Hello All,
This is regarding the previewer white balance coefficients in the case of dm6446.
WB_DGAIN, raw data is multipled by a fixed gain regardless of its color
WBGAIN, raw data is multiplied by a selected gain corresponding to the color of the processed pixel (section 4.3.3.7 of VPFE user guide sprue38e.pdf).
/* prev_config.c setup white balance gain */
params.white_balance_params.wb_dgain = 0x100;
params.white_balance_params.wb_gain[0] = 0x40;
params.white_balance_params.wb_gain[1] = 0x45;
params.white_balance_params.wb_gain[2] = 0x60;
params.white_balance_params.wb_gain[3] = 0x45;
In the above code example, the values assigned wb_dgain and wb_gain are written to the registers (WB_DGAIN, WBGAIN) as it is but how previewer treats them is the question.
How much digital gain does ‘0x100’ corresponds to? Each pixel will be multiplied with 0x100 or some equivalent gain (0x100 = 1.0 eg)
Similarly how 0x40, 0x45, 0x60 are deciphered?
If I have wb coefficients as [1.523 1.000 -1.345], how to assign fractional coefficients to wb registers?
How to select these coefficients optimally?
Please clarify the above queries.
Thanks for your help.
--
Best Regards
S Arun
