• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 99 subscribers
  • Views 511 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.

TDA4VM: RAWFE non-linear LUTs tuning

Zs lei
Intellectual 660 points
Part Number: TDA4VM

Hi TI experts,

I have some questions ragarding RAWFE non-linear LUTs tuning.

1. LUTs config
There are 2 LUTs within RAWFE. 1 is in Decompanding block and another is in Merge block. For our project, WDR merging is done in sensor part. So how to config 2 LUTs in RAWFE? For my understanding, the LUT in Decompanding block is applied and another in Merge block is bypassed.
Where is the config procedure in source code?
2. LUT locations for 20bit input
According to TRM, both LUTs have same concept that supports 3 zones.
For 20bit input,  range 0-4K ->step size is 32.
                          range 4K-2^20 -> step size is 2K.
So the LUT has 4K/32+(2^20-4K)/2K=128+510=638 locations.
Please help check my understanding and correct me if I was wrong.
3. LUT content
For our project, LUT is that look up 20bit input to 16bit output.
Based on locations result in 2, how to define the value of each LUT location? It means to define a LUT with 638 locations to convert pixel 20bit to 16bit. Has it defined already by TI? If yes, could you help indicate where it is? It can be tuned by customer?
Best regards,
Zhishen
  • 0
    Intellectual 660 points

    Hi TI experts,

    No feedback for a week. Could you help check it?

    Thanks.

    Best regards,

    Zhishen

  • 0 in reply to Zs lei
    TI__Guru**** 477695 points

    Hi Zhishen,

    Sorry for the late reply, please find answers to your questions below. 

    #1, Typically the configuration will be part of the DCC profile that internally viss node parses and uses driver to apply it to registers. They are not directly exported to application from the viss node. 

    #2, yes it is corerct.

    I will check with experts on the next questions.

    Regards,

    Brijesh

  • 0 in reply to Brijesh Jadav
    Intellectual 660 points

    Hello Brijesh,

    Thanks for your reply.

    I'm looking forward your reply for question3.

    Thanks.

    Best regards,

    Zhishen

  • 0 in reply to Zs lei
    TI__Guru 59500 points
    >> 3. LUT content
    >> For our project, LUT is that look up 20bit input to 16bit output.
    >> Based on locations result in 2, how to define the value of each LUT location? It means to define a LUT with 638 locations to convert pixel 20bit to 16bit. Has it >> defined already by TI? If yes, could you help indicate where it is? It can be tuned by customer?
    In general, you are correct.
    The LUT, as for all supported sensors in PSDK, combines (1) decompanding sensor raw image to linear 20-bit, (2) removing sensor black level, (3) compressing linear 20-bit to 16-bit with a gamma of 0.7.
    This configuration is located at "imaging/sensor_drv/src/ar0233/dcc_xmls/wdr/AR0233_rawfe_decompand.xml" for AR0233 in PSDK.
    If you have a new sensor, you will need to create all the sensor driver similarly.
    To generate the xml file and the LUT, you may use TI's DCC tuning tool (please contact your local FAE and sales if you don't have it yet).
  • 0 in reply to Gang Hua
    Intellectual 660 points

    Hi Gang Hua,

    Thanks for your reply.

    For our project, WDR fusion is done within sensor. How to modify code to bypass LUT in Merge block?

    Best regards,

    Zhishen

  • 0 in reply to Zs lei
    TI__Guru 59500 points

    Hi ZS,

    It is bypassed by default in PSDK s/w (all SDK supported sensors are in that mode).
    You don't need to do anything.