Part Number: PROCESSOR-SDK-TDAX
Hey,
can you help me on setting a second context for single cam usecase, to enable embedded data passing with datatype (0x12)?
Would be great if you could point to where I have to set these changes.
Best regards,
Nico
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Part Number: PROCESSOR-SDK-TDAX
Hey,
can you help me on setting a second context for single cam usecase, to enable embedded data passing with datatype (0x12)?
Would be great if you could point to where I have to set these changes.
Best regards,
Nico
Nico,
You mean, capture metadata having separate data type?
Rgds,
Brijesh
Hey Brijesh,
I use single cam usecase with own camera and capturing works fine, but now I want to also capture the embedded data. I can´t set the embedded data to any other datatype than 0x12, so I read in the original post that you suggest to create two contexts one for frame data and one for embedded data.
So my question is how do I create this second context and where in the psdk do I have to implement the changes?
Best regards,
Nico
Hi Nico,
I dont remember the post.
i dont think you would require additional context for embedded data. could you please help me understand your data flow? is it that first you are trying to capture data using single cam app?
Rgds,
Brijesh
Hey,
if you look at the original post which is referenced here in the header, you know what I am talking about.
I use singelcam usecase and I am capturing framedata of my sensor, but I also want to capture the embedded data. Since I can´t set the embedded data to the same datatype as framedata the csi receiver will ignore embedded data, because it has to be set to 2D(14bit capturing, I implemented the 14bit support myself). Now I read in the original post (
) that in order to do that I need to add a second channel in the capture driver.
So channel 1 is for frame data ----> csi receiver is set to 0x2D
and channel 2 is for embedded data ---> csi receiver is set to 0x12.
Please have a look in the original post, things should be clear then.
Best regards,
Nico
Hi Nico,
ok, understood, i got confused with the 'context' word, as it has different meaning in TDA3x..
This requires some changes in the OpenVX Capture node, Let me get back to you with the changes..
Rgds,
Brijesh
Ah I see, yes the context word can be confusing.
Thank you, looking forward to it!
Hey Brijesh,
I was able to implement it myself!
Thank you for your time, if I have further questions regarding this issue I will come back to you.
Best regards,
Nico
Hi Nico,
Good to know. Could you please share the changes? it would help..
Rgds,
Brijesh
Hey,
here are the changes I did:
--- increase num_capture_frames to 2 in app_single_cam_demo_main.c.
--- define channels for sensor to 2 in imaging sensor source file.
--- in csirx_drvUdma.c add case for desired datatype within CsirxDrv_getIcnt0(const Csirx_ChCfg *chCfg), in my case it is 0x12.
--- in vx_capture_target.c extend the tivxCaptureSetCreateParams function and implement your desired params for Channel 2(incsidatatype, width, height etc.)
Best regards,
Nico
Hey Brijesh,
so the solution I found is not very convenient. Could you still help me on this and explain how you would implement this?
Best regards,
Nico
Hi Nico,
I am sorry, i could not get it. Are you not able to receive embedded data? Is it not working?
Or do you want some different way for handling metadata?
Rgds,
Brijesh
Hey,
so my problem is the following:
I am able to receive the embedded data, but to achieve this I have to increase the num_capture_frames to 2 in the app_single_cam_main.c
and this leads to the allocation of 3 more buffers, which are used for embedded data, but still have the same resolution as is supposed for the normal frames(1820x940), but I only need 1820x4 buffersize for the embedded data, so my solution is a waste of memory.
Any idea to fix this?
What do you mean with a different way for handling metadata?
Best regards,
Nico
Hi Nico,
oh ok, understood.
it is not clean solution. Currently i don't see anyway of specifying image size for all images in raw_image.
This would require some changes in the raw_image or capture node..
Rgds,
Brijesh
Hey Brijesh,
can you ask someone to help me on this, or how are we going on further with this issue?
Best regards,
Nico
Hi Nico,
Do you want the embedded data to be output in sync with the frame data?
Rgds,
Brijesh
Hi Nico,
ok, currently, i don't have any other solution. This will require some changes in the capture node/interface..
I will discuss internally and will add it.
But i guess this thread can be closed, correct?
Rgds,
Brijesh
Hi Nico,
Sure, will let you know.
For the time being, can we please close this thread?
Rgds,
Brijesh