We have been successfully able to capture from all the MIPI CSI ports on Tegra X1 but now we are interested on capturing also the metadata and not only the video that is being injected in the MIPI CSI Ports. I read in the Technical Reference Manual about potential support for long packets (page 2399):
“The exact use of embedded data is not clear, and therefore in most cases, embedded data is probably not used, or if sent by the transmitter, the embedded data packets can simply be discarded. If for some reason, it is important to receive the embedded data, there are other options that should be supported. Since there are two pixel parsers, if there is only one video source, it is best to direct embedded data to the other pixel parser which does not process the normal image data. In this way, the embedded data can be treated the same way as 8-bit arbitrary/compressed image data and can be output in 8-bit/clock or 16-bit/clock format. If there are two video sources that occupy both the pixel parsers, then embedded data if it has to be stored to memory, needs to be output in the same channel as the image data as 8-bit/16-bit data. But this also means that the module that receives output of CSI module and does image processing must have the ability to differentiate embedded data and not do image processing on this data prior to writing it to memory”
Where can I get information about how to configure the pixel parser to get the metadata? I don’t need support just some guidance of which register I would need to configure to get the metadata. Have someone tried it?