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.

about smooth picture



Dear all,

          Whether  the  smooth picture technology  is used in 0.45 dmd  diamond arrays(912*1140) displaying?

wanna know how its pixels correspond to 1280*800 picture. Thanks.

  • Hi Joyce,

    Can you clarify what you're asking? I don't understand your question. Is there are particular product you're using that has Smooth Picture Technology?

    Thanks,
    Paul
  • Hi Paul,
    My production is DLP Lightcrafter 4500, and its micro-mirror array is diamond by 911*1140. I learned that Smoothpicture technology is used in diamond array to better image displaying and wanna know whether it is used in this product. If so,how the diamond array mirror displays one to one correspondence to 1280*800 image?
    For an accurate image display in 3D measurement, pattern mode is used. and someone told me that creat the image of size 1280*800 and then stretch the image to 912*1140 using image editor and then upload it to flesh memory through the GUI and display by pattern mode. I don't know it is correct or not. So what is the correct way to display the accurate image?

    Thanks,
    Joyce
  • Joyce,

    Smoothpicture is an older technology that was used in TI DLP television sets. It is not used in the Lightcrafter 4500.

    The correct way to display an accurate image is to use pattern sequence mode as you are doing. Due to the diamond configuration, a 912 x 1140 pixel image will be displayed with an aspect ratio of 16:10. If true pixel accuracy is important to you, I would recommend that you work with a 912 x 1140 pixel image to begin with and understand that when it is projected that it will look "stretched."

    The recommendation you received before is not incorrect but it doesn't allow you to control exactly where each pixel is in the image, if that is your highest priority. I say this because you don't control what happens during the resize step in the image editor. You mentioned 3D scanning, so if you're using a binary line pattern, either method should be okay, because that particular pattern isn't terribly sensitive to this issue.

    In any case, read question 3 in the FAQ: e2e.ti.com/.../3425.alc-general-faq and understand that in pattern sequence mode, the Lightcrafter is pixel accurate.

    Hope this helps,
    Paul
  • Hi Paul,
    Appreciate for your response very much.

    That's right. I ever directly input a 912*1140 pixel image with some circles in the flash memory and circles in the image were "stretched" to oval in pattern mode display.

    In 3D scanning, the binary line images are used, however, the calibration is the first step you know. Here we used calibration target with circles, so as above, circle was to be oval. That is not what I expected. High accuracy is very important for our 3d measurement.

    Thanks,
    Joyce
  • Joyce,

    In the case you mentioned above, I would go with the original recommendation you received to create non-distorted circles.

    *Starting from an arbitrary 16:10 aspect ratio image, place your circles where you would like them.
    *In an image editor, create a 912 x 1140 pixel image. The circles will now be ovals.
    *Upload the image to the Lightcrafter. When projected, the ovals will be "re-distorted" back to circles by the diamond pixel configuration.

    -Paul
  • Hi  Paul,

    Thanks again.

    As you said,project image in pattern mode.

    If project image with circles,I need create an 16:10 picture(1280*800) and  get it to 912*1140 in image editor .Then input the lightcrafter and project.That is recoved to circles.

    If line images will be projected,I just need to  creat 912*1140 images and inout the light crafter to projected.

    That  is  right?

    Joyce

  • Hi Paul,
    One more question.
    What is the number of pixels in diamond array in lightcrafer 4500 correspond to a square pixel in an image of 1280*800. Which are these pixels?
    Thanks,
    Joyce
  • Paul,

    I have read question 3 in the FAQ and some questions are still puzzled me:

    *I know that incident illumination is vertical to the hinge axis. and incident illumination direction has any influence on image display?

    *what does greyscale pattern refer to? I don't understand that why it is used, "the “jagged” edges can produce the appearance of an intermediary pixel or “sub-pixel” when viewed or captured through a camera that has greater resolution than the projected pattern. The appearance of these sub-pixels can provide additional resolution in some applications."

    hope to reply
    Joyce
  • paul,

    I have read question 3 in the FAQ and some questions are still puzzled me:

    *I know that incident illumination orientation is vertical to the hinge axis. I wonder that the incident illumination has any influence on image display?

    * What does 'grayscale patterns' refer to?I don't understand why it is used,the “jagged” edges can produce the appearance of an intermediary pixel or “sub-pixel” when viewed or captured through a camera that has greater resolution than the projected pattern. The appearance of these sub-pixels can provide additional resolution in some applications?

    Hope to reply,

    Joyce

  • paul,

    I have read question 3 in the FAQ and some questions are still puzzled me:

    *I know that incident illumination orientation is vertical to the hinge axis. I wonder that the incident illumination has any influence on image display?

    * What does 'grayscale patterns' refer to?I don't understand why it is used,the “jagged” edges can produce the appearance of an intermediary pixel or “sub-pixel” when viewed or captured through a camera that has greater resolution than the projected pattern. The appearance of these sub-pixels can provide additional resolution in some applications?

    Hope to reply,

    Joyce
  • Paul,

    I have read question 3 in the FAQ and some problems are still puzzled me.

    *I have known that incident illumination orientation is vertical to the hinge axis and the diamond array and orthogonal array seem same. But I wonder that incident illumination orientation has any influence on image display?

    *What does the "grayscale patterns" refer to? I just don't understand why it is used, the “jagged” edges can produce the appearance of an intermediary pixel or “sub-pixel” when viewed or captured through a camera that has greater resolution than the projected pattern. The appearance of these sub-pixels can provide additional resolution in some applications. This is an advantage on structure light application using shifted intensity sine waves grey scale patterns?

    Hope to reply
    Joyce
  • Joyce,

    On your first question, I think that you are correct.

    Second question: A 1280 x 800 image will have MORE pixels than a 912 x 1140 image. I can't directly answer your question because it will depend on how the image editor handles the transformation between the two image sizes.

    Paul
  • Paul,

    I think A 1280 x 800 image will have less pixels than a 912 x 1140 image.

    I have read question 3 in the FAQ and some problems are still puzzled me.

    *I have known that incident illumination orientation is vertical to the hinge axis and the diamond array and orthogonal array seem same. But I wonder that incident illumination orientation has any influence on image display?

    *What does the "grayscale patterns" refer to? I just don't understand why it is used, the “jagged” edges can produce the appearance of an intermediary pixel or “sub-pixel” when viewed or captured through a camera that has greater resolution than the projected pattern. The appearance of these sub-pixels can provide additional resolution in some applications. This is an advantage on structure light application using shifted intensity sine waves grey scale patterns?

    For diamond array someone said:"The mirror center-to-center distance in the horizontal direction is twice the distance in the vertical direction. I just need to modify my source image so it accounts for the actual position of each mirror" Is that right?

    Hope to reply
    Joyce
  • Joyce,

    Yes, poor math on my part. 912 x 1140 is more pixels. My main comment is still true though, hard to tell how the mapping will occur - most likely will depend on the image editor.

    The incident illumination orientation has no impact on image display - it just impacts the optical design.

    Grayscale patterns refer to the images being projected without any color or in a single color. I see that you copied the last part of this question from the FAQ. Is there is a question you have about this?

    The mirror center-to-center distance is indeed twice the distance in the vertical direction.

    -Paul
  • Paul,

    About “Grayscale patterns" .

    I still don't clear that "the “jagged” edges can produce the appearance of an intermediary pixel or “sub-pixel” .Please explain it.

    Thanks

    Joyce
  • Hi Paul,
    Sorry to disturb you again.

    I got some information about that diamond array can reduce the number of micro-mirrors to save costs for displaying an image . Yes?
    For 0.45 DMD in the lightcrafter 4500 ,it can display a 1280*800 image, but why it doesn't use orthogonal array with 1280*800,rather than use diamond array with 912*1140. The number of the former is less than the latter. And I don't know what is the advantage of diamond array.

    Thanks very much
    Joyce
  • Joyce,

    The jagged edges comment is primarily a human perception issue. Basically with small pixel sizes and a diamond configuration, the human eye will perceive more detail on the jagged edge than is actually there. This really isn't important for non-display applications and with the Lightcrafter the effect will probably be minimal.

    The primary advantage of the diamond array is that the orientation of the light source allows for a more compact optical design.

    Thanks,
    Paul

  • Paul,

    Sorry to disturb you.

    I know something that  912*2:1140=1280:800=16:10  and  see some word " two frames" somewhere . So I want to know why 912*1140 arrays can display a 16:10 image. what is the hypostasis about the diamond array in lightcrafter 4500.

    And with your recommendation, I carried out some experiment and the result of calibration and 3d measurement experiments were not good. 

    I don't know why? ps. The measurement accuracy of the 3d  measurement system that I hope is 50um. I don't know whether  some wrong with my operation or this  product  is not propitious to my experiment and objective.

    Looking forward to your help.

    Thanks

    Joyce

  • Joyce,

    Not sure what you're asking about in terms of "two frames."

    The 912 x 1140 displays at a 16:10 aspect ratio because of the stretching that occurs on the horizontal axis due to the diamond configuration.

    The 3D scanning demos/TI Designs that we have published probably won't be able to handle 50um resolution due to the optics on the Lightcrafter 4500. If you need high levels of detail, I recommend you get in touch with WinTech. Their 4500PRO board has optics that can handle smaller resolutions.

    www.wintechdigital.com/index.php

    Thanks,
    Paul
  • Hi Paul,

        I noticed some discribe in a paper about DLP DMD chipsets and don't understand very well. (see the underlined part in the figure ).

      Please explain it. Thanks.

    Joyce

  • Joyce,

    The system that is mentioned in the paper you provided sounds like our UHD system. The system uses the speed of the DMD and special optics to use single mirror and make it project multiple pixels. Here is another article that might help you understand.

    This ability is NOT something that the DMDs do natively. It relies on having a complete system design around it to make this possible. We are simply utilizing the fast switching speed of the DMD to make this possible.

    Thanks,

    Paul