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.

Urgent: Image Analyzer error in 3.3

Anonymous
Anonymous

Hi,

 

I would like to ask some questions on CCS 3.3’s Image Viewer.

 

The input is 640 × 480 4:2:2 BT.656 stream, and the starting address for their location in DDR2 memory is 0x81000000. The order is YUYV.

                    

                      

An image frame which can be correctly displayed in CCS 4.1.2 cannot be properly displayed in CCS 3.3. In the first figure above, the right smaller image is an image sensor’s test pattern showing vertical color bars. In CCS 4.1.2, using the settings in “Properties”, although I made a mistake in that U and V mask should actually be exchanged(U before V, not the reverse), the image can basically get correctly displayed. The dimensions of the display is 640 × 480, same as the input stream.

                     

                     

In 3.3 however, I have tried different configurations, and the closest one to the input is when “Index Increment” = 4, and has the bottom half looks like having been contaminated. In addition to this, each of the eight color bars (starting from white in the left) appear as two bars with slightly different shades.

 

When “Index Increment” = 8, the result is like a scaled version of “Index Increment” = 4, replicated in the upper two quadrants. The lower half contains unexplainable color mixes.

 

When “Index Increment” = 2 and 1, no part of the display has a uniform color.

 

I am confused. According to 8-bit BT.656 YUV standard, each pixel in the image would be represented by two pixels, so I expected “Index Increment” = 2, however the experiment result differs largely from this.

 

Whyt there is no  get correct image here in CCS 3.3? Could anyone instruct me on how to configure? Although I have CCS 4 installed, it is very unstable with SEED XDS560 emulator, so I have to switch back and use 3.3 for my project.

 

 

 

 

Thanks,

Zheng

  • Perhaps the data is expected to be stored differently in the two different views?

    In the first, you specify packed bytes and the placement of Y and U and V within that sequence, with stride and masks.

    In the second, you specify three different start addresses. Different start addresses implies to me that the display tool expects separate Y buffer, separate U buffer, and separate V buffer.

    It might be worth a try to see if that makes the difference. There also may be other display options that would more closely match what you want to display.

    Regards,
    RandyP

  • Anonymous
    0 Anonymous in reply to RandyP

    Randy,

    RandyP said:

    In the second, you specify three different start addresses. Different start addresses implies to me that the display tool expects separate Y buffer, separate U buffer, and separate V buffer.

    I searched on the web for information about displaying YUV image in CCS 3.3's image viewer, and there are several of them showing setting "Start Y""Start U""Start V" to the same address. However, this change on my machine didn't solve the problem.

    I am now using CCS 4.2.4 instead with updated emulator from SEED.

     

    Zheng