From: Don on 15 Apr 2006 10:49
On Fri, 14 Apr 2006 09:31:07 -0400, Raphael Bustin <foo(a)bar.com>
>>At 100% magnification (i.e. 1:1) even a JPG image at lowest
>>compression (i.e. highest quality) stands out like a sore thumb when
>>compared to the original.
>Sorry, that's BS. See my earlier post on this subject.
Or an even earlier one (as my "inconsistency alarm" went off):
--- start ---
On Fri, 27 Jan 2006 19:17:58 -0500, rafe b <rafebATspeakeasy.net>
>Lately I have been doing more scanning
>and saving in 48-bit color though it's
>really just caving in to peer pressure.
>I still don't really see the point.
--- end ---
If you're so convinced you can't even tell JPG from TIF then scanning
at 48-bit color makes no sense at all i.e., it's inconsistent.
If someone were as convinced as you now state to be, they would scan
at 8-bit and resist the peer pressure, not succumb to it.
But, be that as it may...
>In a blind test, you could not tell them apart.
Yes, I can.
>Here's a 4000 dpi film scan snippet as JPG:
>and here's the same as a TIF:
>These are each 1000 x 1000 pixel crops straight off
There are far too many unknowns here i.e. it appears you've done two
separate scans which, sort of, defies the whole purpose. So, before
downloading them and then digress and waste time, we need to clearly
define the testing environment.
In the current context, in order to test the JPG conversion and
minimize (eliminate) other irrelevant influences the optimal procedure
is to perform a single *raw* (!) scan using *reliable* (!) scanning
software. Why? Because, in both cases, if the scanning software itself
mutilates the image there's nothing left for JPG to mutilate.
Next, use any software of your choice to turn this *unedited* raw scan
into JPG of "highest quality". That's a flexible term since JPG
quality is not standardized, but never mind... Even with this unknown
factor i.e., one arm tied behind my back, I'm still game.
Finally, we're talking about a reasonable image (e.g. in focus, etc)
which actually contains some fine detail i.e., no underexposed black
cats in tunnels, or overexposed polar bears in snow, etc. One can
always come up with an extreme exception but I think it's quite
self-evident that's not what we're talking about here.
After that, using the *original* (unedited!) raw scan as it came off
the scanner and its JPG version, we can do a meaningful head-to-head
From: Raphael Bustin on 15 Apr 2006 12:50
On Sat, 15 Apr 2006 16:49:12 +0200, Don <phoney.email(a)yahoo.com>
>There are far too many unknowns here i.e. it appears you've done two
>separate scans which, sort of, defies the whole purpose.
What gave you that idea, Don? You're dead wrong.
This is the same bitmap image, saved two ways.
>downloading them and then digress and waste time, we need to clearly
>define the testing environment.
Try not wasting our time by making unfounded
assumptions about what you're looking at.
This is a scan snippet straight off the scanner. Same
scan snippet in both cases, saved two different
ways (ie., as TIF in one case and as JPG in the other.)
I said as much when I presented these images. Why are
you playing stupid? Because the results disprove your
own "expert" opinions?
From: Alan Meyer on 15 Apr 2006 16:46
If I ever get a job as director of the Metropolitan Museum of Art
and need a good man to do scanning, Don is someone I would want
to call. I would know that the scans he did are as perfect as
one could make them.
He's clearly a knowledgeable guy and a perfectionist.
But for those of you who are just trying to scan the collection
of family snapshots, I'm not sure that this kind of perfectionism
repays the effort it costs.
Let's say you do a good job of scanning that 3-1/2x5 black and
white snapshot of your Mom in her girl scout troop. You've
scanned at a resolution that will print at 8x10 without
pixelating, and you've captured pretty much all of the detail
that was there in the print. Anybody looking at your scan will
see a very good facsimile of what was in the snapshot.
It might even look better than the snapshot. You might have
played with it a bit - increasing the contrast and sharpening it,
or you might have used your scanner software's automatic photo
retouching control to do that for you. Then you saved it,
Do you realistically think you will ever go back to this image
and edit it? More than once? More than twice? Will it be
published some day in an art magazine? Or perhaps pored over by
historians who are looking for details of the insignia on Girl
Scount uniform buttons from the 1930's?
If you think all or any of that might happen. Don's your man.
Follow his advice. Scan at ultra high resolution. Save the raw
scanner output, or as close as you can get to it, as well as any
edited version. Use 36-48 bit image encoding and lossless
compression to save the files.
But if you think what will happen to these images is that you and
your family and your descendants may pull them out from time to
time and have a look, then isn't that all overkill?
Personally, I'll be thrilled (in advance of course, since all my
thrills will cease with my demise) if _anyone_ ever bothers to
look over the old photos I'm scanning.
What Don is recommending is absolutely right both for him and for
anyone who shares his goals of preserving the very best possible
scans. That may well include the original poster.
And as Don was the first to say, I'm right too for myself and
people sharing my goals.
If you're going to put many hours into scanning - figure out
which goals you have, and what meets your needs. Then invest the
right amount of time and megabytes to meet them.
From: Raphael Bustin on 15 Apr 2006 17:07
On Sat, 15 Apr 2006 16:46:32 -0400, "Alan Meyer" <ameyer2(a)yahoo.com>
>If I ever get a job as director of the Metropolitan Museum of Art
>and need a good man to do scanning, Don is someone I would want
>to call. I would know that the scans he did are as perfect as
>one could make them.
Hahaha. Don would be too busy "disassembling" the code for
the scanner driver and complaining about the incompetent fools
that designed it.
From: Bart van der Wolf on 15 Apr 2006 20:03
"Raphael Bustin" <foo(a)bar.com> wrote in message
> On Sat, 15 Apr 2006 16:49:12 +0200, Don <phoney.email(a)yahoo.com>
> Try not wasting our time by making unfounded
> assumptions about what you're looking at.
History tells that that might be a little too much to ask for ... ;-)