I am currently trying to use the OMX deinterlacer component using the DM8168 TI board, but I am not getting the output I expect (bypass data and deinterlaced data look exactly the same).
The input video is OMX_COLOR_FormatYCbYCr video from an NTSC analog camera (from memory, not the OMX capture component). The raw video data being input to the DEI component is 720x480 interlaced, which is actually two 720x240 fields stacked vertically into the frame.
1a) Is this the way that the deinterlacer expects the data to be formatted (i.e. will it interleave lines from the two fields together, and then apply a motion compensation algorithm such as CAVT)?
1b) Or, does the deinterlacer expect fields to be input individually, thus for every two 720x240 fields, one 720x480 frame will be output?
2) Does the deinterlacer actually support OMX_COLOR_FormatYCbYCr on the input and output of the deinterlacer?
3) For the deinterlacer component, is the first output port the bypass video, and the second output port the deinterlaced video? Do both of these two ports have to be connected to a data sink (such as display) in order for the deinterlacer to work correctly?
Thanks,
Greg