|
JPEG
vs. Wavelet
Looking for wonders? Be prepared to meet the reality. It looks like new compression standards will not bring a lot of positive changes to our lives. Everyone knows that high quality images require many megabytes of memory. Because of that, there is always the need for efficient compression algorithms. Some news reports and advertisements give us hope that in the near future masterpieces of art will be reduced to 100 bytes. Alas, this will never happen. All up-to-date compression methods try to reduce redundancy that exists in standard representations of graphic data. And we must admit the existing procedures have already succeeded in doing their job. It is unlikely that we will see a breakthrough in this area. Still, at a rough estimate I'd say that volumes of graphic information may decrease by up to 50 per cent. But they will never shrink by several times.
I believe the image is quite suitable for the tests, because it contains uniformly painted areas, details, textures, and regular structures. The results
of my tests can be seen on the second picture. The upper row represents
the conventional JPEG. The images in the middle row were obtained with
the help of the LuraWave.jp2 compression technique. According to the
The leftmost column shows the results of the 30 times compression that gave files of about 3 Kbytes. At such a high compression, various artifacts can be seen on every image. (Mention should be made that five-kilobyte images do not have any visible distortions and look identical to the initial image.) Please note that although LuraWave produced better edges, it significantly blurred the texture of the shirt. The color of the hand also changed. The middle column contains pictures that were compressed by 40 times to fit the two-kilobyte files. You can analyze them yourself. The rightmost column demonstrates what we can get after the picture is compressed by 80 times (1 Kbyte files). In this case, WAVELET outperformed other algorithms. It still gives a blurred but tolerable picture, while JPEG images start to collapse. Now let us draw some conclusions. WAVELET does manage to produce a tolerable picture at high rates of compression. But, alas, in many cases it is of no use to us. A web designer will never choose even a 30-times compressed picture to place it on the site. He is likely to stick with traditional approaches and compress his images by 10 times. Both JPEG and WAVELET result in the same quality at such a compression level. *** LuraWave®
is a registered trademark of
|
|
|
Igor Yefremov, 2002, all rights reserved You
must obtain a written permission from me to use any materials of this
site for any commercial or non-commercial purposes, unless there is an
explicit statement to the contrary. |