Fractal compression's main competitor is the famil
Post# of 1084
Posted On: 05/15/2014 12:03:07 PM
Fractal compression's main competitor is the familiar JPEG standard. This is a compression procedure that makes use of a modified form of Fourier analysis. It also divides up the image into small blocks 8x8 pixels in this case. However rather than try to represent each block as an affine transformation of another (larger) block it uses a set of 2D sine waves.
JPEG does produce good compression results but it is claimed that it is inferior to fractal compression methods. One of the reasons is that fractal methods are not scale dependent. You can decompress a fractal image to any resolution you like by choosing the size of the image that you iterate. Of course beyond a certain resolution the extra detail you will see will be artificial but the argument is that because of its fractal nature it will not be too "obvious" and the image will still look natural.
A second advantage is that fractal compression doesn't share a defect that is inherent in the JPEG and all Fourier methods. The problem is that where there is a discontinuity in the image, i.e. a step from black to white, the sine functions which are used to approximate it simply cannot do a good job.
Why?
Well sine functions are all smooth and curvy and a discontinuity isn't and so trying to fit continuous functions to a discontinuity is like backing a looser!
As long as you don't push JPEG to too high a compression ratio you will not notice the problem but if you do then what you will see is a blurring of the edge and a pattern that looks like an echo of the change. This is the classical Gibb's overshoot or "ringing" effect and there isn't much you can do about it. Of course fractal compression has no such problem representing a discontinuity.
Even so, today Jpeg compression is almost universally used and many of the early claims for the powers of fractal compression made in the 1990s never really came to be. fractal compression takes longer to implement and in most cases the gain in compression ratio over Jpeg isn't significant.
Also the claimed "better quality at higher resolution" also isn't really an issue. Today fractal compression is still being worked on but at the moment its a technology in search of a problem.
WELL GUESS WHAT.......THE PROBLEM HAS BEEN FINALLY FOUND!!!! AND NOW IT IS BEING SOLVED BY TRUDEF AND A WONDERFULLY INNOVATIVE COMPANY CALLED TMMI!!!
Mathemtically the 4X data requirements of the Quad HD/4K screen resolutions presents a serious problem for broadcasters and network providers using current video compression methods. TRUDEF™ Fractal Scaling will allow HDTV broadcasters to offer high quality video for the next generation HDTV resolutions with little or no increase in bandwidth usage.
Upscaling is also valuable for transforming archived content originally shot in lower resolution formats to modern HD formats while maintaining as much of the original detail as possible.
TRUDEF Sterling
tictocktictock
JPEG does produce good compression results but it is claimed that it is inferior to fractal compression methods. One of the reasons is that fractal methods are not scale dependent. You can decompress a fractal image to any resolution you like by choosing the size of the image that you iterate. Of course beyond a certain resolution the extra detail you will see will be artificial but the argument is that because of its fractal nature it will not be too "obvious" and the image will still look natural.
A second advantage is that fractal compression doesn't share a defect that is inherent in the JPEG and all Fourier methods. The problem is that where there is a discontinuity in the image, i.e. a step from black to white, the sine functions which are used to approximate it simply cannot do a good job.
Why?
Well sine functions are all smooth and curvy and a discontinuity isn't and so trying to fit continuous functions to a discontinuity is like backing a looser!
As long as you don't push JPEG to too high a compression ratio you will not notice the problem but if you do then what you will see is a blurring of the edge and a pattern that looks like an echo of the change. This is the classical Gibb's overshoot or "ringing" effect and there isn't much you can do about it. Of course fractal compression has no such problem representing a discontinuity.
Even so, today Jpeg compression is almost universally used and many of the early claims for the powers of fractal compression made in the 1990s never really came to be. fractal compression takes longer to implement and in most cases the gain in compression ratio over Jpeg isn't significant.
Also the claimed "better quality at higher resolution" also isn't really an issue. Today fractal compression is still being worked on but at the moment its a technology in search of a problem.
WELL GUESS WHAT.......THE PROBLEM HAS BEEN FINALLY FOUND!!!! AND NOW IT IS BEING SOLVED BY TRUDEF AND A WONDERFULLY INNOVATIVE COMPANY CALLED TMMI!!!
Mathemtically the 4X data requirements of the Quad HD/4K screen resolutions presents a serious problem for broadcasters and network providers using current video compression methods. TRUDEF™ Fractal Scaling will allow HDTV broadcasters to offer high quality video for the next generation HDTV resolutions with little or no increase in bandwidth usage.
Upscaling is also valuable for transforming archived content originally shot in lower resolution formats to modern HD formats while maintaining as much of the original detail as possible.
TRUDEF Sterling
tictocktictock
(0)
(0)T M M Inc (TMMI) Stock Research Links
TRUDEF Sterling
tictocktictock
tictocktictock
Scroll down for more posts ▼