Re: [OT] ideal file size of a jpeg or gif image

["Alan Shaw" <brindabella@primus.com.au>]

My files are smaller than most railfan sites, which occur mostly because the
dimensions of my images were deliberately designed to be smaller. For
example, the absolute maximum size of any one of my images will be 575 by
450 pixels, which will take up about 54% of the total screen area at a
desktop area of 800 by 600 pixels. A typical image from James' site is 900
by 600 pixels, which at the same desktop area will take up about 113% of the
available area - hence the need to scroll. So my files are about half the
size of James' on a like for like basis. However, in reality my files are
about 20% of the size when taking into account the area presented.
Presumably the difference is because I have used a higher level of
compression than James did.

By the way, these comments are in no way intended to be criticisms of James'
site - I've always liked it: it's clean, well organised and has a wealth of
information on it - I just wish those darn photos wouldn't take so long to
load! However, we evidently have some different philosophies, and having
responded to my earlier post I'd like to discuss some of his comments and
present another view.

Obviously such decisions are very subjective, but when working out the
parameters for my site I did a number of comparative compression tests to
determine what was acceptable quality, given that I wanted quick download
times. There is some minor loss of quality of course, but to me the loss is
not significant, and certainly not when the variations in monitors and
viewing conditions are taken into account. The results are not low quality
by any means, although others of course can disagree! And I believe I
achieved my aim of having an acceptable compromise between quality overall
presentation and download time.

Also, I only want my images to be viewed as presented on the page, and have
never expected them to be used as backgrounds. If this were my intention
then yes, larger images would be needed.

The other thing I had in mind when I set out to design my site was to consid
er what annoyed me as a surfer, and then to avoid repeating those things if
I could. And I have to admit I do not have the patience to wait up to a
minute or more for large files to download, unless I know it is going to be
an extraordinary image. This is my choice of course, but I doubt that I am
the only one there are plenty of other things in life than tapping the
fingers waiting for big files to download! Which comes back to scrolling and
desktop area. Again, my image size was partly determined by deciding to
design my site at 800 by 600, which allows some room for captions and
overall presentation. Why 800 by 600: whether anyone likes it or not it is
still by far the most common desktop area.

It's fair enough to say that all the user needs to do is increase the
desktop area but there are two things here: not all PCs allow this, even
relatively modern ones (where I work for example we have machines by a major
quality manufacturer only a couple of years old but they will simply not
permit adjusting beyond 800 by 600 - why I don't know, but there you are)
and further there seems a bit of elitism to suggest that hardware needs to
be adjusted for the pleasure of viewing someone else's work. Given that
about two thirds of browsers are set at 800 by 600 I felt that using a
higher value could be self-defeating - I want people to look at my site and
come back, not feel like they are not welcome because the standard of their
hardware or their choice of desktop area does not measure up to "my"
standard!

So in this sort of situation large files are a real bugbear for me: they
take too long to load and I have to scroll around. Which is why in my
earlier post I said that I thought it was unnecessary to have such large
images on the web. However, I take James' point that if Windows backgrounds
are required then bigger files will be needed. On the other hand, if I
thought people wanted to use any of my images for backgrounds then I'd
invite them to e-mail me and I would then make larger files available to
them specifically, rather than cluttering up the web with big files for
everyone else generally!

Scrolling to me is a good example of how technology represents a major step
backwards, especially when in some cases it is forced on the viewer, rather
than a choice.  When I look at a print I can immediately take in the whole
composition and then study the detail, but with scrolling I never get that
opportunity. When presented with this situation, my response is to usually
think Thanks, but no thanks, and click elsewhere.

Incidentally, my site will be made public (for better or worse!) around the
end of the year. If anyone though wants to look at it before then and see
whether my approach does indeed result in low quality images, please e-mail
me and I'll give you the URL.

Indeed, I think the discussion about quality can go one step further. There
is clearly a huge leap between Internet images of poor quality and good
quality, but I'm not so sure there is an equal leap between good quality and
fantastic quality, at least overall given variations in screen quality,
colour fidelity and viewing conditions which have to be contended with. If I
really wanted high quality then I'll forget the Net and look for a print!

Alan