Synchronizing a very long video with very long waveforms (8 hours), and playing back a subset

[JKSH]

Hello,

I'd like to record 8 hours worth of waveforms and (low-resolution) video for an experiment. I've seen DIAdem's examples for synchronizing videos and waveforms, but these were all quite short (~3500 samples per channel).

Would DIAdem be able to handle such synchronization for long recordings (~576 000 samples per channel)?

Also, the person who reviews the data wouldn't be watching all 8 hours of video. If we had one 8-hour video and all the data was logged into one TDMS file, is it possible to synchronize and play back only a subset of the recordings (say, 10 seconds at a chosen point)?

Thanks!

[Brad_Turpin · Accepted solution]

Hi JKSH,

I don't think you'll have any trouble in DIAdem with the waveform data, even if the channels get very large.  The difficulty is likely to be a very large video file.  If you replay the video from the beginning at the standard replay rate (like Windows Media Player does), then it should work fine.  The trouble will occur the moment you deviate from standard incremental replay to fast or slow replay or random access with the VIEW cursor.  How much trouble you have will depend on the video format.  If it stores each frame without compression, you should be fine.  The common approach, though, is to store only the deviations from frame N to frame N+1.  This means, though, that in order to jump straight to frame 10,000 you would need the video codec to render 9,999 frame transitions.  Some video formats have "frame markers", which are occasional frames that are stored in their entirety.  In that case you could jump efficiently to the most recent (previous) frame marker, then render the relative deviations from there up to the frame you selected.

In general large video replay in DIAdem can be tricky.

Brad Turpin

DIAdem Product Engineer

National Instruments

[JKSH]

Thanks for your detailed reply, Brad.

I'll create test videos in different formats and see how they perform.