40ms + 121ms = 6761ms which is quite the same as on the picture.
Thanks!