2014-01-28 19:47:15 -0600 | asked a question | DFT_ROWS flag Hello, I am trying to take the DFT of an image, multiply it with a filter, and then take it's inverse DFT. I have got the DFT-IDFT part working (That is, I take the DFT of an image, then take it's IDFT, and obtain the original image). The following code works: // In is the Input Image. dft(In,In,DFT_ROWS); dft(In,In,DFT_ROWS|DFT_INVERSE); However, when I introduce the filter, I get garbage images: dft(In,In,DFT_ROWS); multiply(In,filter,In); dft(In,In,DFT_ROWS|DFT_INVERSE); However, when I use the DFT_REAL_OUTPUT flag instead of the DFT_ROWS flag in the above code(nothing else is changed), the output images are correct. I am using DFT_ROWS, as it is almost 50% faster. Can anyone give me some insight into this problem? Best Regards, |