Thread View: comp.graphics.research
2 messages
2 total messages
Started by Anthony Yen
Thu, 24 Dec 1992 17:58
Scaling Down 24-bit Pix
Author: Anthony Yen
Date: Thu, 24 Dec 1992 17:58
Date: Thu, 24 Dec 1992 17:58
27 lines
1799 bytes
1799 bytes
I was just talking with a friend of mine today, and he asked about the market viability for a package that scales down JPEG format 24-bit or 32-bit pictures to 16-bit or 8-bit depths. I made some noises of dubiousness, pointing out that there are plenty of algorithms with nice implementations (in code frags or even complete programs) out which do exactly that right now. He countered that the package could make it in the market if it used the results of some multi-million dollar government-sponsored Los Alamos project which focused on how to compress high-quality images with little or no data loss. Frankly, I'm kind of doubtful about this. Notwithstanding the standard mathematical proof which shows the infeasibility of outrageous compression claims, I'm not convinced that the package would be a big hit because it is so specialized; its market would be a subset of the already small set of high-end graphics users. Could someone please confirm or refute the possibility of such a piece of software? -- ____ tyen@cs.utexas.edu.....Kowabunga!... /ony -- Claimer: Views expressed here == mine && only mine. Sail tough. -- Moderated by SCRI Vis <> Submissions to: graphics@scri1.scri.fsu.edu Guy, John R. Murray <> Administrivia to: graphics-request@scri1.scri.fsu.edu
Re: Scaling Down 24-bit Pix
Author: Tom_Lane
Date: Thu, 07 Jan 1993 19:17
Date: Thu, 07 Jan 1993 19:17
40 lines
2714 bytes
2714 bytes
tyen@cs.utexas.edu (Anthony Yen) writes: > I was just talking with a friend of mine today, and he asked about the > market viability for a package that scales down JPEG format 24-bit or 32-bit > pictures to 16-bit or 8-bit depths. ... He countered that the package could > make it in the market if it used the results of some multi-million dollar > government-sponsored Los Alamos project which focused on how to compress > high-quality images with little or no data loss. ... > Could someone please confirm or refute the possibility of such a piece > of software? What your friend is talking about is not data compression in the usual sense; it is color quantization, and it is HIGHLY lossy. The name of the game is to make a visually acceptable approximation to a full-color image within the constraints of a color-mapped display. (Incidentally, reduction for 15- or 16-bit direct-color displays has little to do with reduction for 8-bit color-mapped displays.) I believe the Los Alamos algorithm referred to is the one used by the commercial package "Fast Eddie" for Macintosh systems. Fast Eddie is a single-purpose program for 24-to-8-bit color quantization. The blurbs I've seen for Fast Eddie make great claims for its speed and the quality of the resulting images, but I have not seen any independent evaluations. If the underlying research was indeed done at Los Alamos, its results are probably public domain. Does anyone know where more information about the algorithm might be obtained? I would love to try it out for possible inclusion in the free JPEG software. regards, tom lane organizer, Independent JPEG Group PS: No, I don't think your friend would be well advised to try to build a commercial color quantizer, but not because there is no demand. Rather because (a) there are lots of good free ones, and (b) the demand is within larger software packages, not as a stand-alone product. Too bad there is no functioning marketplace in software components yet... -- Moderated by SCRI Vis <> Submissions to: graphics@scri1.scri.fsu.edu Guy, John R. Murray <> Administrivia to: graphics-request@scri1.scri.fsu.edu
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads