Hello,

Has anyone ever tried implementing the IMDCT implementation proposed by Britanak and Rao in the paper, "An Efficient Implementation of the Forward and Inverse MDCT in MPEG Audio Coding"? I tried implementing the 12 point MDCT by following the signal graph shown in figure1 and got it to work. For the IMDCT, we're required to reverse the signal flow graph and do inverse computations. I've been trying to compute the IMDCT and have been struggling in the reverse direction calculations. Does anybody have any experience with this?

Thanks!