"Dave" wrote in message
...
Thanks for the feedback and patience; I'm not sure how many hits
Bellsouth takes before the Bwidth is exceeded. Meantime I'm leaving the
pics up so that, well...."Diffusion of Knowledge" and all that.
Hi Dave, nice site. Regarding the bandwidth issue, the problem is not the
number of "hits", but rather the quantity of data (measured in bytes) that
is downloaded by your visitors. I looked at your site last night (between
"bandwith exceeded" errors) and noticed that your large images are very data
"heavy" (i.e., they have very large data byte sizes).
To reduce the problems with "bandwidth exceeded" errors, you need to
optimize your photos so that they are not so data "heavy". You see, there
are two parameters when discussing the size of digital images: there are the
pixel dimensions (e.g., 800 x 600) which specify the viewable ("physical")
width and height of the image; and there is the data size which specifies
the amount of bits of information (measured in bytes, or 8 bits) in the data
file that represents the image. (The problem you are experiencing is in this
latter category.) Typically, there is a correlation between these two specs:
a "physically" larger image (in terms of pixel dimensions) generally
requires more data (bytes) to represent the image than does a "physically"
smaller image. But there is another relationship between these two
parameters: quality which determines how much data is used to convey a
particular visual artifact within the image. To express this as a simple
equation, you'd have:
dimensions X quality = data size
So there are two ways to reduce the data size of your images (so your web
server does not exceed its bandwidth limits): reduce the pixel dimensions or
reduce the resolution (or both). Most people like to see larger images, so
we'll focus on reducing quality. Here's where the magic happens.
The human eye is an amazing device, but it is not designed to perform
quantitative analysis of an image. In other words, it looks at patterns and
colors and shapes, but does not individually analyze minute differences in
adjoining pixes. So if two adjacent pixels in an image are very close in hue
and/or saturation, we can make them identical and the eye won't be able to
see the difference. In other words, we've (intelligently) reduced the actual
quality of the image (less data), but the subjective quality -- what the eye
can perceive -- is undiminished. And that's how image optimization programs
work: they first reduce the amount of differences in the pixels of the image
(so there is less image data to "describe"), and then they use mathematical
algorithyms to further "compress" the data so that even less bytes are used
to represent that image. Here's a simple example of a compression technique:
If your image data is: 11111100011110000
You could compress that to: 6x1,3x0,4x1,4x0
That simple example may not look like much of a savings, but the reality is
that tremendous data compression can be achieved with these kinds of
methods.
Ok, perhaps you don't want to know all the nitty-gritty that goes into
digital image data compression. But here's what you should know: by using an
image optimization program, you can take an 800 x 600 pixel image that
"weighs" over 200,000 bytes, and compress it so that it only requires 20,000
bytes (for example). That will reduce your web server's bandwidth load by
90%. And the optimization programs are so good, that the human eye probably
won't see much -- if any -- loss of quality (especially when that image is
being viewed on a computer monitor). Of course, the more you compress the
image, the more the visual quality of the image is reduced. But you can
substantially reduce the data size of an image before it starts to look
degraded.
There are a number programs to perform image optimization. I use Adobe
Photoshop, but that is (an expensive) professional-grade graphics tool.
There are some programs that are free and will do a fine job. Since I don't
have a lot of experience with these other free or lower-cost programs, I
can't make any recommendations. But people have mentioned IrfanView as a
free option.
I hope this helps.
- Michael
|