Crazy Homeless Guy Posted March 20, 2005 Share Posted March 20, 2005 Based on a thread from last week, I decided to test jpeg compression results from different apps. http://www.cgarchitect.com/vb/showthread.php?t=9582 I am posting the original file, as well as the images results from them, and their stats. If you use a different app (ACDsee was mentioned), please run the test, and post your results. The initial test. I grabbed an image I took this morning using my canon a80. This is probably not the ideal camera to use for this test, but it is the one I own. The problem with using this camera is that it uses jpeg compression on the photos it takes, not Tiff or RAW like some cameras offer. I had the settings as high as they would go when I took the photos, so their is little compression. I opened the image, and cropped out a square to use as the test area, and then saved it as a tiff file. This will prevent further compression on the image before I run my test. since I am comparing compression results out of different apps, the fact that this image was originally captured as a jpeg shouldn't matter. We start with the same base in each program. I am only because it may be an issue to some. At first my plan was to set the jpeg compression to 50 in each app. I learned two things from this, and decided to change my mind, and run the test differently. The first thing I learned from my initial setup was that the different apps use vastly different algorithms to compress the images. Setting the compression to 50 in one app produces the same results as setting the compression to 10 in another. The second thing I learned was not as important, but only that 50 in any of the apps produced little noticeable compression. for the test results to work best, we need to see visually see breakdown in the image. New test method. I am picking two target file sizes, and making each app produce a file close in size, +/- 1k. No blurring or other special jpeg compression techniques were used. The test target sizes in kilobytes: 26k and 48k The test sizes were chosen to be those sizes based on results from ImageReady. It turns out that Fireworks offers more control the actual file size of the compressed image. Setting the jpeg compression to '0' in ImageReady resulted in a file that was 26k in size. So using this image, that is the smallest jpeg file size I could produce from ImageReady. If I set Fireworks to '0', I could produce an image that was just under 9k. Setting the compression to 100 in both apps produced a 323k file size in ImageReady and a 387k image size in Fireworks. In my opinion this gives Fireworks an advantage because it is able to produce a far wider range of compression quality than the Adobe products can. The apps I am testing in this post are: Macromedia Fireworks Adobe ImageReady It should be noted that I found out that ImageReady and Photoshop 'save for web' produce identical file sizes to the one one hundredth of a kilobyte. We can safely assume that they use the same algorithm because they are both Adobe products. Original Photo - Tiff Format (756k) Macromedia Fireworks 26k (actual size 25.31 w a 36 compression setting) Macromedia Fireworks 48k (actual size 47.54 w a 56 compression setting) Adobe ImageReady 26k (actual size 25.84k w a 0 compression setting) Adobe ImageReady 48k (actual size 48.25k w a 20 compression setting) The best way to compare the results is to download the images, and place them on top of each other in you image editing software. The turn on and off the layers so you can see the differences. Also, zoom in close on the image, and turn on and off layers to see how the compression effects each image. I am attaching each of the sets of images in Photoshop 7 format to compare. Photoshop 7 26k (3.8 megs) Photoshop 7 48k (4 megs) . Link to comment Share on other sites More sharing options...
Crazy Homeless Guy Posted March 20, 2005 Author Share Posted March 20, 2005 Opinion.... Fireworks jpeg algorithm produced a more accurate color when compressed, but blurred the image more than ImageReady did. Fireworks blurring technique removes fine detail from the image to allow it to chunk more of the image into one color than if the fine details were still there. The blurring also helped reduce the noticeably of the blocks that are prevalent in highly compressed jpeg images. ImageReady produced a sharper image, but at the cost of color accuracy. The sharper image played well for keeping contrast, and detail, but took away from the extra kilobytes available for storing color information. Also, the chunks of space were more noticeable in the ImageReady version because the algorithm did not blur the image. I don't know a lot about image interpretation, and whether the mind would respond better to color or sharpness. I think it probably varies a great deal according to what the content of you image is that you are compressing, and personal preference. Bottom line... at medium compression the images are close, and the program did not come into play as much. At high compression I would probably have to go with Fireworks because it gives you more control, and applies a more sophisticated compression techniques to the image (imho). But like I said, I didn't play with the blurring/smoothing option in the compression dialog, and they can greatly effect how an image looks, and what the final file size is. I don't think most people adjust that setting though, just the file quality setting. Not sure if it was worth the effort, but it gave a little view to the differences between the 2 applications. Link to comment Share on other sites More sharing options...
Nasheet Siddiqui Posted March 20, 2005 Share Posted March 20, 2005 superb......I appreciate your logical deduction method...great.. Now....I think JP2 has taken things to more extreme.. I have tested images with jp2 and it gives you half the file size with more or less same quality. problem why still jp2 is not the industry standard I dunno attaching two sample jpg files and then their jp2 equivalent.... Since I cannot upload jp2 files I will upload them as *a.jpg files ,so just right click and save them as *a.jp2 and view them after installing plugin for your browser... I have a 1000 X something pixel file in under 50 kb file size !!!!!! quality is at the border line between visible distortion and approved minimum quality.....relative to my eyes of course one minus point with jp2...no thumbnails preview..but that's just few lines of programming and industry approval away..... still finding a web-site which directly uploads jp2 files.....webshots refuse...will try pixelperfect...... Link to comment Share on other sites More sharing options...
Nasheet Siddiqui Posted March 20, 2005 Share Posted March 20, 2005 Finally I found it imgzhost.com .....so far so good http://www.imgzhost.com/t.php?img=47b8944782.jp2 http://www.imgzhost.com/t.php?img=af87f2a653.jp2 I am so sorry its not working.... Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now