Dear Experts, Why waiting for the update of [the hardware sync tests for OAK FFC 4P + 3xAR0234CS](https://discuss.luxonis.com/d/2303-please-share-the-status-of-hw-sync-for-oak-ffc-4p-ar0234-based-stereo-pair/10), I am making another prototype of OAK FFC 4P + dual RPi global shutter Camera IMX296 modules (supposed to be supported [here](https://github.com/luxonis/depthai-core/commit/47055d5270b13adbb740cb1a897b7f9fe7a98aa5)) as we are curious to see the image quality of the IMX296 sensor compared to our existing solution. The setup is as below : [upl-image-preview url=https://discuss.luxonis.com/assets/files/2023-10-20/1697781347-963826-img-0706.jpg] In terms of hardware settings, I would like to know : - Where can I find the schematic of the SL6996 adapter to verify the below cabling ? [upl-image-preview url=https://discuss.luxonis.com/assets/files/2023-10-20/1697781334-497421-img-0704.jpg] - Should I remove/add some components on the camera modules so that the hardware sync (with external triggering) could work? Thanks in advance and best regards, Khang

Hi @"l4es"#389 , Actually, it seems to be an issue with white balance. You can fix it with ``` cam = pipeline.createColorCamera() cam.initialControl.setManualWhiteBalance(4500) ``` With auto WB: [upl-file uuid=261e3906-3c38-4218-9a87-cb94917f8175 size=7MB]autowb.mp4[/upl-file] WB Fixed to 4500K: [upl-file uuid=0a62481b-00fe-4d97-99c8-9b78120d224a size=6MB]wblock.mp4[/upl-file]

Hi @"l4es"#p13443 , Here's the schematics for the adapter board: https://github.com/luxonis/depthai-hardware/pull/461 You'll likely need to solder some wires to connect FSYNC, as adapter is only 15pins, and AFAIK doesn't pass through the FSYNC line.

Hi @"erik"#p13582 , We also tried with other CCTV lenses we have (with C/CS mount required by the RPi IMX296) but the color still seems to be green shifted : [upl-image-preview url=https://discuss.luxonis.com/assets/files/2023-10-25/1698236332-5692-image.png] Could you share the info of the lens that was used with the RPi IMX296 in your setup if any, please ? Best Regards, Khang

Hi @"erik"#p13632, Thank you for the testing. I think we will let this IMX296 as-is for the moment (even-though I was expecting more with its **fully integrated** status in the [list of supported sensors](https://docs.luxonis.com/projects/hardware/en/latest/pages/articles/supported_sensors/)). The OAK FFC AR0234CS is fairly good enough for us. Thanks again, Khang

RPi Global Shutter Camera IMX296 Module

l4es

Dear Experts,

Why waiting for the update of the hardware sync tests for OAK FFC 4P + 3xAR0234CS, I am making another prototype of OAK FFC 4P + dual RPi global shutter Camera IMX296 modules (supposed to be supported here) as we are curious to see the image quality of the IMX296 sensor compared to our existing solution. The setup is as below :

In terms of hardware settings, I would like to know :

Where can I find the schematic of the SL6996 adapter to verify the below cabling ?
Should I remove/add some components on the camera modules so that the hardware sync (with external triggering) could work?

Thanks in advance and best regards,
Khang

erik

Hi l4es ,
Here's the schematics for the adapter board: luxonis/depthai-hardware461
You'll likely need to solder some wires to connect FSYNC, as adapter is only 15pins, and AFAIK doesn't pass through the FSYNC line.

l4es

Hi erik ,

Thanks for the sharing of the schematic. With that, we were able to retrieve some first frames from the OAK FFC 4P :

However, as you could see, the image is not as "vivid" as expected. Would it be the default ISP setting for IMX296 ?
F.Y.I, we are using the following lens in our setup : https://www.tamron.com/global/biz/products/ipcctv/13fm28ir.html

You'll likely need to solder some wires to connect FSYNC, as adapter is only 15pins, and AFAIK doesn't pass through the FSYNC line.

I will connect the XTR on the back of the camera module to the FSIN pad on the FFC cable.

Best Regards,
Khang

erik

Hi l4es ,
Artifacts as on the img below look like lens chromatic aberration (LCA) issue, which is highly dependent on the lens used. ISP is able to do some compensation for it, but would need custom camera tuning for proper image. Have you tried with other lenses as well?

l4es

Hi erik ,
We also tried with other CCTV lenses we have (with C/CS mount required by the RPi IMX296) but the color still seems to be green shifted :

Could you share the info of the lens that was used with the RPi IMX296 in your setup if any, please ?

Best Regards,
Khang

erik

Hi @l4es ,
We tried it with Arducam's IMX296 module, and besides the bad low-light performance (artifacts due to likely voltage issue - bad design of the module), it does seem it's a bit more greenish, which suggests camera tuning imperfections. We might look into it more, but since IMX296 isn't part of our standard cameras, it doesn't (financially) make sense to do sensor tuning if that is required.

erik

Hi @l4es ,
Actually, it seems to be an issue with white balance. You can fix it with

cam = pipeline.createColorCamera()
cam.initialControl.setManualWhiteBalance(4500)

With auto WB:

autowb.mp4

7MB

WB Fixed to 4500K:

wblock.mp4

6MB

l4es

Hi erik,

Thank you for the testing. I think we will let this IMX296 as-is for the moment (even-though I was expecting more with its fully integrated status in the list of supported sensors). The OAK FFC AR0234CS is fairly good enough for us.

Thanks again,
Khang