1999.07.15 23:47 "RGB Gradient Code, adapted from Andreas Neumann", by Eric B. Middlecamp

1999.07.16 17:39 "Re: RGB Gradient Code, adapted from Andreas Neumann", by Eric B. Middlecamp

I made a simple extension to Andreas Neumann's gradient code, (see http://www.gis.univie.ac.at/strv/strv/leute/andi/computer/graphics/raster-formats/example_tiff_main.html) but the output TIFF image has some odd contamination. Looking at the image, you wouldn't be able to tell -- however a histogram reveals that some levels are omitted and others are doubled.

Has anyone else experienced bizarre anomalies like this?

The code looks fine, and Niles was able to run it with no problem. But you don't say anything about what type of computer, operating system, compiler, image display program, how you produced the histogram or what version of libtiff you used in your message.

Ok, here goes.. I am running on a Power Mac 8100/100 under MacOS 8.1, I am compiling under Codewarrior IDE 2.1, displaying the images in Adobe Photoshop 5.0, producing the histogram in Photoshop with full data sampling (histogram not from cache), and I am using a Mac-compiled version of the old libtiff "tiff-v3.4beta028," although I have compared the code for libtiff version "tiff-v3.4beta037" and not found any major differences.

As an update, I just ran a modified version of "tiffinfo," it verified that the image I have matches what Niles got when he ran it; the problem must be related to the image readers, namely Photoshop. I am also writing some code which reads TIFFs, though, and I have gotten some strange results there as well. Is it possible that I am using a valid, but uncommon/problemmatic, set of tags?

Eric Middlecamp

=========================================================================
sample of modified tiffinfo output on a RGB gradient image:

Strip 0: (00)
  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

Strip 1: (01)
  01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01

Strip 2: (02)
  02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02

...

Strip 253: (fd)
  fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd

Strip 254: (fe)
  fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe

Strip 255: (ff)
  ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
 ========================================================================