• State
  • Locked Locked
  • Replies 7 replies
  • Subscribers 31 subscribers
  • Views 400 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.

Original question:

DLPC350: 8bit pattern sequence result makes border

DLPC350: Unwanted "Edges" in projected gradient images

Tim Betker
Prodigy 40 points
Part Number: DLPC350

Hi,

my question refers to another thread posted before: "DLPC350: 8bit pattern sequence result makes border".

I am observing the same behavior. I am using the LCR4500 EVM with cameras for structured light measurement tasks.

As seen in the image below, there is a visible "edge" in the middle of the image. The image is taken by a camera which is triggered by the projector. Exposure times are matched.

For the usual measurement I am using sine patterns with different frequencies and those "artifacts" are affecting the overall measurement quality. In case of the sine patterns, the "edges" are located at the turning point of the sine wave.

My thoughts so far:

  • This effect seems to occur at the middle of the displayable color depth - ~127 for 8-bit uint.
  • It may be caused by the PWM of the DMD? - I played with the image exposure time and was able to catch the first division of illuminated and not illuminated regions. The edge seems to occur on the border of those regions.

Any thoughts or suggestions on how to approach this?

  • 0
    TI__Genius 12785 points

    Hi Tim,

    Welcome to the E2E Forum!

    Does this edge appear with different output images and settings? For example, does it appear with a splash screen, video mode, and internal patterns or is it specific to the camera triggered image?

    Regards,

    Austin

  • 0 in reply to Austin Snyder
    Prodigy 40 points

    Hi Austin,

    You can even see the effect with your eyes when looking on the projected gradient or sine pattern. It seems to be easier to see when using a rather low led current an a diffuse white surface. I do not need a lot of light since my working distance is at about 20cm.

    I am exclusively using the pattern sequence mode to project the patterns from flash memory. I tested different linear gradient images (horizontal, vertical and diagonal) and different sine patterns of various frequencies and directions. The edge always appears to be on somewhere in the middle of the displayable intensities (0 to 255). I even tried reducing the intensity range by only using values from 2..253 to prevent saturation effects without success.

    With a gradient image ranging from 0 to 127 there was no visible edge - which lead to my assumption that it has something to do with the DMD PWM and intensity encoding.

    Regards,

    Tim

  • 0 in reply to Tim Betker
    TI__Genius 9971 points

    Hi Tim, 

    We will try to reproduce this issue at our end and will share our findings soon.

    Regards,

    Mayank

  • 0 in reply to Mayank Khandelwal
    Prodigy 40 points

    Hi Mayank,

    thanks for your support, let me know if you need more information about the setup.

    Regards,

    Tim

  • 0 in reply to Tim Betker
    TI__Genius 10925 points

    Hi Tim, Could you kindly try experimenting with different LED rise and fall delays and see if it makes the issue better? Kindly see DLPC350 Programmers Guide 2.4.3.3 LED Enable Delay Controls for relevant information.

    Thanks & Regards,

    Hirak.

  • 0 in reply to Hirak Biswas
    Prodigy 40 points

    Hi Hirak,

    I played around with the LED delays inside the GUI but did not notice any changes. I feel like those settings do not affect the projection at all.

    Regards,

    Tim

  • 0 in reply to Tim Betker
    TI__Mastermind 28005 points

    Tim,

    If the issue is appearing at a specific RGB code, then the suggestions from Hirak would be likely to have an impact on system behavior.

    Some other suggestions to try (while modulating the LED delays);

    > Adjust LED current high-low

    > Individually enable-disable each LED and check ramps per color

    > Consider trying different patterns which employ the same "codes" to check if the issue moves.

    Regards,

    Philippe Dollo