MPEG encoding and video compression process

Thread Starter


Joined Nov 13, 2014
I need some help understanding the MPEG video compression process, particularly the motion estimation and motion compensation block. I know that DCT converts the image from time to frequency domain. Then the image is quantized and coded using Huffman Coding (VLC). Can anyone please explain in detail what motion estimation and compensation does and why are they positioned in such a manner according to the image :-

Also, what is the frame memory for? I'd appreciate it if someone can walk me through the steps of MPEG Compression. And by steps I mean baby steps. For example, "first the image goes thru DCT, then blablabla" . Thanks!