• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 27 subscribers
  • Views 677 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.

the max FPS of discovery 4100

liang cai10137
Prodigy 35 points

I have a problem about the max FPS of discovery 4100 if it display a 24bit image. Is this calculate right that the min time from first reset operation to the second reset operation is 30.72us(1.92×16), so if DMD display a 8bit grayscale image, the time is (30.72×256)us, as this way the max frame rate is 42.5fps for 24bit image. That is too little！Where is the mistake of above calculate.

  • 0
    TI__Intellectual 2265 points

    Liang,

    What DMD are you using? By 24-bit, do you mean 8:8:8? Assuming that is the case, depending on the DMD, a fast and easy way to calculate the frames per second is to divide the max binary fps by 3x2^8=768. This give you approximately the same number as what you have above (42.5pfs). However, this approach assumes that each bit of the 8 bits is weighted (each bit is exactly twice as long as the previous bit). This does not have to be the case.

     

  • 0 in reply to WYA
    Prodigy 35 points

    WYA,

    Thanks for your explaination. Could you tell me how to display the 8:8:8 image or video at 60Hz. Is it that after the DMD reset, a clear operation may be run. In this case, the DMD display time reduce half. However,  I can't delineate the sequence in time.

  • 0 in reply to liang cai10137
    TI__Intellectual 2265 points

    Liang,

    First of all, please note that there is development platform that is in preview mode called the LightCommander that would be better suited if you are interested in video projection. You can find out more information online at: www.ti.com/dlplightcommander

    The way to get to faster frame rate is to reduce the LSB time. If the LSB time is smaller, than the frame time is smaller (3x2^8*LSB). What you wrote is one way of getting to a smaller LSB, by using the reset-clear operation. This operation is helpful when you have a lamp based system where you can not turn off the light easily. If you have an LED based system, this operation is not necessary because you can turn off the LED. Note that even though the reset timing can not change, the amount of time you leave the LSB illuminated is the LSB time. So for example, if you want the LSB time to be 10us, and it takes roughly 53us between global resets, you can turn off the LED after 10us. You can then keep the next bit reset time to be still 53us, but leave the LED on for 20us, and so on. In this manner, the reset time for the first 3 bit would be the same, even though the illumination time is twice as long every time. So in your calculation, you would save the total fram time. One important thing for your calculation is that you have to add in the dark time (time when the LED is off, but reset is not done, i.e. in this example, 43us for the first bit, 33us for the second bit, etc.

    If you would like to reach 60Hz, you would need to find out the exact LSB time you would need, plus dark times.