Dave "SpeedHump" Hinson

on how to scan pictures 
and other suggestions on image files

Hi Dario!
         I really thought your message was the usual
Quinto-spam before I opened it up.  As you know we
have been in KL since mid July.  We are enjoying it
very much and are almost settled into our penthouse
apartment.  I have been working full-time at the
Canadian Embassy or I should say "Canadian High
Commission", it being in a Commonwealth country - I
will refer to it in the future as "CHC". When I first
arrived I attended a Saturday Hash (Petaling Hash)
avidly, however it was biting into my spare time too
much and I had to do a lot of work in the apartment.
So - for the last couple of months I haven't hashed.
In the New Year I will get going again.  My brother
was over from Australia for the last month and we have
had some excellent times together - he leaves tonight.
 Suzie has a fairly decent boss and likes the work -
and the country.  It is really a Mecca for Eastern
food - and cheap to live too.  If you ever have any
time off grab an Air Malaysia flight down here - We
will be happy to look after you.

Now your scanning enquiry - always scan your treasured
images at maximum resolution in bmp files.  If you
have a CD burner - and they are cheap these days -
burn copies and put them away.  For web use it is best
to reduce the image size to maximum about 640 x 480.
Banner size type images can be cropped to - say 640 x
200
It is often common to then save them in GIF for web
use
as they - having only 256 colours in gif - will load
up very rapidly.  You can also prepare gif images with
tranparent backgrounds. You will probably need some
decent programs for that.  If your law background does
not hamper you too much, such applications can be
bought down here very cheaply.  When I come in March
next year I will bring you Adobe Photoshop 7.0 and
anything else you like.

Cheers

David

Metafiles

Vector drawing programs such as CorelDraw or Microsoft
metafiles would describe the Japanese flag something
like this:-
" go to the centre of a white field x wide and y high
- from there draw a circle of z diameter and fill it
with red"  These vector images and can be very brief
and can be smoothly scaled up and down without
distortion, but a great deal of skill and hard work is
needed to create realistic images.  This is why in a
competion by Corel a few years ago the winning
creation was a monochrome lifelike portrait of Hedy
Lamarr - now used in all their advertising.  Not so
obvious to the uninitiated but mindblowing to the
cogniti.

Bitmaps describe each pixel one after another - big
files.

Compression files analise the  flag and say "the first
8lines fill white - the next 8 lines up to position
"a" fill with white followed by a red fill up to
position "b" - now fill across with white again"  This
can be quite cryptic - and different methods or
algorythms produce better or worse results.


The above descriptions of course ignore the actual
file size - and of course an image of 320 x 240 pixels
could not ever equal an image of 1600 x 1200 pixels.
The resolution just is not there

Resolution

Images are constructed from pixels.  The pixel's
colour and brightness is described by data.  In the
simplest form it would require only 2 bits of data to
descibe whether a pixel was black - or white. Then
comes early computer images which were in 16 colours
only and with "bright" on - or "off"  With this small
amount of description only rudimentary images could be
made.  Next would be those image file that use 256
colours.  These are GIF images - the number of colours
are often sufficient to construct quite good images
and because they require comparatively small amounts
of data to describe the picture one wants to make,
they are popularly used to transmitted by e-mail.
Hence the name - Graphics Interchange Format - GIF .
No waiting an age to download - on the web these
images pop into view in seconds.  TIFF files use more
colours and are consequently bigger and better defined
because of the use of more tones.  Your Mona Lisa now
looks less like a big daub - but still not quite right
somehow.

Jumping to the other end of the scale bitmap images in
24 bit colour must describe the subtle differences
between over 6 million colours.  36 bit colour even
more.  All of this reqires data and of course the
files can become unmanagebly big - especially for
transmission down phone lines.  Therefore along comes
compression techniques - i.e. jpeg.  This method looks
at a block of pixels at a time and ascribes a value to
the block.  For instance one block of - say 8 by 8
pixels may be simplified by seeing it as in one tone
that gets progressively darker from top to bottom of
the block - naturally this can be more briefly
described than the real situation where each of the 64
pixels were thoroughly described by the data.  Another
block may simplified and described as getting darker
from the top right hand corner down to the bottom left
hand corner - as so on.  This analisis and
simplification is all decided by a particular jpeg
formula - or algorythym.  Some jpeg conversion
programs are better than others.  For instance I
discovered jpeg conversion in "Fauve Matisse" (the
program I gave you) is not very good and I avoid using
it.  Adobe Photoshop is much better.   As a matter of
interest digital cameras too converts the real colour
that come through the lens and simplifies it to a
greater or lesser extent.  Some cameras have very good
jpeg conversion and others use very high compression
formulas in order to get as many pictures in to the
meagre memory banks it has.  The older Sony Mavica
cameras needs to get its data onto a 1.4 meg diskette
and they decided to compromise the image quality for
the number of pictures that would go onto the
diskette.
Later Mavicas allow a choice of bitmap format.  Here
it could be that one picture uses one diskette!

I hope all this helps, Dario.

I will be using this description again for other so
don't think I wrote this especially for you - which I
did initially


Cheers -  David

November 29,  2002