🚀 go-pugleaf

I was concerned about the standardization of JPEG
between two programs I have, so I did some tests.

The original file was a 24-bit true-color file of
size 768 x 768 at 300 dpi, scanned into PhotoStyler.
I saved the file in two formats: 24-bit Targa and
24-bit BMP.

To ensure that both files had the same color
information, I checked the number of unique colors
using Paint Shop Pro. Both TGA and BMP files reported
having 167510 colors.

Then I compressed them using ColorView 2.0 for DOS
and WinJPEG 1.6. (The reason why I used a quality of
74 and 76 for Colorview was because it did not
offer 75 as an option; so I also used those settings
for WinJPEG.) Also, Cview only accepts BMP files.

Original   Software   Quality   Entropy   File
  File       Used     Setting   Optmztn   Size
------------------------------------------------
  TGA      WinJPEG      75        Yes     135909
  BMP      WinJPEG      75        Yes     136256

  BMP      Cview        74        N/A      81795
  BMP      Cview        76        N/A      84487

  TGA      WinJPEG      74        Yes     133183
  BMP      WinJPEG      74        Yes     133519

  TGA      WinJPEG      76        Yes     141489
  BMP      WinJPEG      76        Yes     141800

  TGA      WinJPEG      74        No      143693
  BMP      WinJPEG      74        No      143693


I am very confused by WinJPEG's method of saving JPG's.
WinJPEG takes two original files of the same content but
different format, yet outputs JPEG files of different sizes
(with entropy optimization on).

Cview's JPEG output is much smaller than those of WinJPEG
for a given Quality.

My questions are:

Which program adheres closer to the Quality settings
described in the JPEG FAQ?

Exactly what is entropy optimization and how does it affect
the quality of the compressed image? (Should it be used if
I want a "standard" Q75 JPEG file?)

And for those who have PhotoStyler 1.1a: how does the
quality range in the JPEG export module translate to the
quality range described in the FAQ? (PhotoStyler uses
a range from 0 to 100, with 0 being minimal compression,
100 maximum - just the opposite of convention). What
the heck is the difference between YUV422 and YUV411
SubSampling?

Sorry to ask so many questions, I'm just very confused.
Thanks for any info.

					-Kijin
-- 
 __ Kijin Jung ________________________ To affect the quality of the day, __
|   kxj6@po.cwru.edu                 |  that is the highest of the arts.    |
|   Case Western Reserve University  |                                      |
|__ (216) 754-1101 __________________|________________________ -Thoreau ____|