In signal processing, discrete cosine transform (DCT), discrete fourier transform (DFT) and fast fourier transform (FFT) these three are important. Sometimes for dct gives clear image patterns over dft or fft. I want to know the reasons behind these. With explanation, if any one can give any reference papers as well, it will be better.
Thanks in Advance
In general, DCT isn't "better" or "worse" than DFT or FFT (which is just a special case of DFT with 2^N components, and can be calculated very efficiently as intermediate results can be re-used).
DCT is suited well for audio and video processing, as it captures the common characteristics of such data very well (see, e.g., this paper for a detailed explanation: https://dl.acm.org/citation.cfm?doid=2030613.2030646). Another advantage is that it works on real numbers while DFT/FFT work on complex numbers. Audio and video data commonly are represented as real numbers (or integers, which are easily mapped to [0, 1]), so using a transformation working on reals is a natural choice.
In general, DCT isn't "better" or "worse" than DFT or FFT (which is just a special case of DFT with 2^N components, and can be calculated very efficiently as intermediate results can be re-used).
DCT is suited well for audio and video processing, as it captures the common characteristics of such data very well (see, e.g., this paper for a detailed explanation: https://dl.acm.org/citation.cfm?doid=2030613.2030646). Another advantage is that it works on real numbers while DFT/FFT work on complex numbers. Audio and video data commonly are represented as real numbers (or integers, which are easily mapped to [0, 1]), so using a transformation working on reals is a natural choice.
- Badges
- Science topic
- Similar topics
- Mathematics