How to determine original FPS of video?

Discussion in 'Photography & Video' started by motas, Oct 10, 2018 at 12:24 PM.

  1. motas

    motas Member

    Joined:
    Oct 14, 2009
    Messages:
    2,312
    Hey guys,

    I am guessing the answer is no, but hoping someone might have a trick!

    I need to do a comparison video of two crash tests to prove that one had the data edited. I have an old video filmed at an unknown FPS, and playing at 50 FPS. I also have a new video of a repeat crash test filmed at 1000 FPS and playing to 30 FPS.

    I have made a split screen of the two videos and it looks like the old video was filmed at 240 FPS and that seems to line up well, but I'd like a way to confirm that. Is there any way to confirm this? If I can I can calculate the actual speed of the vehicle, compared to the claimed speed.

    Unfortunately I can't share the old video, but here's another crash test video for your time:



    Thanks!
    Tom.
     
    Last edited: Oct 10, 2018 at 4:28 PM
  2. Cererus

    Cererus Member

    Joined:
    Jan 14, 2002
    Messages:
    1,391
    Location:
    ≥∞
    If the test was shot side on to the vehicle find out what the standard fitted tire size was, workout the circumference then track how many FPS it takes to make one rotation you should then have the speed in m/s convert to kmh and see if that agrees with the claimed speed?
    (not sure if this is on track, just an idea)
    What are you trying to prove, that a test was faked claiming a higher speed than the crash actually happened at.
     
  3. OP
    OP
    motas

    motas Member

    Joined:
    Oct 14, 2009
    Messages:
    2,312
    The test was "conveniently" shot at an angle, and I don't know what speed the video is filmed at, which is the issue I'm having.

    I won't go in to too much detail, but a product was tested in-house by the inventor and approved based on that test. A functionally identical product won't be approved as it does not pass the test.

    It is very obvious that the test is faked. The video is obviously slower than the claimed speed, and the data in the report is blatantly halved midway through the report. It is also physically impossible to pass the requirements of the test based on the equations of motions.

    We have tried to explain this in the past to the approval agency, but we'd like to present the two supposedly identical tests side by side and show the significant difference in results. It is difficult to line the videos up as they are recorded at different speeds, and the best I can do at the moment is guess what looks reasonable.
     
  4. mtma

    mtma Member

    Joined:
    Aug 12, 2009
    Messages:
    4,605
    If you believe that the footage had some fractional slowdown then some hints of this would have been present, but if it's simply faster footage played slower frame by frame you couldn't really tell without using something in the content as reference.

    Things to look at:

    1. LED's, along with some other forms of lighting, will beat at characteristic frequencies and could be used for reference. Also if there are any scanned displays within the video they may also show up banding, characteristic of being shot at fast shutter speeds.

    2. Vibrations: Basically, for example if the test contained the same model cars, there would be technical relationships between the vibrations of the vehicle structures. For example, you will probably find that the bonnets across the same models of Dodge will oscillate at similar natural frequencies when subject to similar (or even dissimilar) collisions. Large flat panels and things also have characteristic vibration modes based on their dimensions, mass and properties, so based on reference information, it is possible to infer a possible timing of the footage.

    3. If there's some sort of lighting reference that would be repeatable, that could be used to determine the exposure time. But of course the exposure time doesn't tell you anything about the actual FPS, aside from that it couldn't have been faster than a certain rate.

    4. File tags. Could be something that carried through in the tags.
     
  5. OP
    OP
    motas

    motas Member

    Joined:
    Oct 14, 2009
    Messages:
    2,312
    Sorry to be so confusing, it would make a lot more sense if I could upload the video, but I don't know that it is a good course of action. This has the potential to end up in a court battle!

    We use sensors in the vehicle to determine the speed not the video itself. I believe the data from the sensors has been adjusted, but the video is correct. It is general practice for the videos to be shown in slow motion, we film at 1000FPS and play at 30FPS so our videos are ~30 times slower than real life. The whole impact is over within about half a second real speed. You can easily see the difference in speed from the severity of the impact on the car, but it would be nice to calculate it for sure rather than by eye. The difference isn't fractional either, our estimation is the "slow" test is around 30-40km/h instead of 60km/h.

    Unfortunately there is nothing but natural lighting, and the video is fairly old and poor quality so I doubt the vibrations will be visible but I will check it out.

    I was hoping for a file tag showing what model camera it was filmed on but unfortunately there isn't any.

    As I said I didn't really expect it to be possible but was hoping there was some trick I didn't know about!

    Thanks for the help!
     
  6. OP
    OP
    motas

    motas Member

    Joined:
    Oct 14, 2009
    Messages:
    2,312
    I'm an idiot, didn't think to check the report.

    The camera was listed, and is capable of 250FPS maximum, so my guess was pretty close.

    Thanks again for the help anyway!
     
  7. Squeezer

    Squeezer Member

    Joined:
    Feb 25, 2002
    Messages:
    6,837
    Location:
    Adelaide

Share This Page