ippiDeinterlaceMotionAdaptive_8u_C1 Example

[Scott_Danahy]

Is there a good example somewhere forippiDeinterlaceMotionAdaptive_8u_C1? I looked in the standard IPP samples and didn't find anything. I'm seeing poor quality (compared toippiDeinterlaceBlend_8u_C1) and want to make sure I'm using the function correctly.

Thank you,

Scott

[Chao_Y_Intel]

Hello,

these two post includes the example and usage for this function? could you help to check it?

http://software.intel.com/en-us/forums/showthread.php?t=66311

http://software.intel.com/en-us/forums/showthread.php?t=78614

Thanks,
Chao

[Scott_Danahy]

Thanks for your reply. I found both of those links already but I was hoping for some complete sample code so I could compare the output and usage to mine.

What doing is using thesefunctions(DeinterlaceBlend andDeinterlaceMotionAdaptive) to perform a 'bob' deinterlace where I convert 29.97 interlaced frames per second to 59.94 byinterpolatingthe missing field. The end of your second link shows how I'm usingDeinterlaceMotionAdaptive:

pSrcPlane[0] = FrameA
pSrcPlane[1] = FrameB
pSrcPlane[2] = FrameC
pSrcPlane[3] = FrameD

ippiDeinterlaceMotionAdaptive_8u_C1(pSrcPlane, ..., topFirst=1, topField=1, ...)
ippiDeinterlaceMotionAdaptive_8u_C1(pSrcPlane, ..., topFirst=1, topField=0, ...)

However, the quality is very poor (lots of stair-steppingand jagged edges) compared toDeinterlaceBlend. Also, the documentation states that "If the processed field is the second field in the frame, the input source planes must be from the previous, current, next, and next after next frames. Otherwise, if the first field is processed, the input source planes must be from the pre-previous, previous, current, and next source frames.".Wouldn'tthis mean that the pSrcPlane array should be different for each call toippiDeinterlaceMotionAdaptive_8u_C1 above since each processed field has adifferenttemporal order?

Thanks for your help,

Scott