Tool/software: TI-RTOS
Hi,
I'm trying to understand the H3A data during 1pass and 2-pass WDR mode in OV10640 MIPI. As per my understanding:
In two pass mode:
The companded data from the sensor is decompanded.
Split into long and short exposures.
Process the long exposure frame(DPC, NSF3 and generate H3A stats) and store the RAW Bayer frame in DDR.
Process the short exposure frame(DPC,NSF3 and generate H3A stats).
Now load the long exposure Bayer frame from memory and short exposure frame and process with WDR merge to get the WDR output.
(I assume output bitwidth of WDR is always 20. If not,What decides the output bitwidth of WDR? )
Then the 20bit WDR is companded to 16 bit and sent to GLBCE.
GLBCE does the tone mappping and the outputis of 16 bit. Since only 12 bits are valid we downshift and pass it through the IPIPE for further processing.
Since both long and short exposure are processed and passed through H3A separately, does it mean stats we receive in the AE plugin(issAewb1Link_algPlugin.c) alternate between the long and short exposure frames?
In one pass mode:
The companded image from the sensor is decompanded to 20 bits and globally tone mapped down to 12it as per this link. Who is responsible for the global tone mapping.(GLBCE?). Since the data is 20 bits after decompanding, is it companded again in IPIPEIF to 16 bit before passing to the GLBCE mode?
Is the H3A data is calculated on the companded and tone mapped data?
Regards,
Navinprashath.R.R
