home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!infonode!tdj
- From: tdj@infonode.ingr.com (Ted Johnson)
- Subject: Peculiar limit?
- Message-ID: <1992Dec17.202531.20920@infonode.ingr.com>
- Organization: Intergraph Corporation, Huntsville, AL.
- Date: Thu, 17 Dec 1992 20:25:31 GMT
- Lines: 29
-
- Is the following true?
-
- g g g g
- (1/b) - (1/a) (1/b) - (1/a)
- lim ------------------ = lim ------------------
- g->+inf. g g g->-inf. g g
- (1/c) - (1/a) (1/c) - (1/a)
-
-
- ln(1/b) - ln(1/a)
- = -----------------
- ln(1/c) - ln(1/a)
-
- If necessary:
- 0 < a <= b <= c <= 1;
- a < c.
- Everything's real.
-
- I'm not so concerned about the first line as with the second. Some
- numerical tests led to this conjecture.
-
- The context is my work on response characteristics for a scanner (of
- photographs). If you're familiar with D-log(E) plots, read "gamma" for
- g and you'll know what I'm referring to in the top line. These two
- equations represent the (limits of the) normalized exposure as computed
- from the transmissivity T of the image, all derived from the linear
- portion of a D-log(E) plot. The bottom expression represents the
- normalized density of an image, density = 1/T, computed from the
- transmissivity of the image.
-