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: TDA4VM WDR MERGE

Part Number: TDA4VM

Hi, Ti experts

  We are facing a wdr requirement that needs TDA4 to do the 3-exposure wdr merge(merge 3 raw bayer arrays with different exposure into one high-bit raw).I have found some instructions that indicate the ability of wdr merging in TDA4, but ask around found no instants or application of using TDA4 to do wdr merge.I want to figure out whether it is possible to do wdr merge in TDA4, and if it is possible,I want to get the detailed instruction about wdr merging int TDA4.

  • Hi,

    In general, we don't recommend using the WDR merge function of TDA4 VISS because merging in sense is typically much better (motion artifacts, color artifacts, etc.).

    merge 3 raw bayer arrays with different exposure into one high-bit raw

    VISS H/W supports 3-frame merging at the beginning of VISS, but output of VISS is always low-bit-depth YUV after tone mapping.
    We won't be able to get high-bit raw from VISS H/W.

    You may find WDR merge and VISS details in TDA4VM TRM.

  • Hi,

    I have read through the wdr merge related part of TDA4VM TRM and it seems that  wdr merge function can only be used in a raw stage such like distributing  registers.Can we get an example in c code form to bring up the merge function more conveniently.

  • Hi yan zhiwei,

    it seems that  wdr merge function can only be used in a raw stage such like distributing  registers

    I am not sure if I understand you correctly.

    What do you need to do on your sensor input in 3 exposure frame format?
    Is it for converting it to local tone mapping 8-bit YUV output image?

  • We typically input 12bit companded wdr raw into cfa module and do the rest ISP when sensor supports mutiple frame WDR merge,but now the sensor we are required to use do not support this companding function,so we are seeking for alternative ways in VISS to fullfill WDR merge and we want a simple demo in C code or any other form.

  • The VISS actually supports WDR merge by calling tivxVpacVissCreate(by remote service, in vx_vpac_viss_target.c) ->tivxVpacVissSetDefaultParams ->tivxVpacVissDefaultMapRfeParams ->tivxVpacVissDefaultMapMergeParams. So the current demo can directly support WDR merge but the problem lies in the sensor driver. You should modify the driver for multiple frame with different exposure settings, manual exposure control and lots of other driver functions. 

    If possible, please use WDR merge in sensor or just the single linear image when the dynamic range is sufficient.