AM62A7-Q1: How to generate cac LUT files

Hi yangrongbo,

CAC can undo shifts on R and B pixels relative to G pixels.
You may follow the FAQ below for generating CAC tuning with DCC tuning tool.
https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1283047/faq-tda4vh-am62a-how-to-create-the-cac-mesh-lut-for-dcc-tuning-tool-input

It looks like you have some data on relative radial shifts between certain wavelength.
For example, 470nm to 550nm shifts may be close to R-G shifts, but won't be exact and may change with lighting.
We don't have much experience about how to use these measurement data and you will have to experiment on that.

HI Gang Hua:

 There are still issues after CAC. Also, I don't think CAC can solve this problem. I tested ISX031 (with the same lens model and sensor), and there was no color difference in the middle area and no CAC was done. As shown in the figure:

Hi Yang,

I see color difference in the raw you emailed. I don't think it is an issue can be solved from ISP. It is better to contact Sony on this issue.

image_v17-35.raw

Regards,

Adam

Hi yangrongbo,

yangrongbo said:

There are still issues after CAC. Also, I don't think CAC can solve this problem.

As we discussed on another thread, there are 2 possible causes for this kind of artifacts.

(1) CA if you the issue does not appear on image center, but around image corner.

Do you have the sample raw/yuv images for both?

(2) WDR tone mapping around bright lights.

This one we would have to check if it is a problem in sensor WDR merge or ISP WDR tone mapping.

Adam Hua said:

image_v17-35.raw

Do you have the WDR companding knee points for this file?