Working with Images

Fred Ridder docudoc at hotmail.com
Sun Feb 3 08:04:16 PST 2008


Pete Rourke asked several questions:
> I am using Snagit (default image resolution 96dpi, and saving as .jpg) to
>  capture screenshots for a end user manual which assumes the user needs
> visual walkthrough of using a desktop application and a pocket pc.
> 
> There are 2 outputs intended:
> 
> 1. PDF leveraging all the indexing and cross-referencing
> 
> 2. A printed manual
 
No problems here. Choose any of the four non-lossy raster image formats.
 
> An issue is the volume of screenshots ~ 200 in a ~ 150 page document.
> 
> Image files saved as .jpg, average file size = 44K 
> 
> So far the total image in the books is 8MB (gag)
 
What's the issue?  8MB is nothing, these days. On a hard disk, that
much storage costs less than $0.01 (with disk prices currently running 
between $.20 and $1.00 per Gigabyte). And archiving to CD or DVD only
costs another penny. 
 
And when you're working with your document, referenced images (as 
opposed to pasted-in images) are only loaded when you view a page,
so the amount of data to be loaded for images shouldn't be an issue 
unless you're working on a *REALLY* slow network. The *number* of 
screenshots is more of a concern in this circumstance than the overall 
space occupied by the files, because each file has to be accessed
separately; but that's something that you, as author, control completely.
 
> So a question is what format, JPG, BMP, PNG, GIF saves the cleanest picture?
 
JPEG is the worst choice for screen shots. Don't use it for that purpose.
The "P" in JPEG stands for "photographic", and continuous-tone photographs 
is what it was designed for. The compression technique the format uses
is area-based and lossy, and relies on two important characteristics of 
photographic images to reduce the visibility of the artifacts and image 
degradation the compression inevitably produces. JPEG inevitably produces
artifacts (a kind of "smudginess" near abrupt transitions between different
colors, but this is OK because photos have relatively few of these and 
because the objects in the photos generally have anough surface texture
to mask the artifacts. But screen shots are all about hard edges and 
abrupt transtions (that describes text characters, for example), and has
*no* texture to mask the artifacts.
 
My preferred file formats for screenshots (in order) are:
GIF: very compact thanks to run-length encoding and indexed color,
   but produces posterized results on graduated color areas or photos.
   Can be used directly one the web and cross-platform.
PNG: a modern superset of GIF supporting full color depth. Can be used
   directly on the web, but some browsers have issues displaying some 
   "flavors" (color formats) of PNG. 
TIFF: good platform portability (usable on Windows, Mac, Unix) but large 
   filer size than PNG or GIF. Not directly usable on the web.
BMP: poor portability (Windows-only); large file size; not web-compatible
> During import I choose 150 DPI, am I insane?
 
No, but if you're really concerned about file size, this is one place you 
can make some improvement. Saving screenshots at 150 dpi rather than 
96 results in image files that are 2.44 times as large. simply because there
are that many more pixels to store.
> I am not sure what resolution is required. Would less than 96 be
> acceptable? 
 
Dropping below 96 dpi loses a lot of quality and doesn't save much file
space (assuming that this really is a legitimate issue). 
_________________________________________________________________
Shed those extra pounds with MSN and The Biggest Loser!
http://biggestloser.msn.com/


More information about the framers mailing list