• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 29 subscribers
  • Views 3129 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

relation between the CCD camera frame rate and the DLP output frequency in the digital fringe projection technique

Xinjun Zhu
Prodigy 170 points

Dear all,

I have a question about the relation between the CCD camera frame rate and the DLP output frequency in the digital fringe projection technique. For example, if the DLP LightCrafter output frequency is set to be 120Hz, then what is the maximum CCD capture frame rate we can use? Is it 120Hz ?

Best wishes,

Xinjun Zhu

  • 0
    TI__Expert 4720 points

    Hello Xinjun,

    Yes, you are correct that the frame rates would ideally the the frame rates would be identical. The issue is not necessarily the output frequency, but the exposure times and delays between the camera and projector. 

    Before I go into further detail about the exposure timings, it is critical that you use triggering for proper synchronization. Either the projector needs to trigger the camera captures or the camera needs to trigger the projector. In other words, synchronizing the projector and camera becomes exponentially more difficult if both the camera and projector are in a "free-run" operation mode.

    As mentioned earlier, ideally there is no delay between the camera and projector which means they can have the same exposure times. The reality is that there will always be some delay between the exposures. In the case that the camera triggers the projector, this means the projector exposure goes beyond the camera's exposure and will miss some of the projected data. In addition to missing data, this also means the the next camera capture will have some of the previous projected pattern! 

    In my experience, I have found the best practice is to have the camera trigger the projector and set the projectors exposure to be less than the camera's. Setting the camera exposure larger than the projector exposure guarantees the entire projected pattern will be captured during the camera exposure. 

    The figure below show's: (top) what synchronization looks like in a perfect world, (middle) what really happens if no adjustments are made, and (bottom) the recommended mode of operation.

    If you are projecting 8-bit patterns with the LightCrafter 4500, the minimum allowable exposure is 8.333 ms which equates to 120 Hz, but your camera exposure probably needs to be closer to 10 ms to ensure the 8-bit pattern is properly captured. This means your overall frame rate would be 100 Hz.

    If you can use a lower bit-depth pattern, say 7-bit, the LightCrafter 4500's minimum exposure is 4.5 ms. With such a configuration your camera could easily be set to 120 Hz and the projector exposure set to 6 or 7 ms to achieve good synchronization. 

    Let me know if you have any more questions or I didn't explain something clearly!

    Best regards,

    Blair

  • 0 in reply to Blair Conner
    Prodigy 170 points

    Dear Dr. Blair,

    Thank you very much for your help.

    I learned a lot from your answers above and thank you for your suggestion on the exposure times setting. But here I have another question.

    As you say, " In addition to missing data, this also means the the next camera capture will have some of the previous projected pattern!". Now in my digital fringe projection experiment, I only use one static fringe pattern. In this case, is it reasonale to infer that the static projected fringe is not changeable at any time, because the "previous projected pattern" is the same as the " present projected pattern"? Suppose that the fringe pattern is not changeable at any time, I think there is no need to synchronize the projector and camera for static fringe pattern. In another words, the camera frame rate can be as large as possible regardless of the exposure time of projector if the projected fringe pattern is static.

    The doubt also relates to another question (also for static fringe pattern projection): How is the transformation between the previous projected pattern and the present pattern. Is it flashing between the two projected patterns.

    Are there somethings wrong with my above statement?

    Best wishes,

    Xinjun Zhu

     

    Best wishes,

    Xinjun Zhu

  • 0 in reply to Xinjun Zhu
    TI__Expert 4720 points

    Hello Xinjun,

    In theory, yes you are correct, however in practice you'll find that the frame rates won't be an exact match and you'll run into the problem I described previously.

    To show you what I mean, I setup my projector to project a simple RGB ramp at 60Hz and my camera at 60Hz. You can see in the GIF below that there is are differences between capture frames even though their frame rates are supposedly equal to each other.

    It is possible to adjust the frame rates until they are as closely matched as possible, but using triggers is the simplest solution is to guarantee proper synchronization.

    For your second question, the DMD constantly updates the mirror positions even when the image does not change.

    Best regards,

    Blair

  • 0 in reply to Blair Conner
    Prodigy 25 points

    Blair,

    Thanks for being so attentive to these boards and giving thorough helpful replies to everyone's questions. It has made learning the DLP and DMD technology much easier.

    I am triggering patterns on the LC 4500 and triggering a camera to capture the exposure of each pattern in an individual frame, similar to what you have described above. I am triggering the patterns and the camera independently (as opposed to using one to trigger the other), but I am generating TTL pulses in Matlab for both so as to maintain proper synchronization.

    My question is, when triggering the DMD patterns externally, how do I control the amount of time the pattern is displayed for? I have seen other posts saying that a single pattern exposure will last for the entire duration between two triggers, but I want something like you have described above, where the projector exposure stops before the camera exposure is finished, and the next pattern will wait to display until the DMD receives the next trigger. For my application, the easiest way to handle this would be to leave the projector on as long as the voltage is being applied, and once the voltage drops back to zero the projector turns off and remains off until the next rising edge. Can this type of trigger be implemented? If not, how do I adjust the projector exposure (and specifically what you have labeled as the safety band in your diagram) using external TTL triggers?

    Thanks!

  • 0 in reply to Matthew Reisman
    TI__Expert 4720 points

    Hi Matthew,

    All of the DLP platforms have exposure and period timing settings for their pattern sequence modes.You just need to set these parameters to the desired value. Please reference the respective chip set user/programmer's guide for specific instructions.

    With regards to your application triggering idea, please start another thread so we don't get too off topic for this particular one. There are a couple different triggering modes, it's possible we could create something similar to what you've described.

    Best regards,

    Blair