To: Ausdauer who wrote (4130 ) 12/23/1998 2:28:00 PM From: Walter Morton Respond to of 60323
Sorry to muddy up the thread with this, but you guys seem to be very knowledgeable technically and may be willing to provide a non-bias opinion on this technology: _________ "I will attempt to itemize the important features, strengths and weaknesses of our compression and how it may differ from others you may be looking at. 1) Compression Ratios We developed the compression routine for the fastest possible access to data. To accomplish this, we sacrifice somewhat on high compression ratio's. Typically, we compress images by a factor of 10 to 20 times. Here are some typical compression results. Image Compression Ratio _ Loss 256 Gray Scale 10:1 _ 3% 24-Bit Color Image 10:1 _ 2% 256 Gray Scale 18:1 _ 6% 24-Bit Color Image 18:1 _ 4% 2) Decompression Speed When it comes to data decompression speeds, we are up to a hundred times faster than other products we have tested. For example A 3.5 GB mosaic of 15 gray scale images, compressed to 300 MB can have any 2.0 GB portion of this mosaic accessed in less than 0.5 seconds. Once zoomed into to under 500 MB, updating the screen is done in less than 0.1 sec (Testing was done on a very average DELL 233 MHZ with 64 MB of RAM, the data was stored locally). On this example, 2 GB of data represents an image of about 45000 x 45000 pixels!!!! Updating that much imagery in under half a second is almost real time decompression and display. These times can not be duplicated by any other decompression software I know. 3) Tiling Images We combine tiled images when viewing, not when compressing. We compress images and store them individually, then piece them together on-the-fly. Other techniques combine all images in a set into one much larger compressed file. There are advantages and disadvantages to both approaches. One larger file is certainly easier to deal with but also limits how the data is maintained, viewed and distributed. One other unique feature is the we can tile many types of images together. Color 10m satellite imagery can be tiled with gray scale 1.0m imagery where available, and the user has complete control over how this is done. Our approach lends itself well to organizations with dynamic datasets, if access to portions of the data changes from user to user, if the user wants control over how the images are pieced together, or if fast display is important. 4) Custom Applications and compete solutions For example, we customized our decompression to rotate the underlying map image so that the direction of travel is always to the top of the screen. Again this is done real-time. Try rotating an image representing 45,000 x 45,000 pixels real time, this is no trivial exercise. Soon to come: Scanned in USGS map sheets. As you zoom in, the demo will automatically replace the 1:250,000 scale sheets with the more detailed 1:100,000 sheets, and eventually the 1:24,000 sheets." I've deleted the company's name so I won't be accused of trying to promote another company on this thread.