Archiving: TIFF or PSP, 16 bit or 8 bit?

  • Thread starter Thread starter Robert A
  • Start date Start date
Kennedy McEwen said:
...
Certainly don't even consider PSP as an archive format, it is a
proprietary coding scheme which may not be supported in the future.

Not true; the Paint Shop Pro file format (at least through v8) is
documented[1], as are its compression schemes (RLE and LZ77[3]). As
proof of this, I have written a PSP format plugin for Photoshop[2]
which is interoperable with PSP 5-8 *and released as open source under
the GPL*. It is a more open format, for example, than Photoshop PSD or
PSB; documentation for those is not freely available.

Arguments for or against archiving in PSP format might perhaps take
into account issues such as metadata and colour profiling. For
interoperability with non-proprietary tools, standardisation and
"future-proofing", TIFF or JPEG seem very good choices.

Toby

[1] http://www.jasc.com/support/kb/articles/pspspec.asp
[2] http://www.telegraphics.com.au/sw/#pspformat
[3] A PSP-compatible FREE implementation of LZ77 exists:
http://www.gzip.org/zlib/
 
I wrote: "I archive 3800x2700-pixel color-corrected images as
level-6-quality (Photoshop) jpeg's. [snip] As for "quality" issues of
jpeg's, I don't believe the differences in jpeg's and tiff's is
noticeable at any normal resolution."

"Phil Rose" responded: "Not noticeable at what jpeg compression and when
standing how far away? Again, such a statement about "noticeable" ought
to be qualified with the intended use, but even for images merely
displayed on a monitor I believe your statement is a "stretch". It does
tend to be true when comparing tiffs with level 10-12 jpegs, but I don't
agree it is true at all for the level-6 that you stated earlier. Toggle
Photoshop's preview button on/off when selecting various jpeg quality
levels--and look at the changes in the image (at 100%); it should be
relatively easy to see the edge-softening and color/noise artifacts that
are generated when going from tiff (or psd) down to a jpeg quality level
of 6."
----------------------

I'm thinking that more pixels (even when jpeg-compressed) are much more
valuable than fewer pixels in tiff format. I began scanning 35mm at
1410ppi (Minolta Scan Dual III) because I thought all I'd ever need was
4x6 prints or screen displays. However, experience showed that
occasionally I wanted an 8x10 print and frequently I needed to crop
images, so 1900x1250 pixels wasn't enough. I changed to the larger scan
size and now I resize down for prints and screen display. If I could
see the difference in Level 6 and Level 12 jpegs, I'd use Level 12. I
can't see any difference on screen at 100% (using the Save-for-Web
preview).

Also note that with these larger images, when I resize to screen size
(1024x788) or 4x6 print size (1200x1800) it is the equivalent of viewing
at 33% to 50%. It only becomes 100% when I have to severely crop an
image.

I've done some tests in the past that showed that for high-quality
halftone-screened 4-color images (in this case advertising images for a
large furniture manufacturer), Level 6 jpeg compression was
indistinguishable from corresponding tiff images (both at 300ppi and
viewed at "reading" distance). I'm thinking for "valuable" images, it
*might* be worth the disk-space investment in losslessly-compressed
images, but for everyday pictures, the added storage space requirement
isn't worth it, particularly when considering that we are frequently
talking about thousands of images.

Preston Earle
(e-mail address removed)
 
Mike said:
Certainly I agree that a higher scan rez extracts more information.

My point is that sharpening is an indispensable step after resampling. That
resampling may be explicit, as when you resize an image in Photoshop, or it
could be implicit, as when a large image is printed at a small size.
But as you can see, from your first quoted line above which you have
conveniently retained throughout this thread, your statement concerned
high resolution scanning, not resampling.
 
Toby Thain said:
Kennedy McEwen said:
...
Certainly don't even consider PSP as an archive format, it is a
proprietary coding scheme which may not be supported in the future.

Not true; the Paint Shop Pro file format (at least through v8) is
documented[1], as are its compression schemes (RLE and LZ77[3]).

As were the technical details of Betamax!

It's specification may be available but it is still a proprietary format
that is, in the main, only supported by JASC software and owned by them.
Remember the GIF format and the Unisys debacle?

Nobody else really bothers with PSP simply because it offers little that
is not already available or bettered in industry standard formats. TIFF,
in particular, supports a wide variety of compression techniques or,
indeed, no compression at all, making it completely immune from Unisys
type action.

Folks will be using TIF and JPG formats long after JASC have gone bust
or attempted to revoke the licences for PSP.
 
: "Scanning at an increased sampling density may
not offer any more resolution in the image, but it certainly cannot make
it any less sharp or softer!" and, later, "But as you can see, from your
first quoted line above which you have conveniently retained throughout
this thread, your statement concerned high resolution scanning, not
resampling."
---------------------

I haven't followed this thread closely, and I don't know anything about
Nyquist frequencies, but aren't you quibbling a little? <g> If an image
is printed at a particular size from two otherwise similar files of
different resolutions, the image from the higher-res original will be
softer and appear less sharp. Whether from scanning or resizing, the
higher-res image will be softer, will it not? [See page 12 of the PDF at
http://www.ledet.com/margulis/PP7_Ch15_Resolution.pdf (or page 306 of
the book).]

Preston Earle
(e-mail address removed)
 
Preston said:
: "Scanning at an increased sampling density may
not offer any more resolution in the image, but it certainly cannot make
it any less sharp or softer!" and, later, "But as you can see, from your
first quoted line above which you have conveniently retained throughout
this thread, your statement concerned high resolution scanning, not
resampling."

Obviously not! :-)
If an image
is printed at a particular size from two otherwise similar files of
different resolutions, the image from the higher-res original will be
softer and appear less sharp.

That is complete rubbish - the higher resolution image will always,
unless you have had to degrade it in some way to meet the constraints of
your printer, be much sharper than the lower resolution image (assuming
that the image contains adequate fine detail with which to observe the
difference in the first place).

Where on earth did you ever get the idea that higher resolution meant
less sharp results?
Whether from scanning or resizing, the
higher-res image will be softer, will it not?

Definitely not. A higher resolution scan will contain finer detail and
sharper edges than a lower resolution scan, if the information is
present in the image in the first place. Even when it is not, it cannot
be less sharp than the lower resolution scan, only as sharp. A resized
image can never contain any more detail or sharper edges than it already
has. How much the resized image softens depends on the algorithm used.
For example, nearest neighbour interpolation will retain apparent edge
sharpness completely, whilst bicubic will soften it slightly and
bilinear more so. None of these algorithms will provide an image as
sharp as a higher resolution scan - again assuming that the image has
higher resolution content to bring out in the first place.
[See page 12 of the PDF at
http://www.ledet.com/margulis/PP7_Ch15_Resolution.pdf (or page 306 of
the book).]
Not the same thing at all. Both images have been resampled to exactly
the same resolution for presentation on the page. This is fairly
obvious if you zoom into the images in the pdf file you referenced -
although the upper image has been scanned at 3x the resolution it has
exactly the same pixel dimensions as the lower image.

What you are looking at here is 3rd (and higher, odd) harmonic
distortion caused by reproducing each sample as a square pixel on the
page. (Recall harmonic distortion in audio - well you get it in images
too!) Each sample, however, only represents the image at an infinitely
small point in space, called a delta function, which has a volume equal
to the average light incident on the CCD sensor centred at that point.
The sample, in reality does not exist anywhere else, however an array of
delta functions is not a particularly useful thing to look at - for one
thing they require an infinite video bandwidth to reproduce on your
monitor, and an infinite dpi printer to represent them. So each delta
function is represented instead by a pixel, which has a finite dimension
but is, in fact, a completely false representation. What should occur
between the samples depends on how the user chooses to reproduce the
delta function in pixel terms - how he *interpolates* between the
samples. Block pixels are simply a uniform square interpolation -
introducing every odd harmonic spatial frequency above what is possible
for the samples to contain, which is simply false information. However
they do make the image look artificially sharp. It is important to draw
a distinction right away between the use of the term interpolation here
and what is normally referred to by the same term in upscaling - this
is simply how each sample is represented by a pixel, in terms of it
size, shape and intensity profile, in the final image.

That conventional square uniform pixel reconstruction process is no more
valid than a linear interpolated pixel, where each pixel is represented
by an intensity at its centre prportional to the volume of the delta
function it represents and which linearly merges to reach the intensity
of the neighbouring pixels at their centres. In the simplest, bilinear,
case each pixel is effectively a square based pyramid (height
representing intensity), with the corners incident on the centre of the
neighbouring pixels. Although bilinear is the simplest version of this
and implements the linear merging only in the horizontal and vertical
axes, you can imagine octagonal interpolation where the intensity of
each pixel merges to the 8 nearest neighbours, or even circularly
symmetric interpolation. Clearly such interpolation schemes cannot be
linear, since the sum of the uniform samples must also be a uniform
illumination field but, nevertheless, such interpolation is possible.
Similarly there are higher order profile pixels which have intensities
which are polynomial curves, even pixels which extend their intensity
profile well beyond their nearest neighbours, although being constrained
to zero at them. Indeed, the ideal pixel reproduction, which introduces
no spatial harmonic distortion on the image at all, would have just such
a profile, extending to infinity in all directions.

What you are doing when you upscale an image using bilinear, bicubic or
any other interpolation method is *simulating* that pixel representation
by using higher resolution pixels to create the intermediate samples.
What you therefore perceive as a reduction in sharpness through bilinear
upscaling is merely the effect of a different pixel reproduction, not a
loss in sharpness over the original lower sampling density original.

The proof of this? Simply upscale using nearest neighbour
interpolation. That gives you a simulation of the square uniform pixel
reproduction using several new pixels to represent each old one but now,
of course, the effect of sharpness is retained.

In short, the effective sharpness of a scaled image is nothing to do
with the resolution, simply how each pixel is represented in the first
place. Quite different from increased scanning resolution, where
information conveying true image sharpness is pulled off of the original
medium, rather than synthetic odd harmonic distortions.

Mike was not referring to how the image was printed or reproduced in his
comments, merely what happened when an image was scanned - hence my
original question.

Nevertheless, since you clearly believe that higher resolution scans are
softer and thus, by default, that lower resolution scans are sharper,
can we expect to see your Minolta film scanner appearing on e-bay whilst
you "trade up" to a sharper 300, perhaps only 100ppi, piece of
antiquity? Why don't you just go the whole hog and flash a single
photodiode at your slides to get an ultrasharp 1x1 pixel rendition of
the entire image on each slide. the next step is just to remove the
sensor completely and type a random character into a file called
"image.raw" and observe the infinite sharpness of it all. ;-)
 
Not likely, if it already appears as a halfway decent image.

16 bits may be useful for the extreme tone-shifting adjustments,
like gamma specifically, but histogram B&W Points or Curve too
(the latter is of debatable benefit, but it is popularly done as 16b).

If we are scanning and saving RAW data (no adjustments done), then 16
bits is good, since all of these operations are still to come. This
would be the purpose of 16 bit data.

But if these operations are already generally done (and archive seems to
imply that), then there would be no point of saving 16 bits.
Additional fine adjustments dont need 16 bits at all.
When I say archive I mean scan - no adjustments whatever - archive. I
don't men archiving after adjustments. :)
 
That is the $64k question. As mentioned, the general consensus is that
after level adjustments are made there is little point in retaining the
additional bits. As mentioned, the challenge is still out there to
provide examples where this is not the case, but I am not aware of
anyone having successfully done that (though quite a few have tried).

I agree. However, there is good reason to retain the file if you do
so *without making any adjustments*. Which is what I do.

It's a bit like archaeology where they never excavate a whole site
because the technology will improve giving the opportunity to find out
more about a given site. If you save at maximum bit depth without
making adjustments you will then always have a raw (or even RAW <g>)
file which may produce better, or different results at a later date
because

a. the technology may have improved and,
b. you may want to make different adjustments to the file than those
you originally thought of.
 
When I say archive I mean scan - no adjustments whatever - archive. I
don't men archiving after adjustments. :)


Unless you have some way to specify a RAW scan (I cant imagine wanting
RAW), then by definition the scanner software has already done gamma.
Gamma does need more than 8 bits, which is why scanners are built that
way, but the scanners do this. And the scanner software will also have
generally made a first try at the histogram end points, at least the
coarse adjustment, so to speak. So there have been adjustments, and that
is what I meant "if it already appears as a halfway decent image".

Other than non-photo contrivances, it has never been convincingly
demonstrated that 16 bit output actually helps photos. Cant hurt however,
other than time and space, and some people do it anyway. Me too, at
times. But keep in mind that other than one of the few 16 bit editor
programs, there is no other use for a 16 bit image.
 
Kennedy McEwen said:
Toby Thain said:
Kennedy McEwen said:
...
Certainly don't even consider PSP as an archive format, it is a
proprietary coding scheme which may not be supported in the future.

Not true; the Paint Shop Pro file format (at least through v8) is
documented[1], as are its compression schemes (RLE and LZ77[3]).

As were the technical details of Betamax!

It's specification may be available but it is still a proprietary format
that is, in the main, only supported by JASC software and owned by them.
Remember the GIF format and the Unisys debacle?

Unlike GIF, unencumbered PSP readers and writers exist, as I have
shown.
Nobody else really bothers with PSP simply because it offers little that
is not already available or bettered in industry standard formats. TIFF,
in particular, supports a wide variety of compression techniques or,
indeed, no compression at all, making it completely immune from Unisys
type action.

Folks will be using TIF and JPG formats long after JASC have gone bust
or attempted to revoke the licences for PSP.

All true... which is why I suggested them in my posting.

The rest of my post was merely intended to correct some wild
misconceptions. (In that vein, understand that Jasc cannot "revoke
licenses" for free code that reads and writes PSP as it has been
documented to date. Unisys' patent was on the specific LZW compression
method. Jasc has no such hold on LZ77, for instance.)

Toby
 
Hecate said:
If you save at maximum bit depth without
making adjustments you will then always have a raw (or even RAW <g>)
file which may produce better, or different results at a later date
because

a. the technology may have improved and,
b. you may want to make different adjustments to the file than those
you originally thought of.
Unless you evolve some new technology to replace your eyes then "a" is
irrelevant. However, if you are going to save without making any
adjustments at all then do so at the highest bit depth available.
 
Toby Thain said:
Unlike GIF, unencumbered PSP readers and writers exist, as I have
shown.
As did many readers and writers for GIF, until Unisys decided to enforce
their IPR which they had previously been quite happy for everyone to use
openly.
In that vein, understand that Jasc cannot "revoke
licenses" for free code that reads and writes PSP as it has been
documented to date. Unisys' patent was on the specific LZW compression
method. Jasc has no such hold on LZ77, for instance.

Patents are not the only form of IPR and Jasc certainly do own IPR in
the PSP format. What they choose to do with that in the future is
anyone's guess, just as nobody would have predicted Unisys enforcing
their IPR in the GIF format.
 
Kennedy McEwen said:
Unless you evolve some new technology to replace your eyes then "a" is
irrelevant. However, if you are going to save without making any
adjustments at all then do so at the highest bit depth available.

Just guessing but, if "Hectate" was thinking of a VueScan Raw (64-bits), it
is possible to benefit from e.g. improvements that Ed Hamrick makes on his
IR-cleaning method, and he's working on his "Curves adjustment option". It
happened in the past, IR-cleaning improved and all I had to do was rerun VS
on the Raw file and never had to touch the film untill I got a better
scanner.
I am also thinking about improved tonescaling or High Dynamic Range
compression based on Raw scan data.

But you are right, once adjustments have been applied, and no significant
new adjustments are anticipated, there's little benefit in keeping more than
24-bits color.

Bart
 
You might take a look at both the TIF and PNG formats. I think they'll both "be around"
for quite a while.

PNG has many of the capabilities found in the TIF / TIFF format ... but not(yet) the usage
in the photo and publishing worlds. Additionally ... by design and because of the Unisys
fiasco ... it has a compression scheme that is not covered by patents. It's usage can
easily extend past the 'web only' realm, and might be a good candidate for archiving. The
only thing, I don't think that it can handle CMYK (yet).
 
Other than non-photo contrivances, it has never been convincingly
demonstrated that 16 bit output actually helps photos. Cant hurt however,
other than time and space, and some people do it anyway. Me too, at
times. But keep in mind that other than one of the few 16 bit editor
programs, there is no other use for a 16 bit image.

Understand. It's purely for image Editor use i.e. Photoshop.
 
Unless you evolve some new technology to replace your eyes then "a" is
irrelevant. However, if you are going to save without making any
adjustments at all then do so at the highest bit depth available.

I meant the technology for making image adjustments. <shrug>

And yes, that is why I save them at the highest bit depth - it makes
no sense to "degrade" your "negative".
 
Kennedy said:
I have built systems (monochrome as it happens, but that shouldn't
influence the result) with 12-bit ADCs on the video channel output. The
difference cannot be perceived, but it is something the marketing folks
like to exploit. ;-)

I believe, humans should never be limited to what dumb computers
typically provide.
My eyesight (and/or hearing) isn't digital and never will be.

:-)

Uni
 
Are you scanning medium format? I wonder if you are getting any additional
resolution over, say, a 25 or 50 meg scan. Have you experimented and can
you see the difference on your prints? You may even be losing sharpness by
scanning at too high a resolution. Scanning at a high resolution introduces

My experience has been the higher resolutions such as 4000 Vs 2000 can
introduce a grain effect, while the lower resolution gives a softer
image with less detail.

I had some aerial photos of the big bridge between the lower and upper
peninsulas of Michigan. I could do far more to the lower resolution
scan before the grain effect turned up.

It doesn't matter if it's E6 or Kodachrome, that grain effect is there
and annoying as can be.
softness, which you must then compensate for by sharpening.

But back to your question. Volumes have been written on the topic of 8 bits
versus 16, and I have contributed some bulk to that discussion.

Scanning at 16 bits on the LS5000 ED appears to give a better image
than at 8 bits when it comes to post processing. This was readily
apparent in some extreme enlargements of tiny parts of the photograph.
The grain effect was noticeably less at 16 bits.

That is just my experience from some experimenting today. I normally
scan at 8 bits except for some problem slides.

Were I scanning just my present day work I'd scan at 16 and then go to
8, but with the volume and age of the slides I'm doing it's strictly 8
bit and those create 66 meg un-cropped images and about 53 cropped to
a rectangle (get rid of the round corners of the slide mounts). 16
bit is twice the 8 bit or 106 to 132 megs each.
My personal conclusion is that 8 bits per channel is plenty for today's
technology, and the evidence I offer is that (for a gamma 1.8 or greater
image) it is impossible to tell by looking, and looking, after all, is what
we do with photographs.

But there are those for whom that argument is not convincing, and the act of
throwing away any image data is not something they can justify. Whether I
agree with the technical reasons for this extra data, I have to say many of
these people do create better photographs and prints than I do.

So, pick which side of the fence you want to be on. Above all keep your
originals in a safe place - scanners can only continue to get better and
better.

I figure at their age a lot of the originals will be useless in a few
more years.

Roger Halstead (K8RI & ARRL life member)
(N833R, S# CD-2 Worlds oldest Debonair)
www.rogerhalstead.com
 
Bart van der Wolf said:
Just guessing but, if "Hectate" was thinking of a VueScan Raw (64-bits), it
is possible to benefit from e.g. improvements that Ed Hamrick makes on his
IR-cleaning method, and he's working on his "Curves adjustment option". It
happened in the past, IR-cleaning improved and all I had to do was rerun VS
on the Raw file and never had to touch the film untill I got a better
scanner.
I am also thinking about improved tonescaling or High Dynamic Range
compression based on Raw scan data.

But you are right, once adjustments have been applied, and no significant
new adjustments are anticipated, there's little benefit in keeping more than
24-bits color.
Yes - that is specifically why I did not respond to Hectate's reason
"b".
 
Hecate said:
I meant the technology for making image adjustments. <shrug>
The latest couple of versions of Photoshop have all of the processing
options for 16bpc that I believe I need. Sure, there are some plug-ins
available which automate some of those functions for pulling detail out
of the deep shadows, but the capability is there in the application
itself if you are prepared to do it. The only improvement likely in the
technology is true 16bpc processing as opposed to PS's 15-bit
approximation, but 1 bit isn't gonna make a great deal of difference.
 
Back
Top