• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 27 subscribers
  • Views 475 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DLP4500: Diffraction from Coherent Source Different Between Pattern Sequence and HDMI

Alexander Sludds
Prodigy 120 points
Part Number: DLP4500
Other Parts Discussed in Thread: DLPC350

Hi everyone,

I am using a coherent source (633nm) to illuminate the DMD array. If I use the pattern generation in the GUI with a standard image (a lattice of one pixels for example) the diffraction pattern off of the DMD matches a model that I created. However, if I pipe the same pattern over HDMI using opencv to display the pattern in fullscreen on the "second monitor" DMD I don't get the same diffraction pattern. This makes me think that over HDMI I am not getting a pixel perfect match with the image I am sending in.

I was wondering if anyone could recommend a way of getting pixel perfect control of the DMD mirrors over HDMI. I will have numpy arrays that I want to "display" on the DMD and I have to have precise spatial control of them.

Thank you for your time.

  • 0
    TI__Mastermind 23070 points
    Hello Alexander,
    Welcome to DLP forum and thank you for your interest in DLP technology.

    Could you please details of GUI settings? A screen shot will be very helpful.

    You may also want check the following:
    1. In case of video, the input data undergoes series of image processing functions.
    2. The DLP4500 has diamond pixel and input data needs to match pixel addressing.

    regards,
    Vivek
  • 0 in reply to Vivek Thakur
    Prodigy 120 points

    Hi Vivek,

    I have attached a copy of the settings I am using for the DMD.

    Can you be more specific about what image processing the DMD is doing?

    The display area listed in the TI software is 1280 by 800, but I am unclear what that is corresponding to. If I create a box in OpenCV and move it around I find that the corners of the DMD are at (0,0) (1140,0) (0,912) (1140,912) which matches the DMD array size. So it is unclear to me what this 1280 by 800 display area is.

  • 0 in reply to Alexander Sludds
    TI__Mastermind 23070 points

    Hi Alexander,
    In the video mode , the EVM is considering the input image as a video signal and doing several image processing before displaying on to DMD. As result the image displayed on DMD is not one to one pixel map from input.

    The figure 13 on page 28 of DLPC350 datasheet has function block diagram of the controller. It is lists image processing functions of the device.
    www.ti.com/.../dlpc350.pdf


    Another key difference is that system is treating input as 1280x800 image (in Manhattan grid) and converting it to diamond pixel grid (912x1140) of DLP4500 DMD. Please refer to figure 7, page 25 of DLP4500 DMD datasheet for pixel mapping/grid.
    This step requires image re-sampling. As result the image displayed on DMD is not one to one pixel mapped to input image.


    In the video mode of GUI , the display area parameters are referring to input image and cropped area is referring to target pixels in DMD.

    For pixel perfect control of the DMD mirrors over HDMI, you need to run in pattern mode.

    I hope this explains.

    regards,
    Vivek