TDA3: Using GLBCE in 12-bit linear mode on TDA3x.

Tom Korah Manalody

Part Number: TDA3

Hi,

On tda3x vision_sdk2_10 the operating mode is ISSM2MISP_LINK_OPMODE_12BIT_LINEAR, with ov10640 sensor. Is it possible to use GLBCE properly in this mode of operation?

If possible how could we enable it properly? Where/how could we set the parameters like glbce strength(irStrength)?

Thank You

over 8 years ago

0 Brijesh Jadav over 8 years ago

TI__Guru**** 475745 points

Hi Tom,

Well, it can be used, but it is not supported as of now.

Why do you want to use GLBCE in linear mode?

Rgds,

Brijesh

0 Alex Bashkov over 8 years ago

TI__Genius 13425 points

Hi Tom Korah Manalody ,

If you need to write your custom library then, GLBCE is capable of working with 12bit linear data. Please refer to TDA3x ISS addendum (SPRUHV2). The global parameters and strength parameters are explained in "Table 631. ISS ISP GLBCE Core Key Parameters". You can also refer to "3.3.2.4 ISS ISP GLBCE Programming Model" for hints how to enable the module properly.

Thanks,
Alex

0 Tom Korah Manalody over 8 years ago in reply to Brijesh Jadav

Prodigy 210 points

Hi Brijesh Jadav,

We are trying to use GLBCE for improving low light performance.

We managed to see some changes when GLBCE is enabled for 12-bit linear, not sure its correct or not,But For monochroma path and 16bit WDR mode, why it is required to shift up the input before GLBCE and shift down by 4 bits after GLBCE, is this required for 12 -bit linear also?

0 Tom Korah Manalody over 8 years ago in reply to Alex Bashkov

Prodigy 210 points

Hi Alex Bashkov ,
Thank you for your references.

0 Brijesh Jadav over 8 years ago in reply to Tom Korah Manalody

TI__Guru**** 475745 points

Hi Tom,

Yes, we need to shift up by 4 bits before GLBCE and shift down by 4 bits after GLBCE.

Regards,
Brijesh

0 Tom Korah Manalody over 8 years ago in reply to Brijesh Jadav

Prodigy 210 points

Thank you Brijesh Jadav for your replay.

Can you tell why the we are shifting the input value by 4 bits? We are doing the same for 16bit wdr also.

0 Alex Bashkov over 8 years ago in reply to Tom Korah Manalody

TI__Genius 13425 points

If I may hop in again,

The 16bit wdr 4 bit shift is called WDR Splitting, this is to support image processing of more than 12bit data, IPIPEIF can generate pseudo long exposure image and short exposure image from WDR input. For this purpose, IPIPEIF has saturation function, and down-shift function at the input. To generate pseudo-short-exposure image, IPIPEIF downshift the input image by a certain bits. If the input is 16-bit image, it is down-shifted by 4bits to generate 12-bit “short exposure” image.

Thanks,
Alex

0 Brijesh Jadav over 8 years ago in reply to Alex Bashkov

TI__Guru**** 475745 points

Hi Alex,

WDR splitting is used for generating pseudo short and long exposure images from the PWL merged images. This is done in IPIPEIF. But GLBCE can be used even for single pass flow.

Essentially GLBCE works only on 16bit input. For the two wdr flow, the output after merge can be upto 20bits, this 20bit pixel will be compaded to 16bit using companding module in IPIPEIF and provided to GLBCE. GLBCE output is again 16bit, where lower bits are not valid, so we remove lower 4bits by downshifting.

In Linear Mode, IPIPEIF merge block will not be used, so 12bit input pixel needs to be shifted up by pixels before feeding it to GLBCE. and then after GLBCE it should be shifted down.

Tom,

Shift up and down operation is done ISP_SYS3 module and it is taken care in the ISP driver.

Regards,

Brijesh

0 Alex Bashkov over 8 years ago in reply to Brijesh Jadav

TI__Genius 13425 points

Hi Brijesh Jadav ,

Got ya, thanks for explaining further.

Hi Tom Korah Manalody ,

Does that answer your questions? Could you please verify the answer?

Thanks,
Alex

0 Tom Korah Manalody over 8 years ago in reply to Alex Bashkov

Prodigy 210 points

Brijesh Jadav , Alex Bashkov .

Thank you for your replay and support.

Processors

Processors forum

TDA3: Using GLBCE in 12-bit linear mode on TDA3x.