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.

TDA4VM: After being processed by CFA module, the dynamic range of the image is reduced. How to solve it?

Part Number: TDA4VM

24bit -> 12bit , it will result in the loss of half of the dynamic range of the image. We need help to reach the dynamic range of 120dB.

  • Dynamic range compression is done by tone mapping (GLBCE).

    Please refer to the 120dB IMX390 sensor settings available in PSDK.

  • we found that the IMX390 sensor GLBCE settings and ours were the same after comparing the two settings. 24bit->12bit in the CFA plugin means 12bit data loss of the picture, is that ture? I think this operation will lead to the reduction of dynamic range.

  • CFAI input is the GLBCE output.

    I am not sure which sensor you use and I suppose your sensor output raw image is companded to 12-bit before sending to TDA4.
    GLBCE has a local tone mapping algorithm which compresses the dynamic range of its input image to roughly 12-bit.

    IMX390 works in 20-bit dynamic range mode while yours is in 24-bit.
    Careful tuning of ISP and AE must be done for 24-bit sensors.

  • The sensor is OX08B, and 24-bit raw pictures are conpanded to 12-bit before sending to TDA4.

    There are two questions:

    1. CFA input is 16-bit in the VISS dataflow, but 4th description says GLBCE output is 12-bit. 

    2. How to use the gamma value in the CFA plugin? 

  • 1.

    CFAI is 12-bit in VPAC1 and 16-bit in VPAC3.
    GLBCE output is physically 16-bit, but its intention is to compress dynamic range to the 12MSBs.
    We typically use the LUT at CFAI input to re-linearize color by inverting the gamma used for compressing 20b or 24b to 16b for VISS pipeline.

    2.

    Please refer to the plugin guides for more details.