Back to Scanners and Comparisons.
Gregory Blank
food for thought.
http://largeformatphotography.info/scan-comparison/#
--
"To announce that there must be no criticism of the President,
or that we are to stand by the President, right or wrong,
is not only unpatriotic and servile, but is morally treasonable
to the American public."--Theodore Roosevelt, May 7, 1918
www.gregblankphoto(dot)com
Neil Gould
> Here is an excellent comparison of scanners, which I did not know
> existed food for thought.
>
> http://largeformatphotography.info/scan-comparison/#
>
Thanks for the link, Greg! I like the way the comparisons are set up... it
is quite enlightening, even if I'd like to have more details about who and
how the scans were done, especially the Scitex vs. Tango versions.
Neil
Gordon Moat
Both the Scitex and the Tango are quite old. The one that surprised me
is the Polaroid, since I would not have thought that old a scanner could
still hold up so well. The ICG 350 is still quite good, though their
newer products have moved onwards and upwards. However, I think the
f19.0 aperture used for the original film image might be more of a
limitation than the capabilities of either the Heidelberg Tango or ICG
350. That is not to state that the scan samples from those two are
really quite good, as should be expected considering the pricing.
Ciao!
Gordon Moat
A G Studio
<http://www.allgstudio.com>
Neil Gould
>
> Both the Scitex and the Tango are quite old. The one that surprised me
> is the Polaroid, since I would not have thought that old a scanner
> could still hold up so well. The ICG 350 is still quite good, though
> their newer products have moved onwards and upwards. However, I think
> the f19.0 aperture used for the original film image might be more of a
> limitation than the capabilities of either the Heidelberg Tango or ICG
> 350. That is not to state that the scan samples from those two are
> really quite good, as should be expected considering the pricing.
>
time to thoroughly "mix and match" the various models, but in the little
time that I spent, the color fidelity and detail of those two scanners
stood out. They were significantly better than the consumer models. It
demonstrates, once again, that there is a lot more information on film
than most people are getting if they aren't using equipment of that
calibre or better.
Regards,
Neil
Noons
> demonstrates, once again, that there is a lot more information on film
> than most people are getting if they aren't using equipment of that
> calibre or better.
Bingo!
Gordon Moat
What I found interesting is that I could actually see differences in the
JPEGs. Obviously, we both know that a JPEG is a poor way to judge how a
printed item will turn out, though the fact that differences are visible
in such a low quality representation of scans is interesting. I have
been mostly opposed to these internet challenges mostly due to the
degradation of images by posting and viewing JPEGs.
Anyway, certainly the Heidelberg Tango more than five years after it
came out is still a good system. Shame that Heidelberg got out of that
market, since I would imagine a replacement would have been even more
developed. At least ICG is still in this market, though I rarely find
places that use these.
The latest Creo scanners are even better than the old EverSmart
represented in the tests. Hopefully Kodak will not stagnate further
development of the current iQSmart and EverSmart products. Probably some
speculation, but I think even the somewhat low end (for high end gear)
iQSmart 1 would do a better job in this test.
I agree that there is lots of information on film, though the problem is
that few people will ever have their films scanned on a high end drum
scanner. Even for commercial work, the Creo flatbed scans are much more
common, more cost effective, and often a faster turnaround. A
professional might consider getting a new Creo iQSmart 1 for under
$9000, but would rarely ever consider an ICG for over $30000.
Gregory Blank
Gordon Moat <mo...@attglobal.net> wrote:
> I agree that there is lots of information on film, though the problem is
> that few people will ever have their films scanned on a high end drum
> scanner. Even for commercial work, the Creo flatbed scans are much more
> common, more cost effective, and often a faster turnaround. A
> professional might consider getting a new Creo iQSmart 1 for under
> $9000, but would rarely ever consider an ICG for over $30000.
>
> Ciao!
>
> Gordon Moat
> A G Studio
> <http://www.allgstudio.com>
Ah humbug- whats another 30k :-)
rafe b
wrote:
>What I found interesting is that I could actually see differences in the
>JPEGs. Obviously, we both know that a JPEG is a poor way to judge how a
>printed item will turn out, though the fact that differences are visible
>in such a low quality representation of scans is interesting. I have
>been mostly opposed to these internet challenges mostly due to the
>degradation of images by posting and viewing JPEGs.
And it's been demonstrated again and again that
your fears and concerns on this score are quite
unfounded. They might be founded if the creator
of the site had been stupid enough to use extreme
JPG compression, but fortunately he knows better.
>I agree that there is lots of information on film, though the problem is
>that few people will ever have their films scanned on a high end drum
>scanner. Even for commercial work, the Creo flatbed scans are much more
>common, more cost effective, and often a faster turnaround. A
>professional might consider getting a new Creo iQSmart 1 for under
>$9000, but would rarely ever consider an ICG for over $30000.
The inability to scan LF film, with the quality it deserves,
with a reasonably priced scanner, is the main reason
my LF kit sits idle at the moment.
You'll notice that the Microtek scan in the comparison
was in fact done by me. I am/was under no delusions
that it could match a Howtek, Tango or ICG drum scan.
The Microtek has since been sold.
Now, if Nikon made a (modern, current) LF film scanner,
I'd snap it up in an instant.
rafe b
www.terrapinphoto.com
Neil Gould
>
> You'll notice that the Microtek scan in the comparison
> was in fact done by me. I am/was under no delusions
> that it could match a Howtek, Tango or ICG drum scan.
> The Microtek has since been sold.
>
software did you use, and what were the settings? It looks like somewhat
radical processing was done.
Neil
rafe b
Microtek's ScanWizard. Nothing special on my part,
but I don't know what Leigh Perry might have done
with the scans.
What is it that looks "radical" to you?
IMO, it's clearly sharper than any of the Epsons, but
not as sharp as any of the drums. It's closest to the
Polaroid 45, which isn't surprising -- the Polaroid is
probably another Microtek design.
rafe b
www.terrapinphoto.com
Neil Gould
> On Tue, 06 Dec 2005 11:13:56 GMT, "Neil Gould"
> <ne...@myplaceofwork.com> wrote:
>
>> Recently, rafe b <rafebATspeakeasy.net> posted:
>>>
>>> You'll notice that the Microtek scan in the comparison
>>> was in fact done by me. I am/was under no delusions
>>> that it could match a Howtek, Tango or ICG drum scan.
>>> The Microtek has since been sold.
>>>
>> I did notice that, and wondered how you did that scan? What scanner
>> software did you use, and what were the settings? It looks like
>> somewhat radical processing was done.
>
> Microtek's ScanWizard. Nothing special on my part,
> but I don't know what Leigh Perry might have done
> with the scans.
>
ScanWizard has some strengths and weaknesses, like any other software. I
don't find the "stock" settings to be all that useful, but mostly because
there aren't any stock settings for some of the film that I shoot. ;-)
> What is it that looks "radical" to you?
>
Comparing the text sample and clothing in the frame below it, the color
balance looks like it has way too much red, and the contrast seems rather
high, even in comparison to the other Microtek i900, which is much softer,
but has a color balance and contrast closer to the other scanners.
> IMO, it's clearly sharper than any of the Epsons, but
> not as sharp as any of the drums. It's closest to the
> Polaroid 45, which isn't surprising -- the Polaroid is
> probably another Microtek design.
>
The Polaroid also looks like it has too much red.
Regards,
Neil
Gregory Blank
rafe b <rafebATspeakeasy.net> wrote:
> On Mon, 05 Dec 2005 11:17:36 -0800, Gordon Moat <mo...@attglobal.net>
> wrote:
<Snip>
>
> The inability to scan LF film, with the quality it deserves,
> with a reasonably priced scanner, is the main reason
> my LF kit sits idle at the moment.
BS-could be, I won't delve too far into personal motivation
because from time to time I shy away from carrying my LF
camera (because yes it can be an ordeal) and this is the MF newsgroup.
However there are other options. One can have the images scanned, or
optically printed you just have to be choosy which one you do. Secondly
one could shoot with a roll film back, it accomplishes two things,
one-you can take many more shots on a journey, and bracket & still have
VC movements.
two-you can scan the originals in your Nikon scanner.
> You'll notice that the Microtek scan in the comparison
> was in fact done by me. I am/was under no delusions
> that it could match a Howtek, Tango or ICG drum scan.
> The Microtek has since been sold.
>
> Now, if Nikon made a (modern, current) LF film scanner,
> I'd snap it up in an instant.
>
>
> rafe b
> www.terrapinphoto.com
Gregory Blank
rafe b <rafebATspeakeasy.net> wrote:
I agree although why you and I write Nikon and convince them
there is a market.
rafe b
"Gregory Blank" <greg@greg_____photo.com> wrote in message
news:greg-3F1779.0...@news.isp.giganews.com...
> BS-could be, I won't delve too far into personal motivation
> because from time to time I shy away from carrying my LF
> camera (because yes it can be an ordeal) and this is the MF newsgroup.
> However there are other options. One can have the images scanned, or
> optically printed you just have to be choosy which one you do. Secondly
> one could shoot with a roll film back, it accomplishes two things,
>
> one-you can take many more shots on a journey, and bracket & still have
> VC movements.
>
> two-you can scan the originals in your Nikon scanner.
I have considered a roll-film-back but in some regards it's the worst of
both cases -- the bulk, weight and inconvenience of LF coupled with the
"smaller" film area of MF. Of course, the one big advantage would have been
to scan the resulting images on the Nikon.
For the nature/landscape photography that I do, most of the camera movements
I need are for focus, not for perspective correction.
In terms of sheer megapixels, I get almost as many from scanning 6x7 on the
Nikon as I used to get from scanning 4x5" film on the Microtek.
Drum-scanning 4x5... I haven't found a place that will do that for less than
$50 for a 2500 or 4000 dpi scan. Optical printing is not an option (for me.)
I keep tabs on eBay auctions of drum scanners, hoping to find one selling at
a decent price and within driving distance. For some odd reason I can't
bring myself to buy an Epson 4990. The Microtek (for which I paid $1K) just
wasn't up to the job.
rafe b
Gregory Blank
"rafe b" <ra...@foobar.com> wrote:
> "Gregory Blank" <greg@greg_____photo.com> wrote in message
> news:greg-3F1779.0...@news.isp.giganews.com...
>
> > BS-could be, I won't delve too far into personal motivation
> > because from time to time I shy away from carrying my LF
> > camera (because yes it can be an ordeal) and this is the MF newsgroup.
> > However there are other options. One can have the images scanned, or
> > optically printed you just have to be choosy which one you do. Secondly
> > one could shoot with a roll film back, it accomplishes two things,
> >
> > one-you can take many more shots on a journey, and bracket & still have
> > VC movements.
> >
> > two-you can scan the originals in your Nikon scanner.
>
> I have considered a roll-film-back but in some regards it's the worst of
> both cases -- the bulk, weight and inconvenience of LF coupled with the
> "smaller" film area of MF. Of course, the one big advantage would have been
> to scan the resulting images on the Nikon.
And the camera movements, and increased sharpness by being able to stop
down to 32 or 45. And don't quote the old then your into lens
diffraction (BS) I know the image is hugely sharper at 45 using my Apo-
Symmar.
>
> For the nature/landscape photography that I do, most of the camera movements
> I need are for focus, not for perspective correction.
Which is very important in some scenarios.
>
> In terms of sheer megapixels, I get almost as many from scanning 6x7 on the
> Nikon as I used to get from scanning 4x5" film on the Microtek.
>
> Drum-scanning 4x5... I haven't found a place that will do that for less than
> $50 for a 2500 or 4000 dpi scan. Optical printing is not an option (for me.)
Like I said, must be choosey.....or have a defined purpose.
> I keep tabs on eBay auctions of drum scanners, hoping to find one selling at
> a decent price and within driving distance. For some odd reason I can't
> bring myself to buy an Epson 4990. The Microtek (for which I paid $1K) just
> wasn't up to the job.
Yeah I know the 4990 looks good on some fronts - and pee poor on others-
I have a huge pile of 35mm I keep telling myself I need to scan and
market. I have thought about the option of buying the dedicated Nikon
35mm scanner for around 975 and the 4990 for MF and LF. I would like to
convert my 8x10 camera to 4x10 and then scan the color films I would be
shooting.
Wanta go halfs on a drum scanner?
rafe b
<greg@greg_____photo.com> wrote:
>Wanta go halfs on a drum scanner?
Possibly... I sometimes wish I'd kept (and tried harder to fix)
the ScanMate that I briefly owned last year.
One thing I know is that I will only buy it locally. I'll have to
see it working first, and bring it home myself. Somewhere
within a day's drive of Boston.
Already, you have the problem (with most drum scanners)
that the cost of maintaining one will be prohibitive. Ie., you
might be able to find a decent used machine for, say, $2K,
but it could cost you that much or more to fix it, first time it
breaks. I heard a figure of $5K for a PMT assembly, for
example. Oh yeah, and figure anywhere from $500 to
$2K for a decent driver, most likely.
They are large, heavy, fragile, precision mechanical-
electrical- optical devices. The ScanMate was a "desktop"
model that still weighed around 80 lbs.
From the cost POV, I'm sure a wet darkroom is much
more cost effective. No quarrel there. I have an old
Omega B22XL in the basement...
David J. Littleboy
"rafe b" <rafebATspeakeasy.net> wrote:
>
> From the cost POV, I'm sure a wet darkroom is much
> more cost effective. No quarrel there. I have an old
> Omega B22XL in the basement...
But that only does up to 6x6...
David J. Littleboy
Tokyo, Japan
rafe b
<dav...@gol.com> wrote:
>
>"rafe b" <rafebATspeakeasy.net> wrote:
>>
>> From the cost POV, I'm sure a wet darkroom is much
>> more cost effective. No quarrel there. I have an old
>> Omega B22XL in the basement...
>
>But that only does up to 6x6...
Yeah, I know. The real question is what the hell
I'll ever do with it. Not sure why I kept it -- mostly
it's a relic, but a very nice piece of engineering
for its day.
There's not much likelihood of my ever setting
up a wet darkroom again.
rafe b
www.terrapinphoto.com
Gregory Blank
"David J. Littleboy" <dav...@gol.com> wrote:
Omega used to make a C760 that covered
6x7,....about 700.--with the color lamp house but
no more, to get 6x7 you would have to buy a 4x5
enlarger, unless you could find a good work C760
which should be rather difficult because the chassis
had a number of mechanical issues.
All told i have about five figures in optical
darkroom equipment. I have a color 4x5 enlarger
dichroic lamp hse, and a Fuijimoto roller color paper
processor and a lot of other stuff. I also have an B&W 8x10
enlarger.
Gregory Blank
rafe b <rafebATspeakeasy.net> wrote:
You should have kept your fathers camera and ditched the enlarger ;^)
less lost space.
Gordon Moat
> On Mon, 05 Dec 2005 11:17:36 -0800, Gordon Moat <mo...@attglobal.net>
> wrote:
>
>
>
>>What I found interesting is that I could actually see differences in the
>>JPEGs. Obviously, we both know that a JPEG is a poor way to judge how a
>>printed item will turn out, though the fact that differences are visible
>>in such a low quality representation of scans is interesting. I have
>>been mostly opposed to these internet challenges mostly due to the
>>degradation of images by posting and viewing JPEGs.
>
>
> And it's been demonstrated again and again that
> your fears and concerns on this score are quite
> unfounded. They might be founded if the creator
> of the site had been stupid enough to use extreme
> JPG compression, but fortunately he knows better.
>
>
Perhaps if you read a bit closer to what I stated; I don't print from
JPEGs, and my concern with image files is their printing quality.
Obviously, storing TIFF files would take up a huge amount of space, and
is not practical for websites. The JPEG algorithm also functions with a
sharpening effect, even at the least compressed settings.
I stand by what I stated: "a JPEG is a poor way to judge how a printed
item will turn out". That has nothing to do with "fear", and I highly
disagree in your assertion that my "concerns" are "unfounded". If you
want to believe otherwise, you are entitled to your opinion.
>
>>I agree that there is lots of information on film, though the problem is
>>that few people will ever have their films scanned on a high end drum
>>scanner. Even for commercial work, the Creo flatbed scans are much more
>>common, more cost effective, and often a faster turnaround. A
>>professional might consider getting a new Creo iQSmart 1 for under
>>$9000, but would rarely ever consider an ICG for over $30000.
>
>
>
> The inability to scan LF film, with the quality it deserves,
> with a reasonably priced scanner, is the main reason
> my LF kit sits idle at the moment.
>
> You'll notice that the Microtek scan in the comparison
> was in fact done by me. I am/was under no delusions
> that it could match a Howtek, Tango or ICG drum scan.
> The Microtek has since been sold.
>
> Now, if Nikon made a (modern, current) LF film scanner,
> I'd snap it up in an instant.
>
Nikon does fairly well with medium format scanners. Of course, that
would mean no larger than 6x9 images, though I don't think that is a bad
use of a view camera.
I actually thought the Polaroid made a very nice compromise. That with
the latest SilverFast Ai might make a good combination. The biggest
problem is that it is a used scanner.
Gordon Moat
>. . . . . .
>
> I keep tabs on eBay auctions of drum scanners, hoping to find one selling at
> a decent price and within driving distance. For some odd reason I can't
> bring myself to buy an Epson 4990. The Microtek (for which I paid $1K) just
> wasn't up to the job.
>
>
A good source for used drum scanners and service:
<http://genesis-equipment.com>
They have a few Howtek models at reasonable prices. The downside of many
older drum scanners is a lack of modern software. That can mean a need
to run a dedicated older computer and OS just for scanning.
rafe b
wrote:
>Perhaps if you read a bit closer to what I stated; I don't print from
>JPEGs, and my concern with image files is their printing quality.
>Obviously, storing TIFF files would take up a huge amount of space, and
>is not practical for websites. The JPEG algorithm also functions with a
>sharpening effect, even at the least compressed settings.
>
>I stand by what I stated: "a JPEG is a poor way to judge how a printed
>item will turn out". That has nothing to do with "fear", and I highly
>disagree in your assertion that my "concerns" are "unfounded". If you
>want to believe otherwise, you are entitled to your opinion.
Hmm.. Your words were: "I have been mostly opposed to these
internet challenges mostly due to the degradation of images
by posting and viewing JPEGs." That's the statement I was
refering to.
Also, you're dead wrong about sharpening being implicit in
"the JPG algorithm."
My contention is that, in a fair test, you would not be able
to distinguish between the same image as TIF vs. high-
quality (low-compression) JPG. Whether you view it on
screen or print it -- makes no difference.
rafe b
www.terrapinphoto.com
Scott W
rafe is right and here are the images to back it up.
These are down sampled to get a good clean image.
http://www.sewcon.com/temp/DSC02676%20sized.tif
http://www.sewcon.com/temp/DSC02676%20sized.jpg
jpeg does no shapening.
Scott
Gregory Blank
"Scott W" <bip...@hotmail.com> wrote:
>
> rafe is right and here are the images to back it up.
> These are down sampled to get a good clean image.
> http://www.sewcon.com/temp/DSC02676%20sized.tif
> http://www.sewcon.com/temp/DSC02676%20sized.jpg
>
> jpeg does no shapening.
>
> Scott
What you and perhaps Rafe are failing to see is that Gordon works with
offset printers, and clients that send cheap low res jpegs they pull
out of their yin yang to save actually hiring someone to re shoot and
supply the printer with a respectable size file. The problem with jpegs
are that you can't manipulate them to a great extent and so therefore
once in the hands of a print shop where they may need to be adjusted for
the requirements of the specific device, they become a problem, do to
that nature. ....that is the more they are opened the quicker they
degrade for printing purposes, yes they may look fine but in print they
will suck. You can perhaps get away with giving people clients lower res
jpegs or and its all well and good until they need a bigger file like a
tif.
So you can look at jpegs all you like on screen and state there's no
issue but i believe your misinformed.
rafe b
<greg@greg_____photo.com> wrote:
>What you and perhaps Rafe are failing to see is that Gordon works with
>offset printers, and clients that send cheap low res jpegs they pull
>out of their yin yang to save actually hiring someone to re shoot and
>supply the printer with a respectable size file. The problem with jpegs
>are that you can't manipulate them to a great extent and so therefore
>once in the hands of a print shop where they may need to be adjusted for
>the requirements of the specific device, they become a problem, do to
>that nature. ....that is the more they are opened the quicker they
>degrade for printing purposes, yes they may look fine but in print they
>will suck. You can perhaps get away with giving people clients lower res
>jpegs or and its all well and good until they need a bigger file like a
>tif.
>
>So you can look at jpegs all you like on screen and state there's no
>issue but i believe your misinformed.
You are asking us to argue with a strawman.
Gordon does not talk about multiple cycles of
JPG encode/decode. Separate issue.
Gordon's argument is that even *one* cycle
of JPG encode/decode produces a result
visibly inferior to TIF. And that is only the case
when high levels of JPG compression are used.
Furthermore, the context of our discussion is
(to use Gordon's phrase) "these internet challenges"
presumably, full-res film scan or digicam output,
and the like (like the stuff on my scan snippets
site, or Leigh Perry's site, which is the subject
of the thread...)
In all such cases, high-quality (low-compression)
JPG settings are used.
It's a silly thing to argue about, really, because it
costs nothing but a bit of time to verify all this for
one's self.
rafe b
www.terrapinphoto.com
Scott W
> In article <1134007422.9...@g44g2000cwa.googlegroups.com>,
> "Scott W" <bip...@hotmail.com> wrote:
> >
> > rafe is right and here are the images to back it up.
> > These are down sampled to get a good clean image.
> > http://www.sewcon.com/temp/DSC02676%20sized.tif
> > http://www.sewcon.com/temp/DSC02676%20sized.jpg
> >
> > jpeg does no shapening.
> >
> > Scott
>
> What you and perhaps Rafe are failing to see is that Gordon works with
> offset printers, and clients that send cheap low res jpegs they pull
> out of their yin yang to save actually hiring someone to re shoot and
> supply the printer with a respectable size file. The problem with jpegs
> are that you can't manipulate them to a great extent and so therefore
> once in the hands of a print shop where they may need to be adjusted for
> the requirements of the specific device, they become a problem, do to
> that nature. ....that is the more they are opened the quicker they
> degrade for printing purposes, yes they may look fine but in print they
> will suck. You can perhaps get away with giving people clients lower res
> jpegs or and its all well and good until they need a bigger file like a
> tif.
>
> So you can look at jpegs all you like on screen and state there's no
> issue but i believe your misinformed.
> --
So I have posted a tiff and jpeg, try to do any edit that you might do
before printing that will show a difference between them. In fact they
are are same to within one level, in other words in the noise.
Scott
rolan...@hotmail.com
I've checked carefully and I do not see any difference in the detail.
Gordon Moat
JPEG is a lossy compression algorithm. It functions by retaining more
information in higher resolution areas, then compressing lower
resolution areas. The effect is that contrast along high resolution
edges can change. Very difficult to see this in smaller images, though
it becomes more noticeable in larger images. This is different from
sharpening as most consider it, though if you recall I stated it has the
effect of sharpening. Follow this next paragraph for an example of this.
Simple test for your two images. Take the JPEG, copy it, then paste it
onto a layer above the JFax TIFF you provided. Then choose Difference as
the layer blending. If they were exactly the same, the result would be
completely black. If it looks that way, then zoom into the image to
enlarge the view, and spots will be apparent. Next flatten the image to
create one layer, which looks like the black spotted image. Then select
Image>Adjust>Threshold and type in "2" as the value. This will give you
a field of white spots on a black background. If you view that at
exactly 100% magnification, and compare to the originals, you will see
the ghost image of the ship and rigging. Those spots represent the
changed edge contrast near the high resolution parts.
Okay, so lots of people here probably go "so what" or "who cares". If
you only print inkjet or some of the newer LightJet, Chromira or similar
digital/chemical processes, you should rarely see a problem, or much
difference; the dot gain will usually mask any small aberrations. This
also assumes you would only work with RGB files, letting the printing
software convert to CMYK (or CcMmYk, or other) instead of doing your own
CMYK files.
Gregory Blank got it exactly right. The difference becomes more
noticeable when handling files for offset printing. This is even more
true as the printed sizes grow larger. Do you have to look really close
to find it . . . sure, and therein might be the trouble . . . probably
many just won't look that close, or will accept any printed difference.
All of us have stated it many times, the technology, technique and
resolution will become less important when the image is of a compelling
enough nature.
Scott W
>
> JPEG is a lossy compression algorithm. It functions by retaining more
> information in higher resolution areas, then compressing lower
> resolution areas. The effect is that contrast along high resolution
> edges can change. Very difficult to see this in smaller images, though
> it becomes more noticeable in larger images.
samples are 920 x 720, you trying to tell me that if these were larger
I could see a differance?
> Simple test for your two images. Take the JPEG, copy it, then paste it
> onto a layer above the JFax TIFF you provided. Then choose Difference as
> the layer blending. If they were exactly the same, the result would be
> completely black. If it looks that way, then zoom into the image to
> enlarge the view, and spots will be apparent. Next flatten the image to
> create one layer, which looks like the black spotted image. Then select
> Image>Adjust>Threshold and type in "2" as the value. This will give you
> a field of white spots on a black background. If you view that at
> exactly 100% magnification, and compare to the originals, you will see
> the ghost image of the ship and rigging. Those spots represent the
> changed edge contrast near the high resolution parts.
You are talking about a change of 2 levels, and if you look at the
differance you
will see that what you are looking at is noise, no edge detail.
> Okay, so lots of people here probably go "so what" or "who cares". If
> you only print inkjet or some of the newer LightJet, Chromira or similar
> digital/chemical processes, you should rarely see a problem, or much
> difference; the dot gain will usually mask any small aberrations. This
> also assumes you would only work with RGB files, letting the printing
> software convert to CMYK (or CcMmYk, or other) instead of doing your own
> CMYK files.
>
> Gregory Blank got it exactly right. The difference becomes more
> noticeable when handling files for offset printing. This is even more
> true as the printed sizes grow larger. Do you have to look really close
> to find it . . . sure, and therein might be the trouble . . . probably
> many just won't look that close, or will accept any printed difference.
> All of us have stated it many times, the technology, technique and
> resolution will become less important when the image is of a compelling
> enough nature.
I would be willing to give heavy odds that if printed these two photos
would produce the same print to anybody's eye. But then your point
was that jpeg was so bad that you did not feel you could judge
differences in scanners if all you had to look at was the jpeg, this is
just not true. In fact put each into its own layer and blink between
them at 400% zoom, you can't tell one from the other.
Jpeg can be bad, very bad, if the compression is set too high. But for
a low compression jpeg images there is no visable change in the image,
even if it is a large image.
Scott
Gregory Blank
rafe b <rafebATspeakeasy.net> wrote:
> and the like (like the stuff on my scan snippets
> site, or Leigh Perry's site, which is the subject
> of the thread...)
>
Since I started this thread I am well aware of the subject.
& to correct you its not Leigh Perry's site its -
largeformatphotography.info is operated on a volunteer basis by Q.-Tuan
Luong and Tom Westbrook, using server space generously donated by Brian
Reid, refered to us by Tim Atherton.
Gregory Blank
"Scott W" <bip...@hotmail.com> wrote:
> I would be willing to give heavy odds that if printed these two photos
> would produce the same print to anybody's eye.
Ok. Thats a very general statement and truly open ended.
Gregory Blank
rolan...@hotmail.com wrote:
>
> I've checked carefully and I do not see any difference in the detail.
That was quick for outputting on an Imagesetter.
rolan...@hotmail.com
Gregory Blank wrote:
> In article <1134030831....@g14g2000cwa.googlegroups.com>,
> rolan...@hotmail.com wrote:
>
> >
> > I've checked carefully and I do not see any difference in the detail.
>
> That was quick for outputting on an Imagesetter.
Oh, I left it behind today. How remiss.
I think you are talking BS in this case. I don't see any difference in
the images and I don't believe there will be a difference in the print,
unless the file format is handled differently.
Scott W
> In article <1134040677.6...@g14g2000cwa.googlegroups.com>,
> "Scott W" <bip...@hotmail.com> wrote:
>
> > I would be willing to give heavy odds that if printed these two photos
> > would produce the same print to anybody's eye.
>
> Ok. Thats a very general statement and truly open ended.
> --
The point is that there are people who will say you can't tell
anything looking at jpegs because the jpegs change the image too much.
This can be true but only if the jpeg is poorly done, if the jpeg is
done right then you can't tell the difference between the jpeg and a
tiff.
It seems a bit funny to be worried about the fact that the images are
jpegs but not the fact that they have be down sampled to 2400 ppi. For
a scanner test I would have thought that leaving the images at 4000 ppi
might not have been a bad idea. I have always held that there is not
much detail past 2000 ppi in film, but it seems that if you are going
to give a scanner a work out go for the high numbers.
I know the Nikon SUPER COOLSCAN 9000 ED can get lots of detail in a
4000 ppi scan and it is a CCD scanner. In the end if drum scanners can
really produce high resolution scans then this test did test them close
to their limits.
Scott
Scott
Gregory Blank
rolan...@hotmail.com wrote:
Well there certainly is a big issue I can say from personal printing
that is having my work published. Jpegs do not hold the contrast
within the range a tiff file can. Therefore jpegs are losing information
from the scene.
I also know that opening and closing jpegs multiple times creates issues
whether you say so or not, greater minds than yours and mine have
qualified this.
Gregory Blank
"Scott W" <bip...@hotmail.com> wrote:
> Gregory Blank wrote:
> > In article <1134040677.6...@g14g2000cwa.googlegroups.com>,
> > "Scott W" <bip...@hotmail.com> wrote:
> >
> > > I would be willing to give heavy odds that if printed these two photos
> > > would produce the same print to anybody's eye.
> >
> > Ok. Thats a very general statement and truly open ended.
> > --
>
> The point is that there are people who will say you can't tell
> anything looking at jpegs because the jpegs change the image too much.
> This can be true but only if the jpeg is poorly done, if the jpeg is
> done right then you can't tell the difference between the jpeg and a
> tiff.
Perhaps once the image is successfully printed, my point was that
providing jpegs versus tifs presents issues that a technician would
rather not deal with & a lot of clients have no knowledge of why those
issues should be addressed before hand. Personally i don't have a
working knowledge of every four color press in existence -do you?
Which was my reason for stating the sentence above.
>
> It seems a bit funny to be worried about the fact that the images are
> jpegs but not the fact that they have be down sampled to 2400 ppi. For
> a scanner test I would have thought that leaving the images at 4000 ppi
> might not have been a bad idea. I have always held that there is not
> much detail past 2000 ppi in film, but it seems that if you are going
> to give a scanner a work out go for the high numbers.
I think you should ask the webmaster why they were down sampled.
> I know the Nikon SUPER COOLSCAN 9000 ED can get lots of detail in a
> 4000 ppi scan and it is a CCD scanner. In the end if drum scanners can
> really produce high resolution scans then this test did test them close
> to their limits.
rafe b
> In article <subfp1tidvpldo5up...@4ax.com>,
> rafe b <rafebATspeakeasy.net> wrote:
>
>> and the like (like the stuff on my scan snippets
>> site, or Leigh Perry's site, which is the subject
>> of the thread...)
>>
>
> Since I started this thread I am well aware of the subject.
> & to correct you its not Leigh Perry's site its -
>
> largeformatphotography.info is operated on a volunteer basis by Q.-Tuan
> Luong and Tom Westbrook, using server space generously donated by Brian
> Reid, refered to us by Tim Atherton.
The scanner comparison effort was organized by
Leigh Perry and the scans provided by volunteers
that he recruited. The scanned image, the presentation
format, etc. -- are Leigh's.
Incidentally, the results are now being hosted as part
of the largeformatphotography site.
While I was corresponding with Leigh and doing
my scan(s) for him, he was hosting the web results
independently.
The main thing interesting about Leigh's effort is that
it focuses on scanners for LF film -- drum scanners,
Creo-Scitex, Fuji, and at the low end, Epson/Canon.
In that regard, Leigh's site is a bit off-topic on rpe-MF.
Indeed, for MF, one of the best scanners has been
excluded from consideration.
rafe b
www.terrapinphoto.com
rafe b
> Well there certainly is a big issue I can say from personal printing
> that is having my work published. Jpegs do not hold the contrast
> within the range a tiff file can. Therefore jpegs are losing information
> from the scene.
>
> I also know that opening and closing jpegs multiple times creates issues
> whether you say so or not, greater minds than yours and mine have
> qualified this.
These observations are irrelevant to the validity of the
scan-comparison site that is the subject of this thread.
Being the OP, you should understand that.
Opening/closing a JPG can be done an infinite number
of times without degrading the file. If that weren't so,
the web as we know it could not exist.
It's the JPG encoding process that loses data, but that's
quite independent of file-open or file-close.
rafe b
www.terrapinphoto.com
Gregory Blank
"rafe b" <ra...@foobar.com> wrote:
Never the less- appropriate to the previous discussions and
when have you been terribly concerned with how relevant
a thread is to rec.Mf? :-)
Gregory Blank
"rafe b" <ra...@foobar.com> wrote:
> "Gregory Blank" <greg@greg_____photo.com> wrote in message
> news:greg-1F4410.1...@news.isp.giganews.com...
>
> > Well there certainly is a big issue I can say from personal printing
> > that is having my work published. Jpegs do not hold the contrast
> > within the range a tiff file can. Therefore jpegs are losing information
> > from the scene.
> >
> > I also know that opening and closing jpegs multiple times creates issues
> > whether you say so or not, greater minds than yours and mine have
> > qualified this.
>
>
>
> These observations are irrelevant to the validity of the
> scan-comparison site that is the subject of this thread.
> Being the OP, you should understand that.
I do understand my statement thanks for pointing it out -
however, you and others seem to not fathom the issues
behind the simple statement that one cannot judge the differences
between a jpeg and a tiff based on website viewing and those issue apply
to printing four color image using offset printers. However that is not
to saying that relative comparisons between scanners cannot be made
using jpegs.
> Opening/closing a JPG can be done an infinite number
> of times without degrading the file. If that weren't so,
> the web as we know it could not exist.
>
> It's the JPG encoding process that loses data, but that's
> quite independent of file-open or file-close.
I sense a contradiction.
Scott W
> I do understand my statement thanks for pointing it out -
> however, you and others seem to not fathom the issues
> behind the simple statement that one cannot judge the differences
> between a jpeg and a tiff based on website viewing and those issue apply
> to printing four color image using offset printers. However that is not
> to saying that relative comparisons between scanners cannot be made
> using jpegs.
>
tell if a given scan of a photo will make a good print. You are not
trying to say that the only why to tell is to print it are you? If
this were the case then we would be in a rather uncomfortable position
of not knowing how our magazine, ad or what ever is going to look until
it is too late.
So how do you tell? The whole print industry is based on evaluating
images in there digital form before sending them off to be printed. I
would be interesting to hear just how you evaluate a tiff to tell how
well it will print.
Scott
rafe b
>> Opening/closing a JPG can be done an infinite number
>> of times without degrading the file. If that weren't so,
>> the web as we know it could not exist.
>>
>> It's the JPG encoding process that loses data, but that's
>> quite independent of file-open or file-close.
>
> I sense a contradiction.
No contradiction. Just being careful about terminology.
You and Gordon are raising issues (four-color offset
printing, etc.) that are way outside the given topic of
"Scanners and Comparisons."
You open a JPG file whenever you view a JPG image.
You close a JPG file whenever you select another image
to view, or close the viewing application. Neither of
these operations affects or degrades the file.
These operations occur whenever we view any of the
scan samples on the scan-comparison site that we're
discussing. Indeed, they happen whenever we view
any JPG image on any web site.
rafe b
www.terrapinphoto.com
Gregory Blank
"Scott W" <bip...@hotmail.com> wrote:
Well first off unless I am the one doing the final color management
which I am usually not, I would stick with Abode 1998 as a profile.
Secondly I would do color correction and as minimal sharpening
as possible and I would adjust contrast with levels and curves to look
good on my screen, realizing the art dept probably will make some
adjustments to suit their tastes or requirements.
Since most publications adjust the imagery anyway.
Now if I was responsible for the whole color management process
I would either work directly with the printer and request a proof, or
I might say I think I know how its going to look simply based on
a supplied profile from the print company by looking at the CMYK
conversion on my humble screen- if I was still uncertain as in the case
where I was spending a huge some of money I would request a proof to
confirm my perceptions. This just to use basic procedures, but what I
have always wanted to learn is the whole soft-proofing PDF sequencing of
desktop to final print. I also have always wanted to work in the sheet
feed higher end printing industry.....but it may be a little late in my
life for that.
& You could also send the printer if working directly an original (film)
or a guide print to match and let them use their scanner.
Lastly I would supply a file double the requirement as a safe guard
in case the print shop has to re spec my job.
Scott W
> Well first off unless I am the one doing the final color management
> which I am usually not, I would stick with Abode 1998 as a profile.
> Secondly I would do color correction and as minimal sharpening
> as possible and I would adjust contrast with levels and curves to look
> good on my screen, realizing the art dept probably will make some
> adjustments to suit their tastes or requirements.
>
> Since most publications adjust the imagery anyway.
>
> Now if I was responsible for the whole color management process
> I would either work directly with the printer and request a proof, or
> I might say I think I know how its going to look simply based on
> a supplied profile from the print company by looking at the CMYK
> conversion on my humble screen- if I was still uncertain as in the case
> where I was spending a huge some of money I would request a proof to
> confirm my perceptions. This just to use basic procedures, but what I
> have always wanted to learn is the whole soft-proofing PDF sequencing of
> desktop to final print. I also have always wanted to work in the sheet
> feed higher end printing industry.....but it may be a little late in my
> life for that.
>
> & You could also send the printer if working directly an original (film)
> or a guide print to match and let them use their scanner.
>
> Lastly I would supply a file double the requirement as a safe guard
> in case the print shop has to re spec my job.
> --
The point is you can pretty much tell is a given file will be suitable
for printing, by looking at it on your screen. Sure it is posible to
slip up between the photo and the offset press but that is going to be
an issue of color managment and not so much the content of the image.
To say that you can't tell by looking at a jpeg image how suitable the
scanner is for creating prints would require that you can in fact see a
differance between the jpeg image and the tiff of the same image.
Gordon would have us believe that jpegs do more then just lose
information, he claims that they also do effectively sharpening of the
image. If this were the case should not a jpeg look different, even
when view on a screen.
Scott
Gregory Blank
"Scott W" <bip...@hotmail.com> wrote:
Scott I can't speculate for Gordon, I know there are issues in four
color printing "sometimes" the equipment is beyond what I can afford
and I can't say what one can see on a high end monitor. I do however
Look at what the print companies tell me and what Gordon says
corresponds with that.
Gregory Blank
"rafe b" <ra...@foobar.com> wrote:
> No contradiction. Just being careful about terminology.
>
> You and Gordon are raising issues (four-color offset
> printing, etc.) that are way outside the given topic of
> "Scanners and Comparisons."
>
> You open a JPG file whenever you view a JPG image.
> You close a JPG file whenever you select another image
> to view, or close the viewing application. Neither of
> these operations affects or degrades the file.
>
> These operations occur whenever we view any of the
> scan samples on the scan-comparison site that we're
> discussing. Indeed, they happen whenever we view
> any JPG image on any web site.
>
>
> rafe b
> www.terrapinphoto.com
I think what we are talking about is relatively minor for most people,
but I do know the talk at last years Fuji-film demo of digital file prep
for lab printing did cover the opening and closing file issue and
supported my stance, and that stance has been carried through the photo
industry for quite sometime are you saying its blatantly wrong?
Lets suppose this, a file was never opened but saved to a CD media
that only allowed files to be read not saved over and a CD of course
that had unlimited years to exist before the media might have otherwise
degraded. Now let's look at opening & closing a jpeg from the same file
on a disk subject to fragmentation once the image is worked on and saved
4 times at its highest compression value. Do you believe that image to
be of equal quality to say a tif of the same original file that was
equally manipulated?
rafe b
> I think what we are talking about is relatively minor for most people,
> but I do know the talk at last years Fuji-film demo of digital file prep
> for lab printing did cover the opening and closing file issue and
> supported my stance, and that stance has been carried through the photo
> industry for quite sometime are you saying its blatantly wrong?
We're drifting further and further afield from the
topic of "Scanners and Comparisons," aren't we?
I'm not privy to what went on in the Fuji-film demo
that you're refering to, but I can only guess that you
either mis-heard or they were trying to sell so some
nice $$ Fuji-made backup system or service.
If files could be degraded simply from being read,
the entire microcomputer industry would collapse
within minutes, as would the Web and most of
the smarter electrical appliances in your home.
We know media can go bad (hard disks, CDs,
DVDs, etc.) but for the purposes of this discussion
the issue is irrelevant.
I'm saying that you can go to the URL in the OP's
post (oops, that's you) and view the images as
often as you like, and they will appear identically
each and every time -- asssuming the server's
hardware remains intact.
Fact: except in pathological cases (viz., media
failure) reading a file does not affect the file.
rafe b
www.terrapinphoto.com
Gordon Moat
I thought it might be good to add some comments.
Gregory Blank wrote:
> In article <1134066857....@g43g2000cwa.googlegroups.com>,
> "Scott W" <bip...@hotmail.com> wrote:
>
>
>>Gregory Blank wrote:
>>
>>>I do understand my statement thanks for pointing it out -
>>>however, you and others seem to not fathom the issues
>>>behind the simple statement that one cannot judge the differences
>>>between a jpeg and a tiff based on website viewing and those issue apply
>>>to printing four color image using offset printers. However that is not
>>>to saying that relative comparisons between scanners cannot be made
>>>using jpegs.
>>>
>>
>>So now we come to an interesting question, how do you Gregory Blank
>>tell if a given scan of a photo will make a good print. You are not
>>trying to say that the only why to tell is to print it are you? If
>>this were the case then we would be in a rather uncomfortable position
>>of not knowing how our magazine, ad or what ever is going to look until
>>it is too late.
>>
>>So how do you tell? The whole print industry is based on evaluating
>>images in there digital form before sending them off to be printed. I
>>would be interesting to hear just how you evaluate a tiff to tell how
>>well it will print.
>>
>>Scott
>
>
> Well first off unless I am the one doing the final color management
> which I am usually not, I would stick with Abode 1998 as a profile
This is not a bad choice if you are doing so on the basis of
consistency. You may want to investigate using ProPhotoRGB instead, if
you are delivering RGB files. This choice gives more room for
adjustments to conversion onwards to CMYK. However, if you are not sure
of how your files will be handled going to CMYK, then you might be
better sticking with AdobeRGB.
.
> Secondly I would do color correction and as minimal sharpening
> as possible and I would adjust contrast with levels and curves to look
> good on my screen, realizing the art dept probably will make some
> adjustments to suit their tastes or requirements.
>
> Since most publications adjust the imagery anyway.
Quite true, but they don't always do a proper job of it. The better
choice in this case is delivering CMYK files, though that involves more
work on your part, and the need for more information. SWOP Offset
Sheetfed is one choice in CMYK which is a good compromise, though many
use one of the SWOP Web Offset choices. A sheetfed press offers some
better paper choices than a web press, so results can look better. The
idea of using a slightly better CMYK choice is better end printed
result. The ideal is to get the CMYK profile settings, and use a custom
space with those input parameters within your image editing software.
>
> Now if I was responsible for the whole color management process
> I would either work directly with the printer and request a proof, or
> I might say I think I know how its going to look simply based on
> a supplied profile from the print company by looking at the CMYK
> conversion on my humble screen- if I was still uncertain as in the case
> where I was spending a huge some of money I would request a proof to
> confirm my perceptions. This just to use basic procedures, but what I
> have always wanted to learn is the whole soft-proofing PDF sequencing of
> desktop to final print. I also have always wanted to work in the sheet
> feed higher end printing industry.....but it may be a little late in my
> life for that.
The standard is still a hard copy proof. Soft proofing is okay in some
instances, such as anything on newsprint. Soft proofing is sometimes
used when time constraints, or lack of proper hard proofs is available.
The best hard copy proof is a press proof, though it slows down the
printing process and costs more.
>
> & You could also send the printer if working directly an original (film)
> or a guide print to match and let them use their scanner.
There are a few choices in SWOP certified inkjet proofing systems. These
are now reasonable choices, with prices on the HP 130nr and Epson 4000
(4800) at attainable levels. You can check on <http://www.swop.org> to
find out which printer, RIP, and papers combination is certified.
SWOP is a minimum standard, and the majority of print shops will try to
do better than SWOP. The upcoming and evolving GraCOL specifications aim
for even better outputs.
>
> Lastly I would supply a file double the requirement as a safe guard
> in case the print shop has to re spec my job.
There are a few approved soft proof monitors, though they are not
without problems. EIZO makes one system, and that can be matched with a
viewing system from Kodak (or a few other companies). While these can
save time in production environments, they are not as good as a hard
copy proof, and the entire system actually costs more than a few of the
mid range inkjet choices from HP or Epson.
Another option with delivering RGB is to make a separate soft proof
image using ColorMatchRGB. That space choice use to be for the old
Radius PressView monitors. The new idea now is to roll your own CMYK
conversion, then take that CMYK result and make an RGB file from that,
then use Profile to Profile in PhotoShop to create a ColorMatchRGB image
file. The idea is that while the ColorMatchRGB file might not look as
good as an sRGB or AdobeRGB image, it should look at least as good (or
slightly worse) than a print out. When the final print looks better than
a soft proof, then you are playing it safe. This concept comes from Seth
Resnick, founder of D-65, long time photographer, and digital guru.
The other soft proof of using a PDF works well in some situations. If
nothing else, the PDF proof is the best soft proof choice when needing
to show how text or vector graphics appear in a page layout. Choose
Device Independent in your settings to play it safe.
Gregory Blank
"rafe b" <ra...@foobar.com> wrote:
You 'Could" just answer the question I posed instead of rerouting it.
You misread it..... here it is once more, read it slowly-then develop a
precise two or three letter response.
Gregory Blank
Gordon Moat
There are currently some newer slightly competing views in the industry.
What is muddying the waters a bit is delivery of RGB files, and a huge
lack of understanding about CMYK from photographers doing their own
digital file preparation (either capture, or scanned film).
Andrew Rodney and Seth Resnick offer some opposing view, and a
perspective from the photographic industry. I suggest reading more from
each of these two people, then formulating your own methods. There is a
bit of reality and some good concepts from both, though I tend to think
that what Seth Resnick <http://www.sethresnick.com> writes is more in
line with better and consistent end printed results. Probably more
people on this newsgroup would like (or agree) with the views of Andrew
Rodney <http://www.digitaldog.com>.
JPEG is a bad choice, and with fairly cheap storage and fast electronic
file delivery rarely an issue, I don't see why people are still using
JPEGs for anything outside of website display. I had one client in the
past year who insisted on me delivering four images as JPEG files for
printing in their publication. I was actually pissed about the end
printing quality, since I felt it unfairly represented my photography.
I have been delivering print ready CMYK files since 1996, with almost no
problems. In fact, the only issue I had was one press operator giving me
a high total ink number (intended for one paper), then when the job went
to print the paper was switched without an adjustment to the total ink
level. The result that GCR caused was grey splotches within parts of the
dark background. Sure, I was unhappy about that, though the client did
not notice it until I pointed it out, and then the client stated it was
no big deal and still thought the print was nice. Lesson learned here is
that "good enough" is the target to achieve, and playing it safe on the
settings would have been fine.
If Scott is running one of the latest EIZO monitors, a SWOP certified
viewing system, and doing soft proofs that way, then maybe even then the
difference might be so slight that it might not matter. I suspect none
on this newsgroup are running a high end soft proofing system. No matter
how good we think our current computer monitor might be, the reality is
that the on screen image will not be a good representation of offset
printed results. Sure, we still should run a calibrated monitor, and we
still should be using the colour picker tool (eyedropper) in PhotoShop
to check the values on our images. Do those things, and you can get
"good enough" results 90% of the time.
Please understand that when I do this for work, I cannot afford to be
wrong on image files 10% of the time, so I need to do More. I wish that
everyone doing graphic design put as much effort into getting good
prints, but they don't. Places like to hire temp workers who sort of
know graphic design and software, but are often rushed for time or just
don't know how to do things properly. Delivering press ready CMYK files
can avoid some of the problems.
Gordon Moat
I think it was Seth Resnick who gave a nice example of this. In his
example, all he did was open and then save again the same JPEG file.
When he got close to 100 saves, he got an End of File error message, and
was unable to re-open the JPEG.
What I have not heard of was server mirroring and back-ups causing
degradation of JPEG files. It might be possible, but I think it might
take much more than 100 transfers. I think a separate issue is how many
saves/opens/transfers or other actions it would take to actually see a
difference on a computer monitor. My guess would be that the number
would depend upon the level of compression within the JPEG file.
Gordon Moat
> Gordon Moat wrote:
> >
>
>>JPEG is a lossy compression algorithm. It functions by retaining more
>>information in higher resolution areas, then compressing lower
>>resolution areas. The effect is that contrast along high resolution
>>edges can change. Very difficult to see this in smaller images, though
>>it becomes more noticeable in larger images.
>
> First off this is not how jpeg does it compression, and second my
> samples are 920 x 720, you trying to tell me that if these were larger
> I could see a differance?
>
>
>>Simple test for your two images. Take the JPEG, copy it, then paste it
>>onto a layer above the JFax TIFF you provided. Then choose Difference as
>>the layer blending. If they were exactly the same, the result would be
>>completely black. If it looks that way, then zoom into the image to
>>enlarge the view, and spots will be apparent. Next flatten the image to
>>create one layer, which looks like the black spotted image. Then select
>>Image>Adjust>Threshold and type in "2" as the value. This will give you
>>a field of white spots on a black background. If you view that at
>>exactly 100% magnification, and compare to the originals, you will see
>>the ghost image of the ship and rigging. Those spots represent the
>>changed edge contrast near the high resolution parts.
>
>
> You are talking about a change of 2 levels, and if you look at the
> differance you
> will see that what you are looking at is noise, no edge detail.
>
So, just a guess here, but you did not follow those steps and view the
results? The "noise" is in a pattern that very much matches the rigging
on the ship. If you don't see it, then I guess the difference between
JPEG and TIFF should not matter to you; and you can just use JPEGs to
keep your file sizes and storage requirements lower.
Perhaps you have another explanation of why the data is different
between the JPEG and the TIFF. The use of Threshold was only to make the
image "noise" higher contrast to make it easier to see. If you instead
did a convert to greyscale step, you should still see an outline rough
shape of the ship and rigging, but it is much more difficult to make
that out.
>
>
>>Okay, so lots of people here probably go "so what" or "who cares". If
>>you only print inkjet or some of the newer LightJet, Chromira or similar
>>digital/chemical processes, you should rarely see a problem, or much
>>difference; the dot gain will usually mask any small aberrations. This
>>also assumes you would only work with RGB files, letting the printing
>>software convert to CMYK (or CcMmYk, or other) instead of doing your own
>>CMYK files.
>>
>>Gregory Blank got it exactly right. The difference becomes more
>>noticeable when handling files for offset printing. This is even more
>>true as the printed sizes grow larger. Do you have to look really close
>>to find it . . . sure, and therein might be the trouble . . . probably
>>many just won't look that close, or will accept any printed difference.
>>All of us have stated it many times, the technology, technique and
>>resolution will become less important when the image is of a compelling
>>enough nature.
>
>
> I would be willing to give heavy odds that if printed these two photos
> would produce the same print to anybody's eye.
Which is why I stated that it is "very difficult to see this in smaller
images". Basically, you answered your own question . . . many people
just will not see a difference, or if there is a difference it might not
be important enough to worry about it.
But then your point
> was that jpeg was so bad that you did not feel you could judge
> differences in scanners if all you had to look at was the jpeg, this is
> just not true. In fact put each into its own layer and blink between
> them at 400% zoom, you can't tell one from the other.
Which tells me what about printed output on a sheetfed press?
Hey, I don't want to convert your working habits. If you have a system
you like using, and you are happy / satisfied / ecstatic about the
results, then what reason would you have to change anything?
What is the best colour space choice for internet file display: sRGB
What is one of the worst choices for printed file outputs: sRGB
What colour space choice is used on most internet images: sRGB
Of course, colour issues are substantially more subjective, and even
tougher to compare than resolution issues. Viewing at any percentage, on
most computer monitors, there would be nearly no easily discernible
difference in colour . . . so I guess that makes it a non-issue for most
people judging images on their computer monitors. Seems to me like there
should be nothing else to discuss.
>
> Jpeg can be bad, very bad, if the compression is set too high. But for
> a low compression jpeg images there is no visable change in the image,
> even if it is a large image.
>
Hey, if you say so . . . then okay. If it is not visible to you, or if
it does not matter to you, then why argue the point? Be happy with your
workflow, and just ignore what I say, since what I write probably would
not apply to your work, nor would it have much impact on what you are
now doing.
rafe b
"Gregory Blank" wrote:
> You 'Could" just answer the question I posed instead of rerouting it.
> You misread it..... here it is once more, read it slowly-then develop a
> precise two or three letter response.
>
> Now let's look at opening & closing a jpeg from the same file
> on a disk subject to fragmentation once the image is worked on and saved
> 4 times at its highest compression value. Do you believe that image to
> be of equal quality to say a tif of the same original file that was
> equally manipulated?
Of course not. I simply don't enjoy responding
to totally irrelevant hypotheticals.
rafe b
www.terrapinphoto.com
rafe b
"Gordon Moat" <mo...@attglobal.net> wrote >
> What I have not heard of was server mirroring and back-ups causing
> degradation of JPEG files. It might be possible, but I think it might take
> much more than 100 transfers. I think a separate issue is how many
> saves/opens/transfers or other actions it would take to actually see a
> difference on a computer monitor. My guess would be that the number would
> depend upon the level of compression within the JPEG file.
Aside from sharing small images on the web and
via email and the like, JPG is a reasonable format
for images *in their final state.*
It is obviously not a good format for images still
in-progress.
My archiving regimen for both scans and digicam
captures is to save TIFs and RAW files on multiple
backup media (eg., DVDs.) The corresponding TIFs
on my hard drive are then converted to best-quality
JPGs and left on the hard drive.
In all but the most critical cases, printing is done
from these JPGs. If it's an image that I intend to
work on over a span of time (eg., several evenings)
then I'll fetch the TIF from the archives save as TIF
until I'm done.
It's not like you and Gregory have told us anything
new. Yes, one can misuse JPGs. So what?
My point was and remains this: a single instance
of JPG encoding, using "high quality" settings,
results in an image that is indistiguishable by eye
from the corresponding TIF, in print or on-screen.
rafe b
www.terrapinphoto.com
Scott W
snip A whole bunch of stuff dealing wtih color.
Seems to me like a bit of a switch of topic here. What does color
space have to do with the differences between jpeg and tiff, either one
can be in sRGB and either one can be in Adobe 1998 RGB.
But you were not talking about the color space, the scanner test in
fact does not cover color and said so up front. What the test is
covering is detail captured, you position is that jpeg degrades the
quaility to such a degree that you can't really tell about detail when
viewing a jpeg image, this is clearly not the case.
I don't use tiffs in my workflow, I don't know why people do, I use psd
files, I find them a bit more to my liking. I do use jpeg when showing
a result of an image, to use a psd or tiff would not add anything at
all.
Scott
Q.G. de Bakker
> Seems to me like a bit of a switch of topic here. What does color
> space have to do with the differences between jpeg and tiff, either one
> can be in sRGB and either one can be in Adobe 1998 RGB.
Yes.
And "TIFF" is about tags with additional information 'packaging' the image
data. And the image data may be in one of many forms, included
jpeg-compressed.
So a TIFF-file may well be a 'lossy-compressed' JPEG-image (with some added
extras) too.
Q.G. de Bakker
> I don't use tiffs in my workflow, I don't know why people do, I use psd
> files, I find them a bit more to my liking. I do use jpeg when showing
> a result of an image, to use a psd or tiff would not add anything at
> all.
That's not true.
JPEG is a compression method that does indeed causes lots of artifacts to
appear, degrading image quality. By how much depends on the compression
settings, but it's always there.
What matters, of course, is how obvious the degradation is. In a visual
medium, what you do not see is not there. Right? ;-)
So a carefully monitored JPEG-compression, tailored to the end use (mainly
displaying on computer screens) may be quite acceptable.
TIFFs and PSDs can contain much more than JPEGs. So while JPEGs may be good
as an final-format, they are utterly useless as long as the 'workflow' is
still flowing.
And that's not even considering that every time you save a reworked image,
(more) quality is thrown away.
TIFFs are often preferable to PSDs, because, they are more universally
'understood'. Just more cross-platform and cross-application 'compatible'.
David J. Littleboy
"Gregory Blank" <greg@greg_____photo.com> wrote:
>
> I think what we are talking about is relatively minor for most people,
Bingo!
What's going on here is that there are enormous and obvious differences
between different formats and between crops of scanned and digital originals
with the same number of pixels, but certain people refuse to admit that
those differences exist and sidetrack the discussion to extremely minor
issues.
The difference between 645 and 6x7 is large, a lot larger than the
difference between 645 scanned with a Nikon 8000 and 645 scanned with the
world's greatest drum scanner. There's no way to make a 16x20 from 645 that
looks as good as a 16x20 made from 6x7, Provia 100F, and a Nikon 8000.
The claim that you can't evaluate the differences on the screen or with
jpegs is completely ridiculous; the differences are major.
> but I do know the talk at last years Fuji-film demo of digital file prep
> for lab printing did cover the opening and closing file issue and
> supported my stance, and that stance has been carried through the photo
> industry for quite sometime are you saying its blatantly wrong?
Basically, yes. The problem with jpeg is that if some application in your
workflow defaults to anything lower than Photoshop's best quality jpeg, your
image gets trashed. That's a valid thing to worry about, but it has nothing
to do with whether Photoshop's best quality jpeg is at all problematic,
which it's not. Or whether even some lower quality jpeg is more than
adequate to see the differences between scanners.
> Lets suppose this, a file was never opened but saved to a CD media
> that only allowed files to be read not saved over and a CD of course
> that had unlimited years to exist before the media might have otherwise
> degraded. Now let's look at opening & closing a jpeg from the same file
> on a disk subject to fragmentation once the image is worked on and saved
> 4 times at its highest compression value. Do you believe that image to
> be of equal quality to say a tif of the same original file that was
> equally manipulated?
Basically, yes. Jpeg can represent a subset of images that tiff can
represent, so once you save as jpeg, opening and resaving makes either no
changes to the file or changes that are trivial even in comparison to the
already minimal changes between the tiff an the first jpeg. (And file
fragmentation on the disk has nothing to do with the contents of the file.)
David J. Littleboy
Tokyo, Japan
Gregory Blank
"rafe b" <ra...@foobar.com> wrote:
> Aside from sharing small images on the web and
> via email and the like, JPG is a reasonable format
> for images *in their final state.*
>
> It is obviously not a good format for images still
> in-progress.
>
> My archiving regimen for both scans and digicam
> captures is to save TIFs and RAW files on multiple
> backup media (eg., DVDs.) The corresponding TIFs
> on my hard drive are then converted to best-quality
> JPGs and left on the hard drive.
>
> In all but the most critical cases, printing is done
> from these JPGs. If it's an image that I intend to
> work on over a span of time (eg., several evenings)
> then I'll fetch the TIF from the archives save as TIF
> until I'm done.
>
> It's not like you and Gregory have told us anything
> new. Yes, one can misuse JPGs. So what?
>
> My point was and remains this: a single instance
> of JPG encoding, using "high quality" settings,
> results in an image that is indistiguishable by eye
> from the corresponding TIF, in print or on-screen.
>
>
> rafe b
> www.terrapinphoto.com
Jeez that actually sounds like some sort of concession :-)
David J. Littleboy
"Q.G. de Bakker" <q...@tiscali.nl> wrote:
> Scott W wrote:
>
>> I don't use tiffs in my workflow, I don't know why people do, I use psd
>> files, I find them a bit more to my liking. I do use jpeg when showing
>> a result of an image, to use a psd or tiff would not add anything at
>> all.
>
> That's not true.
> JPEG is a compression method that does indeed causes lots of artifacts to
> appear, degrading image quality. By how much depends on the compression
> settings, but it's always there.
Your "lots of artifacts" is problematic: artifacts you can't see really
don't count as being "lots of artifacts".
> What matters, of course, is how obvious the degradation is. In a visual
> medium, what you do not see is not there. Right? ;-)
If you overlay a tiff and a best quality jpeg from that tiff, blow up to
400%, and switch back and forth, you won't be able to find any differences.
So the idea that jpeg is problematic for prints is patently ridiculous.
> So a carefully monitored JPEG-compression, tailored to the end use (mainly
> displaying on computer screens) may be quite acceptable.
Displaying on computer screens is far more critical than printing; printing
is 200 to 300 ppi, computer screens are 70 to 120 ppi with much higher
contrast.
> TIFFs and PSDs can contain much more than JPEGs.
Not really. Real images contain much less information than random values, so
tiff is ridiculously inefficient way to represent the information.
> So while JPEGs may be good
> as an final-format, they are utterly useless as long as the 'workflow' is
> still flowing.
The only problems with jpegs during the workflow are (1) some application
may default to a higher compression rate than you want, and (2) you really
ought to keep intermediate steps in a 16-bit representation to prevent
posterization.
> And that's not even considering that every time you save a reworked image,
> (more) quality is thrown away.
This is simply wrong. Jpeg only loses information when it tries to encode
something it can't. So recoding a previously jpeg encoded image doesn't lose
information, because there's nothing in the image jpeg can't encode.
rafe b
<greg@greg_____photo.com> wrote:
>Jeez that actually sounds like some sort of concession :-)
You may flatter yourself into believing that, but almost
everything that I've said in this thread is to counter some
mis-statements uttered by Gordon in his first post, directed
at Neil.
Later in the thread there were other silly ideas I addressed.
Like the idea that simply reading or "opening" a JPG
degrades it.
Nobody denies that JPG is lossy. We seem to disagree
on the extent of loss. Folks make silly claims about
extreme loss, or speculate about losses in hypothetical
scenarios of obvious misuse. WTF?
rafe b
www.terrapinphoto.com
rafe b
<greg@greg_____photo.com> wrote:
>Never the less- appropriate to the previous discussions and
>when have you been terribly concerned with how relevant
>a thread is to rec.Mf? :-)
To the extent that scanners relate to MF (and being the
OP you must believe they do) -- it's important to point
out the specific domain of scanners in Leigh's survey.
This domain of scanners just happens to exclude
one of the best choices for serious MF work.
This in no way discredits the work. Leigh's survey
explicitly covers LF scanners only.
I didn't have permission to chop Leigh's original
down so that I could scan it on the Nikon. But I was
sorely tempted!
rafe b
www.terrapinphoto.com
rafe b
Your assertion surprised me so I put it to the test.
You can view the results here:
test1.tif is the "original", downsampled to 675x450
from a digicam shot, with a touch of sharpening.
test1.jpg through test10.jpg are successive cycles
of: open, edit, save in Photoshop, using Quality 12.
The edit in each case was to apply a tiny new spot
of black in the upper left corner, with the paintbrush
set at 1 pixel.
I figured this was enough to constitute an "edit" --
ie., it would force a new encoding of the existing
bitmap with each save.
Conclusion: nothing lost, through ten generations,
in terms of detail. However, there's some poster-
ization that's pretty evident in the 10th gen JPG.
The losses are just barely visible in the 5th gen.
Interestingly, the file size increases about 2 or 3
percent in each generation.
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> Your "lots of artifacts" is problematic: artifacts you can't see really
> don't count as being "lots of artifacts".
That's right.
That's why i wrote what i wrote next.
> > What matters, of course, is how obvious the degradation is. In a visual
> > medium, what you do not see is not there. Right? ;-)
>
> If you overlay a tiff and a best quality jpeg from that tiff, blow up to
> 400%, and switch back and forth, you won't be able to find any
differences.
The "best quality" JPEG compression actually repacks JPEG image data in a
JPEG image file, without performing compression.
> > TIFFs and PSDs can contain much more than JPEGs.
>
> Not really. Real images contain much less information than random values,
so
> tiff is ridiculously inefficient way to represent the information.
That "much more" includes layers (image data, or 'meta data'), and other
usefull stuff.
> The only problems with jpegs during the workflow are (1) some application
> may default to a higher compression rate than you want, and (2) you really
> ought to keep intermediate steps in a 16-bit representation to prevent
> posterization.
The big problem is that everytime you save a JPEG-file as a JPEG-file, the
algorithm starts anew and messes up the image data once more.
So you really need to keep the times you open, rework and save the JPEG file
to a minimum.
This big and real problem is easily avoided by not using JPEG format while
the workflow is still in flow. Use it as a final-format.
> > And that's not even considering that every time you save a reworked
image,
> > (more) quality is thrown away.
>
> This is simply wrong. Jpeg only loses information when it tries to encode
> something it can't. So recoding a previously jpeg encoded image doesn't
lose
> information, because there's nothing in the image jpeg can't encode.
Very true. But you obviously haven't considered what that means.
If you open a JPEG, *do not change the image information*, and resave as
JPEG, there will be no problem. Like you said, the JPEG algorithm has no
problems decoding what it first encoded, nor recoding it the exact same way.
But as soon as you change a bit, it has to recode that, and it does recode
the changed image data *including the artifacts* caused during the first
compression.
So if you only open to change nothing (why save again?), there is no
problem, and you are right.
But as a working file format JPEG stinks. Quality will suffer.
Q.G. de Bakker
> test1.jpg through test10.jpg are successive cycles
> of: open, edit, save in Photoshop, using Quality 12.
Make it a usefull test, and let the JPEG algorithm actually do something...
rolan...@hotmail.com
Gregory Blank wrote:
> In article <1134045548....@g47g2000cwa.googlegroups.com>,> > > In article <1134030831....@g14g2000cwa.googlegroups.com>,
> > > rolan...@hotmail.com wrote:
> > >
> > > >
> > > > I've checked carefully and I do not see any difference in the detail.
> > >
> > > That was quick for outputting on an Imagesetter.
> >
> > Oh, I left it behind today. How remiss.
> >
> > I think you are talking BS in this case. I don't see any difference in
> > the images and I don't believe there will be a difference in the print,
> > unless the file format is handled differently.
>
> Well there certainly is a big issue I can say from personal printing
> that is having my work published. Jpegs do not hold the contrast
> within the range a tiff file can. Therefore jpegs are losing information
> from the scene.
Yes they are, but I think they tend to hit the blues hardest and for
skies this can sometimes improve the sky since you don't want to see
much detail in a sky.
> I also know that opening and closing jpegs multiple times creates issues
> whether you say so or not, greater minds than yours and mine have
> qualified this.
Thanks for ending your post with this. I like to be cheered up on
Fridays to put me in a jocular mood for the weekend.
rolan...@hotmail.com
You might like to read this article. I'm still having a chuckle.
http://www.michaelfurtman.com/jpeg_myths.htm
Q.G. de Bakker
> You might like to read this article. I'm still having a chuckle.
> http://www.michaelfurtman.com/jpeg_myths.htm
This Myth Buster starts with proclaiming a believe in another myth, that
TIFF files are uncompressed, and lossless.
TIFF files can contain compressed image data. One of the compression methods
it allows is JPEG compression...
Credibility?
;-)
rolan...@hotmail.com
But he was talking about the TIFF file in his camera and rightly said
it was not compressed and I would think that is true of every digital
camera.
> Credibility?
> ;-)
Who said "Send in the clowns"? This should be a good weekend.
rafe b
wrote:
>rafe b wrote:
Eh? What did you have in mind?
I should clarify a bit. Each cycle consisted of open, edit,
save, and close. The edit of test1 became test2... etc.
I'm not 100% certain, but it seems to me there would have
had to be a new JPG encoding of the entire image with
each generation.
In any case, there were observable effects after the first
few generations -- in image areas far removed from the
actual edit.
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> >Make it a usefull test, and let the JPEG algorithm actually do
something...
>
>
> Eh? What did you have in mind?
Use setting "6", for instance.
rafe b
>It seems a bit funny to be worried about the fact that the images are
>jpegs but not the fact that they have be down sampled to 2400 ppi. For
>a scanner test I would have thought that leaving the images at 4000 ppi
>might not have been a bad idea. I have always held that there is not
>much detail past 2000 ppi in film, but it seems that if you are going
>to give a scanner a work out go for the high numbers.
This is a good question, and I've thought about it myself a bit.
Anyone presenting images from scanners at "native" resolution
has to explain (and know!) precisely what that resolution is -- or
else resample each submitted image either up or down to some
fixed resolution designated as the "standard."
Viewers must also know and carefully consider these numbers,
and/or do their own resampling or resizing (they might elect
to do it in their head...)
I had one person (with a high end Howtek drum scanner) who
was seriously confused about just what the native res was of
the images he sent me. Consequently, his scan samples were
never posted.
>I know the Nikon SUPER COOLSCAN 9000 ED can get lots of detail in a
>4000 ppi scan and it is a CCD scanner. In the end if drum scanners can
>really produce high resolution scans then this test did test them close
>to their limits.
Like I said elsewhere... I truly wish I could have scanned
Leigh's chrome on my Nikon. A shame.
Here are the film scanners I've owned and used over
the years:
1998-1999: Microtek 35t+, 1950 dpi, 35mm only
1999-2001: Polaroid Sprintscan 35+, 2700 dpi, 35 mm only
2000-2001: Epson 1640SU (for MF) 1600 dpi, all formats
2001-present: Nikon LS-8000, 4000 dpi, 35mm, MF
2004: ScanMate 5000 (drum, 5000 dpi), all formats
2004-2005: Microtek 2500, 2500 dpi, all formats
(A bit over $6K total. Ouch.)
Each new scanner yielded improved results,
though the differences get subtle near the high
end. The ScanMate was in fact visibly better
than the Nikon -- but the difference was small.
(See Dave King's A:B comparison on my
site -- same ScanMate, but Dave's Nikon.)
Comparing the ScanMate against the Microtek
on an LF scan I noticed something interesting
and in hindsight, predictable. In some regions of
the scan, scanner A was sharper, and in other
regions, scanner B was sharper. The sharpest
regions of A and B were very similar looking.
In both cases the film was air-mounted.
This tells me that the results were limited
by film flatness.
In terms of tonality, drum scans are clearly
ahead of the game, followed by the Nikon,
and then far behind are the CCD scanners
that use cold-cathode lighting. The Nikon
approach (monochrome CCD with RGB
LED illumination) gives a lot of bang for the
buck. On C41 media, it's nearly indistinguish-
able from a drum scan, IMO, and digital ICE
is nothing short of magic.
In my experience, each of the CCD scanners
with cold cathode illumination had problems
with banding in dense media.
rafe b.
www.terrapinphoto.com/jmdavis
www.terrapinphoto.com/drumscansaga
rafe b
wrote:
>rafe b wrote:
Why bother?
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> Why bother?
Because showing that the things involved with JPEG compression have a
miniscule effect by letting the JPEG thing do next to nothing is utterly
pointless.
So unless you let it actually do something, you might as well have not
bothered at all.
Gregory Blank
rafe b <rafebATspeakeasy.net> wrote:
> 1998-1999: Microtek 35t+, 1950 dpi, 35mm only
>
> In my experience, each of the CCD scanners
> with cold cathode illumination had problems
> with banding in dense media.
That was my first scanner, I still own it. I never use it
but its still down in the basement in a box. In any event
when I first got it, it seemed quite good, however the scan wizard
software was tricky to use by comparison to the Epson 2450 interface.
I only notice problems with single band lines appearing in the scan close
to the end of when was using it. I bought
a Microtek 120tf last fall but had unresolved issues with the scanner &
company, the first scanner I received did produce very nice results on
par with the Nikon scans I have done at a rental Darkroom facility. And
its shame that the companies/including the reseller took certain
customer unfriendly stances-becuse i liked the price and the scanner.
Gregory Blank
"Q.G. de Bakker" <q...@tiscali.nl> wrote:
Is that not a function of the Image editor?, I would try emailing it
back and forth 100 times before I judged based on purposefully degrading
the image.
rafe b
"Q.G. de Bakker" <q...@tiscali.nl> wrote in message
news:439999f4$0$9254$dbd4...@news.wanadoo.nl...
There was most certainly a JPG encode step with each encode.
The first encode (from the TIF) using best-quality settings,
yielded nearly 10:1 compression.
My point was, and remains, that a single low-compression
JPG encode step causes very little image degradation. As
a rule, I only use low-compression JPG settings in critical
work.
rafe b
www.terrapinphoto.com
rafe b
"Gregory Blank" <greg@greg_____photo.com> wrote in message
news:greg-F63622.1...@news.isp.giganews.com...
> In article <439999f4$0$9254$dbd4...@news.wanadoo.nl>,
> "Q.G. de Bakker" <q...@tiscali.nl> wrote:
>
>> rafe b wrote:
>>
>> > Why bother?
>>
>> Because showing that the things involved with JPEG compression have a
>> miniscule effect by letting the JPEG thing do next to nothing is utterly
>> pointless.
>> So unless you let it actually do something, you might as well have not
>> bothered at all.
>
> Is that not a function of the Image editor?, I would try emailing it
> back and forth 100 times before I judged based on purposefully degrading
> the image.
And what would be the point of that?
We've already covered this. Images do not
degrade in the reading, opening or (accurate)
transmission. On the internet, TCP/IP protocol
ensures accurate transmission.
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> > Because showing that the things involved with JPEG compression have a
> > miniscule effect by letting the JPEG thing do next to nothing is utterly
> > pointless.
> > So unless you let it actually do something, you might as well have not
> > bothered at all.
>
> Is that not a function of the Image editor?, I would try emailing it
> back and forth 100 times before I judged based on purposefully degrading
> the image.
That's what you do everytime you save image date in a JPEG file format.
By how much, and whether it matters is another matter.
But what's happening here is much like showing that, say, a car's engine can
stand high revs by letting it run low revs for many hours. Silly.
Gordon Moat
> Q.G. de Bakker wrote:
>
>>rolan...@hotmail.com wrote:
>>
>>
>>
>>>You might like to read this article. I'm still having a chuckle.
>>>http://www.michaelfurtman.com/jpeg_myths.htm
>>
>>This Myth Buster starts with proclaiming a believe in another myth, that
>>TIFF files are uncompressed, and lossless.
>>TIFF files can contain compressed image data. One of the compression methods
>>it allows is JPEG compression...
>
>
> But he was talking about the TIFF file in his camera and rightly said
> it was not compressed and I would think that is true of every digital
> camera.
>
The problem with TIFF in most D-SLR is that it is limited to sRGB colour
space. I think Kodak might be the exception in that. In some cases, TIFF
is not the full bits possible when shooting RAW, so in some ways, on
certain cameras, JPEG and TIFF might be much closer than one might think.
Some medium format digital backs do handle direct TIFF files differently
than D-SLRs. There are also sometimes more colour space choices,
including directly capturing to CMYK files, something than can
streamline workflow. Anyway, I only mention medium format digital backs
because that is the newsgroup in which we are having this discussion.
I looked through the article, and one thing really stands out, which is:
"To test this, I took an image and saved it 30 times during a single
session – a number far greater than anyone would normally ever do. I
then produced 100% crops of the two images and placed them side by side.
While there was some slight change in contrast and hue (both fixable),
there was little, if no, degradation of detail."
So the important (to me) words that crop up in that are "change in
contrast and hue". Colour is very important, though I understand many
people might put more importance on resolution. It baffles me why anyone
would accept changes in contrast and hue, both from a practicality and a
workflow viewpoint.
I can save a 16-bit TIFF, but I cannot do that with a JPEG. Note: some
D-SLRs are only 10-bit or 12-bit, though PhotoShop works in either 8-bit
or 16-bit modes. I can also save a CMYK TIFF. These workflow advantages
seem to be good reasons to use TIFF. JPEG takes up far less disk
storage, but with disk storage at such reasonable prices, is this false
economy?
Ciao!
Gordon Moat
A G Studio
<http://www.allgstudio.com>
rafe b
"Gordon Moat" <mo...@attglobal.net> wrote
> The problem with TIFF in most D-SLR is that it is limited to sRGB colour
> space. I think Kodak might be the exception in that. In some cases, TIFF
> is not the full bits possible when shooting RAW, so in some ways, on
> certain cameras, JPEG and TIFF might be much closer than one might think.
There's very little point in saving images as TIF on a D-SLR
or any other digicam. This would require the camera to do the
de-mosaicing and the resultant TIF would be enormously
inefficient relative to either the RAW or JPG equivalent.
> Some medium format digital backs do handle direct TIFF files differently
> than D-SLRs. There are also sometimes more colour space choices, including
> directly capturing to CMYK files, something than can streamline workflow.
> Anyway, I only mention medium format digital backs because that is the
> newsgroup in which we are having this discussion.
CMYK workflow? Get with the program.
That's so early-nineties. CD Tobie would laugh.
> I looked through the article, and one thing really stands out, which is:
>
> "To test this, I took an image and saved it 30 times during a single
> session – a number far greater than anyone would normally ever do. I then
> produced 100% crops of the two images and placed them side by side. While
> there was some slight change in contrast and hue (both fixable), there was
> little, if no, degradation of detail."
>
> So the important (to me) words that crop up in that are "change in
> contrast and hue". Colour is very important, though I understand many
> people might put more importance on resolution. It baffles me why anyone
> would accept changes in contrast and hue, both from a practicality and a
> workflow viewpoint.
It baffles my why anyone would do multiple saves and opens
on a critical file in JPG format. That's simply bad practice.
JPG usually (though not necessarily) works by decimating
color into, while retaining detail. It does this by storing the
image in Yuv color space (lightness plus two complementary
color axes.) It's an intelligent compromise because it exploits
a major weakness in human vision. We see detail or color,
but not much detail in the color.
> I can save a 16-bit TIFF, but I cannot do that with a JPEG. Note: some
> D-SLRs are only 10-bit or 12-bit, though PhotoShop works in either 8-bit
> or 16-bit modes. I can also save a CMYK TIFF. These workflow advantages
> seem to be good reasons to use TIFF. JPEG takes up far less disk storage,
> but with disk storage at such reasonable prices, is this false economy?
10:1 compression with no visible image degradation isn't
"false economy" in my book.
It doesn't take long to fill a 160 Gbyte hard drive with
film-scans and digicam captures in 16-bit TIF or RAW
format. CMYK workflow only makes sense for a few
professionals in the biz -- and hardly ever for those printing
from desktop inkjet printers.
rafe b
www.terrapinphoto.com
Gordon Moat
> Gordon Moat wrote:
> snip A whole bunch of stuff dealing wtih color.
>
> Seems to me like a bit of a switch of topic here. What does color
> space have to do with the differences between jpeg and tiff, either one
> can be in sRGB and either one can be in Adobe 1998 RGB.
Not a switch, this goes back to my comment that started this round of
replies, in which I stated looking at a JPEG on a computer monitor is
not a good indication of a final print. That relates directly to colour,
though I understand more people would compare resolution. Like I stated,
it (colour) is somewhat subjective.
Okay, so the majority of monitors only display in sRGB. While using some
other RGB working space would change the range of available colours, on
the majority of monitors it would be difficult to see that change. In a
similar manner, since monitors only work in RGB, viewing a CMYK image is
really only a simulation.
Both sRGB and AdobeRGB are limited RGB spaces. While the view I have
about ProPhotoRGB is still a little controversial, I notice that
currently more writers of digital imaging are embracing that space. I
have been using that since 1999, with a good track record of
commercially printed results. Now I find a few writers, or digital
gurus, talking about the benefits of that.
So can colour space choices affect resolution, contrast, or apparent
sharpness? How does edge contrast change when colour space changes? Will
anyone be able to see that on an sRGB monitor? The last one might be the
limiting parameter, which is namely that our least common denominator
might be masking the differences.
>
> But you were not talking about the color space, the scanner test in
> fact does not cover color and said so up front.
I think it is a nice test even without testing colour. However,
resolution is only one part of scanning. Clients tend to notice errors
in colour after something is printed much more than they notice any lack
of sharpness or resolution. Of course, your mileage may vary, and you
might find people who do care more about the resolution. I don't think I
would use a scanner that had good resolution, but not so good colour
capture ability, unless I was only scanning B/W films.
What the test is
> covering is detail captured, you position is that jpeg degrades the
> quaility to such a degree that you can't really tell about detail when
> viewing a jpeg image, this is clearly not the case.
Maybe you cannot tell on a computer monitor. Maybe you cannot tell on a
small print, an inkjet print, or some digital/chemical prints. However,
why not use the best choices in quality, and avoid any possible
degradation due to a lossy format?
>
> I don't use tiffs in my workflow, I don't know why people do, I use psd
> files, I find them a bit more to my liking.
Everything I have that has more than one layer remains as a PSD. Things
that went to print are often stored as CMYK TIFF files, though there is
always an RGB PSD file somewhere. Obviously, this can take up lots of
storage space.
I do use jpeg when showing
> a result of an image, to use a psd or tiff would not add anything at
> all.
>
> Scott
>
Using a PDF is another viewing choice. As I mentioned in one of my
posts, it can be a nice way to soft proof, though not better than having
a hard copy proof.
Scott W
> Not a switch, this goes back to my comment that started this round
of
> replies, in which I stated looking at a JPEG on a computer monitor is
> not a good indication of a final print.
So let us have a look once more at what you did say
>"What I found interesting is that I could actually see differences in the
>JPEGs. Obviously, we both know that a JPEG is a poor way to judge how a
>printed item will turn out, though the fact that differences are visible
>in such a low quality representation of scans is interesting. I have
>been mostly opposed to these internet challenges mostly due to the
>degradation of images by posting and viewing JPEGs. "
But you have been unable to show that there is any meaning full
degradation of the images. You clearly made it sound like jpeg was
going to ruin the image so much as to be of little if any use. You
claim that two images that look exactly the same on the screen, even
when viewed at 400%, are somehow going to magically appear differently
in the final print.
You are trying to tell us that jpeg compression ruins a image to the
point that you can no longer use it to judge image quality, and yet
look as close as you might you will not see a difference.
You claim a sharpening effect from saving as jpeg, but you can't show
us this effect and I do believe that I have shown there is no
sharpening effect.
Scott
Gordon Moat
>. . . . . . . . . CMYK workflow only makes sense for a few
> professionals in the biz -- and hardly ever for those printing
> from desktop inkjet printers.
>
>
Gee, how 'bout that . . . I work in the pre-press and printing industry,
and I don't use desktop inkjet.
Gordon Moat
Look Scott, if you are happy with JPEGs, I doubt anything I would ever
type, nor anything anyone else might state, would ever change your mind.
If you think JPEGs are better than holes in Swiss cheese, then keep
using them. If it mattered any, or bothered you any, then do some
research and reading on your own. There are many book writers, digital
gurus, photographic workshops and other choices where you can hear or
read many opinions about JPEGs. That there is no one universal agreement
probably has more to do with habits and comfort levels people have
developed over the years.
My writing style can sometimes seem extreme or harsh, though I think
that makes people think about why I would write something. I encourage
anyone to investigate on their own, and never trust any one single
source. Read as many sources as you can, then make your own decisions.
I would rather see everyone figure out their own opinions on these
matters. In fact, it would bother me if people just blindly accepted
what I wrote, or never questioned what I wrote. If you dismiss what I
wrote, then you don't need to go any further; if you question what I
wrote, then do your own investigating. There is no trophy girl for this
newsgroup, and no prize for a "winner" of any discussion.
Scott W
> Look Scott, if you are happy with JPEGs, I doubt anything I would
ever
> type, nor anything anyone else might state, would ever change your mind.
> If you think JPEGs are better than holes in Swiss cheese, then keep
> using them.
You still don't get it, this is not about what I use in my workflow.
I am challenging your statement that jpeg images are unsuitable for
judging the quality of scans. When backed into a corner you start
talking about workflow and trying to imply that I prefer to use jpeg
files in my workflow, which I in fact stated I do not.
Part of what I do for a living is writing image-processing software.
Not too surprising I have to deal with jpeg as the input image and as
such I do know a thing to two about just what is in a jpeg file.
Can you show me a tiff from a film scan that can not be converted into
a jpeg where the jpeg and the tiff look exactly the same on the screen?
Now let me be very clear on what I am not saying, I am not saying that
the jpeg output from a scanner will always look as good as the tiff,
the scanner software might do a poor job of producing the jpeg. I
always scan as a tiff and then use Photoshop to convert to a jpeg when
I need a jpeg.
Scott
Scott W
> I can save a 16-bit TIFF, but I cannot do that with a JPEG. Note:
some
> D-SLRs are only 10-bit or 12-bit, though PhotoShop works in either 8-bit
> or 16-bit modes. I can also save a CMYK TIFF. These workflow advantages
> seem to be good reasons to use TIFF. JPEG takes up far less disk
> storage, but with disk storage at such reasonable prices, is this false
> economy?
You keep going back to your work flow, a bit of miss direction if ever
I saw one.
Did anyone say you should not store you photos as 16 bit/color tiff if
you wish to?
But then this is not an agument about workflow now is it?
The thing that is nice about the internet and digital images is that it
should be hard to BS. If you say your scans are sharper then another
guys then you can both show your scans. But then there will always be
someone who will state that you can judge the quality form the jpeg
images that are posted. In one example not too long ago a drum scan
and a scan from a CCD scanner looked very much the same, someone tried
to claim that you could not tell if they were or not because the images
were jpegs, pure BS.
The reason that jpegs are used for these comparisons is that they do
not degrade the image quality enough to matter and they greatly shorten
download times.
Every review site that reviews any kind of image equipment, whether it
be scanners or digital cameras uses jpeg to show the images. The use
of jpegs for evaluating images is well established, I would say the
burden of proof is on you to give evidence that this practice is as
flawed as you seem to believe it is.
Scott
rafe b
wrote:
>So can colour space choices affect resolution, contrast, or apparent
>sharpness? How does edge contrast change when colour space changes? Will
>anyone be able to see that on an sRGB monitor? The last one might be the
>limiting parameter, which is namely that our least common denominator
>might be masking the differences.
Fact is, there are precious few digital print
methods that remotely approach the overall
volume of the much-maligned sRGB space.
Yes, the printer will exceed sRGB limits in
certain corners of Lab space, but the overall
volume of sRGB still dwarfs that of most
printers -- including inkjet, CMYK offset and
LightJet.
You can see this for yourself at
http://www.iccview.de.
Put sRGB in one column and any ICC print
profile (or one of your own) in the other.
Let 'er rip. (Requires VRML plugin for the
most entertaining show.)
The notion that a print is the best way to judge
either sharpness or tonality... is hard to justify
based on gamut or color space.
rafe b
www.terrapinphoto.com
Gordon Moat
Whatever Rafe. If you like sRGB, certainly don't let me talk you out of
using it. You will be in the fine company of camera engineers who design
most of the current D-SLRs, since they have a preference towards sRGB.
Meanwhile, I have no problem using ProPhotoRGB, so let's just leave it
at that . . . agree to disagree.
Gordon Moat
> Gordon Moat wrote:
> > I can save a 16-bit TIFF, but I cannot do that with a JPEG. Note:
> some
>
>>D-SLRs are only 10-bit or 12-bit, though PhotoShop works in either 8-bit
>>or 16-bit modes. I can also save a CMYK TIFF. These workflow advantages
>>seem to be good reasons to use TIFF. JPEG takes up far less disk
>>storage, but with disk storage at such reasonable prices, is this false
>>economy?
>
>
>
> You keep going back to your work flow, a bit of miss direction if ever
> I saw one.
> Did anyone say you should not store you photos as 16 bit/color tiff if
> you wish to?
>
You selectively crop my message, then pull a note off the end as the
only thing to make a statement onto. Did you either not read the entire
post, nor understand what I posted? Did you notice that I posted in
answer to someone other than you? If you feel you are being misled, or
misdirected, that was not my intention.
In a roundabout way it was implied that images should be stored as
8-bit. Storing as a JPEG is only 8-bit, not 16-bit. I don't recall you
stating you stored images as JPEG, but honestly I don't feel like
digging through the messages to find out if you did or did not state
anything like that.
So see if you can follow that, I stated that I can save a 16-bit TIFF
file. Sure seems like a benefit of using TIFF over JPEG, unless of
course you think there is no benefit to 16-bit files. Of course this
thread has wandered through many areas, including the relative merits
(or lack) of JPEGs. Hard to discuss JPEG without mentioning other file
formats.
> But then this is not an agument about workflow now is it?
Sure it is . . . how do you get to a JPEG without involving some sort of
workflow. How can I give any contrast or comparisons of JPEGs and other
choices in files without mentioning workflow.
>
> The thing that is nice about the internet and digital images is that it
> should be hard to BS. If you say your scans are sharper then another
> guys then you can both show your scans. But then there will always be
> someone who will state that you can judge the quality form the jpeg
> images that are posted.
I think you meant cannot, rather than "can".
In one example not too long ago a drum scan
> and a scan from a CCD scanner looked very much the same, someone tried
> to claim that you could not tell if they were or not because the images
> were jpegs, pure BS.
Someone?
Masterfully done CCD scan, poorly done drum scan, other way around,
heavy compression, resampled images. Surely a "miss" direction if I ever
heard one. ;-)
>
> The reason that jpegs are used for these comparisons is that they do
> not degrade the image quality enough to matter and they greatly shorten
> download times.
>
Well, we all like shorter download times. :-)
> Every review site that reviews any kind of image equipment, whether it
> be scanners or digital cameras uses jpeg to show the images. The use
> of jpegs for evaluating images is well established, I would say the
> burden of proof is on you to give evidence that this practice is as
> flawed as you seem to believe it is.
>
Why would the burden of proof be on me? Sounds to me like you are
already convinced of all that you do in digital imaging. There would be
little point in me trying to "prove" anything to you. Besides, if how I
do things is so terribly wrong, then you should be dancing for joy that
you have things so right, and that your printed images should be vastly
better than mine. I think you have already convinced yourself that you
are completely right.
rafe b
wrote:
>Whatever Rafe. If you like sRGB, certainly don't let me talk you out of
>using it. You will be in the fine company of camera engineers who design
>most of the current D-SLRs, since they have a preference towards sRGB.
>Meanwhile, I have no problem using ProPhotoRGB, so let's just leave it
>at that . . . agree to disagree.
It's not a matter of "liking" a color space. It's not my pal.
A color space is a mathematical space, a concept.
There are an infinite number of such spaces -- and some
are more useful than others, for any given purpose.
If you'd like to counter my arguments with facts, we
might have something to talk about.
FWIW, I use AdobeRGB and don't fret my working
space all that much.
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> It's not a matter of "liking" a color space. It's not my pal.
> A color space is a mathematical space, a concept.
> There are an infinite number of such spaces -- and some
> are more useful than others, for any given purpose.
>
> If you'd like to counter my arguments with facts, we
> might have something to talk about.
"Facts"?
"There are an infinite number of such spaces"?
;-)
rafe b
wrote:
>rafe b wrote:
Indeed. deBakkerRGB. Imagine that.
A dozen or so of the most useful ones are
discussed here, and these don't even include
the ICC variants that we all know and love.
<http://www.poynton.com/PDFs/ColorFAQ.pdf>
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> >"Facts"?
> >"There are an infinite number of such spaces"?
> >;-)
>
>
> Indeed. deBakkerRGB. Imagine that.
You do not see the irony in your plea for sticking to facts, while engaging
in dramatic exaggeration, leaving the realm of fact, yourself?
I do. ;-)
rafe b
wrote:
>rafe b wrote:
It's no exaggeration.. Anyone can go into Photoshop's
Color Settings->Custom dialog and create a custom
ICC color space. That's a fact.
The first such that I ever heard of was BruceRGB.
Later there was EktaRGB, then WideGamutRGB, etc.
For that matter, the profile for any real input or output
device defines a unique color space -- generally
not suitable as a working space, but perhaps you
get my point.
What do you suppose AppleRGB is? Or
ColorMatchRGB? One was designed around
legacy monitor settings; the other around the
Radius PressView and similar pro-grade gear
of the time (ca. 1998.)
rafe b
www.terrapinphoto.com
Q.G. de Bakker
> It's no exaggeration.. Anyone can go into Photoshop's
> Color Settings->Custom dialog and create a custom
> ICC color space. That's a fact.
That is indeed a fact.
The exaggeration lies in the "infinite".
Scott W
Well since a color space is a mathematical construct one could say that
there were in infinite number of them.
Ah for the good old days were we just argued between using x, y, Y or
u, v, Y or u', v', Y. This argument would get rather heated as I
recall. All of these color spaces had a nice feature, they could
represent any color the human eye could see. There were not the most
useful for working on images however since they needed a lot of bits to
get the colors accurate.
A bit of trivia, if a color falls within the sRGB space it will be more
accurately encoded in that space then in Adobe RGB. Of course if it
fall outside of the sRGB space then there are problems. Adobe RGB
encompasses a larger volume, and so if you have the same number of bits
to define the location of a color in that volume the distance between
colors in Adobe RGB will be larger then sRGB. As always there are trade
offs and you need to know which one works best for a give situation.
I use to design CRTs for military aircraft, we would have a spec for
the color space of the tube. Since there are variations from tube to
tube no two would have exactly the same color space, the specs allowed
for a range of spaces. One might say we had in fact an infinite number
that we could hit and still have the tube meet spec.
Scott
Q.G. de Bakker
> Well since a color space is a mathematical construct one could say that
> there were in infinite number of them.
Not so.
First, "mathematical construct" and "infinite" do not come together, as if
they are somehow synonymous. So much for your "since". ;-)
Second, the entire concept of colour space itself is an expression of the
limited and limiting (i.e. finite) nature of colour, both colour perception
and colour reproduction.
So "infinite number" and "colour spaces" just do not go together.