Minolta 5400 or Coolscan 5000

[Kennedy McEwen]

Then we can put it to rest because that was never the issue.

The issue was (and I'll repeat it once again):


You lose a lot of quality this way.

To say "lose quality" is wrong, to say "lose a lot of quality" is "a
lot" wrong and shows a lack of very rudimentary understanding.

And to say that "the truth is exactly the opposite" is wrong and shows
more than a little arrogance, especially as you did not even bother to
question Toby as to what he meant by his statement or how he reached it.
Since he has now done so and explained the condition that he was
referring to (and it correct in that situation) it is even more clear
that your absolute statement was wildly inaccurate.

 

[Kennedy McEwen]

Processing not, but the ability to employ yes. You just can't achieve
the same result looking at the image in the tiny Preview windows even
if you used Photoshop or any other image editing software.

I explained before that you can, if you are so concerned about the
minimal probability cases, preview at full scale. I currently use a
1600x1200 pixel display, so that is only 8 or so preview panes to view
every pixel at full resolution. But then, statistically, that is
completely unnecessary - a 1200x900 pixel preview is a statistically
valid proportion of the image on which to assess *all* of the controls
available except for sharpening.

I never claimed that PS or any other image
editing software has superior algorithms to scanner software. They
may, but that's not what this is about.

By implication, the statement that scanning raw and processing in PS,
which is the "exact opposite" of optimum scanning for input to PS, is
endorsement of PS as superior to the same processes performed by the
scanner software. You have no evidence that this is so.

I said, among other things, that:

- stating that *relative* user friendliness of "auto-everything"
directly translates into *absolute* technical superiority over
knowledgeable "everything-off" user is wrong.

- stating that relying on tiny Preview window and a limited subset of
tools is superior to full display and a full set of tools is wrong.

- stating that auto is "a lot" better than manual is wrong, except in
the singular and very unlikely and rare case of a "lucky shot" - when
it's at best merely equal - which "lucky shot" is beyond the scope of
this discussion.

You have said a lot of things in this thread Don, but those particular
words do not appear in any of your articles prior to today.

 

[Kennedy McEwen]

Very good!!! Tests are essential, although one has to prepare tests
very carefully otherwise one will arrive at wrong conclusions...

Case in point: The reason it's so windy in Holland is *not* because
all those windmill blades are spinning... ;o)

Nor is it because of all that natural gas being consumed by car engines!

 

[Kennedy McEwen]

This is not really a question of 8 vs. 16 bits. If you make changes in
your scanner software (regardless of bit depth) you are not getting a
"pure" scan from the scanner.

Analogue gain!!

 

[Kennedy McEwen]

BTW (just to keep on topic) I recently returned from a 2000 Km bicycle
ride, but I flew home so I lost all my enviro points.

You could have done worse:

Train : see http://www.futurepundit.com/archives/002197.html

Hybrid car(!) : see http://www.trainweb.org/midtown/us_hybridcar.htm

So flying home was probably the next best thing to cycling home!

But as an FotE spokesman said when presented with this data "the only
environmentally friendly journey is the one that is not taken at all" -
so, according to the greenies, you lost all your enviro points by taking
part in the cycle ride in the first place. Just think of all that CO2
your lungs unnecessarily pumped into the atmosphere, the produce you
consumed en-route, you environmental vandal you! ;-)

 

[Toby]

Hi Don,

I understand that the only difference between 8 and 16 bit scans has to do
with the amount of information in the final scan and not with the appearance
of the scan. Basically an 8 bit scan and a 16 bit scan with identical
exposure opened in PS are going to appear the same, because we do not
perceive the subtle differences engendered by the extra 65300-odd luminance
levels in each channel. However once with start screwing around with the
white and black point, gamma, curves, chroma, etc, we are in effect taking
bugger chunks out of the tonality in 8-bit files because a given pixel has
only 256 levels of choice in 8 bit but 65536 in 16. Yes, you still get a
gappy histogram and tonal discontinuities but the gaps are finer in 16 bit.

If, and *only* if you will be working with files in PS at a lower bit depth
than your native scanner bit depth, I still contend that you are better off
doing adjustments at the higher bit depth in the scanner before the file
gets interpolated down to 8 bit from 14, for example. If you set B & W
points intelligently so as not to clip information, and even curves to
change the tonality of some pixels, you are better doing so in 14 bits where
there are more levels to choose from and interpolating down, than taking a
raw scan that you are obviously going to have to make the same adjustments
to in PS and doing so in 256 levels. The gaps will be finer.

That's how I see it, anyway.

This absolutely does not apply to working in PS in a bit depth equal to or
higher than the scanner bit depth.

Right? Wrong?

Anyway this has been a most useful conversation for me, and has clarified my
rather confused thinking on the issues.

Toby

 

[Bruce Graham]

...

Oooh... Tell us more! ;o)

from Melbourne where I live due North to Longreach, Western Queensland.

The ride took a month with rest days and was about 70% bushland and 30%
pasture on a sealed road with almost no traffic apart from a road train
about every 20 minutes on average and some grey nomads with their
caravans.

In the Cobb & Co stage coach days (last coach 1928), the towns were about
100 Km apart (usually with an exchange of horses halfway in those days).
In Queensland especially, most of the old towns still exist.

Without exchanging the horses, 100 Km is a comfortable days ride on a
loaded touring bike today and many of the old pubs are still there,
sometimes just truck stops where you preorder your dinner by UHF radio 10
Km down the track, but usually the pub is the watering hole for the local
cattle and sheep people who are very friendly to an odd bike rider or
two.

Our theme was "The Last Great Aussie Pub Crawl" because we stopped at
just about every pub. Some were pretty basic, like the one that offered
free rooms to good drinkers (we booked in early for $10) but others were
almost Ritzy, like the Corones in Charleville.

I took an EOS30 with just a 28-105mm and this has kept my scanner busy.

On topic for a moment, I can't get excited by most of the arguments in
this thread. I take a middle road as follows.

I rarely use the Canon Filmget software with my FS4000 because it tends
to clip the histogram ends a bit much for my liking. I do use the IR
channel for cleaning routinely, so that and expense eliminated the
Silverfast option, leaving Vuescan as my standard. I scan 16 bit TIFF,
letting Vuescan do the mask removal for negatives and using "lock image
colour" so that I get control of the histogram end points for each
channel (another Vuescan secret handshake). I leave these fairly wide so
that I can scan a batch without previewing more than a few at the start.
The 130 Mbyte files go straight to Photoshop where I use 16 bit adjust
levels as an initial step, working on each channel. This produces
smooth histograms because of the 14 bit data. On completion of editing I
save as 8 bit jpeg (15-25 Mbyte depending on detail and remaining noise).

Now, is there a problem with that? (lets not discuss the jpeg save at the
end).

Bruce

 

[Toby]

Hi Kennedy,

Yes, in a nutshell, analog(ue) gain.

BTW if one cannot vary the RGB LED intensities on the Nikon scanners does
that mean that the RGB controls in NikonScan are software controls only? I
suppose that they would have to be...

So I take it that the only actual hardware controls are analogue gain,
focus, multiscan and ICE.

Toby

Analogue gain!!

--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when

replying)

 

[Toby]

Hi David,

I didn't realize that about interpolation. However I still contend (see my
letter above) that--if and only if you are going to be using less bit depth
in PS than is native to your scanner--it is better to have it interpolated
in 14 bits in the scanner (or whatever) and then convert the file to 8 bits,
than it is to have the raw scan interpolated down to 8 bits without
correction and then make the corrections in 8 bits in PS.

Aside from my hazy thinking, which has now been cleared up thanks to you,
Don and Kennedy, I was really talking about folks who work on 8 bit files in
Photoshop. My argument doesn't apply to working on 16 bit files in PS. There
is also, as Kennedy points out, the question of where to set analog(ue)
gain, assuming that the dynamic range of the film you are scanning exceeds
the scanner's dynamic range.

Toby

 

[Kennedy McEwen]

Hi Kennedy,

Yes, in a nutshell, analog(ue) gain.

BTW if one cannot vary the RGB LED intensities on the Nikon scanners does
that mean that the RGB controls in NikonScan are software controls only? I
suppose that they would have to be...

So I take it that the only actual hardware controls are analogue gain,
focus, multiscan and ICE.

That is correct - and ICE is partly software as well, only the defect
detection is implemented in hardware.

 

[DavidTT]

Toby, one problem with all the tutorials and books is that they do not
tell the whole truth about how things actually work, either because the
authors are ignorant or are blinded by religion. "Get the best scan
possible in the scanner" or "Re-scan if the histogram looks bad in PS"
fuddled me for a long time. Rarely did I find sources that talk about
the division of labor between a scanner's hw and sw. (See my references
in this thread to the book and links that address this.) With Don's
support, I'm glad to be able to get this off my chest.

BTW, it you compare two recent revisions of Haynes' PS Artistry (both
great books), you will notice a subtle difference on this topic. In the
PS 7 revision, there is an advice on "Re-scan if the histogram looks bad
in PS" accompanied by ugly looking histograms. In the PS CS revision,
this section is removed.

It also took me a long time (a slow learner) to realize what is
happening with histograms. If PS can offer another histogram that can
show the effect of accumulative edits, it may demonstrate how blunt some
of the tools are, and the difference between 8 and 16 bit edits.

If you are not working in PS CS which has full support of 16 bit edits,
there are tutorials on how to work around this.

 

[Toby]

Thanks David,

I'm going to check the links you refer to.

Toby

 

[Don]

Which is also not true and a contradiction of your earlier statement
that these were hybrid vehicles with minimal emissions. Natural gas and
propane emit just as much CO2 per kW of produced power as octane
(gasoline).

When I lived and worked in the Netherlands (about 3 years ago) a guy
came in one morning and said his car broke down, the natural gas part,
so he had to switch to gasoline. He commented that he forgot how much
more "kick" gasoline gave to his car.

Natural gas? I asked, not being a driver and not knowing these things.
Sure, he replied, most cars in the Netherlands are hybrids.

Anecdotal? Of course, but he was not an environmentalist and had no
agenda, so I believed him.

Could he have been misinformed? Yes, but I can unequivocally state
that this very company (I contracted for) alone had more than 500
hybrid cars (which were part of the pay package for all full time
employees) let alone the rest of the country. So, your figures are
either incorrect or out of date, or both.

Anyway, let's get my *full* original statement in here for
recalibration:

On a serious note though, most cars in Europe, for example, are hybrid
vehicles with minimal emissions, certainly far less than 4WD monsters
common in North America. So an environmentalist driving such a low
emission vehicle would not be a contradiction.

In particular, the main point of my statement:
"certainly far less than 4WD monsters common in North America".

So even if the auxiliary "most" was incorrect (and I still maintain it
wasn't) it doesn't change the gist of my statement. Since you like to
parse, focus on "certainly far less...".

An average car in Europe is considerably smaller than an average car
in North America, with mileage to match.

In addition to the Netherlands I have also lived in Belgium and
Germany, to name only a couple of extra countries where I lived for
more than 1 year. I don't ever recall seeing a single SUV. Go to a
suburb of any large North American city and pretty much all you will
see are SUVs, not to mention 2-3 cars per family.

Due to higher population density in Europe the duration of an average
trip is far shorter than in North America.

Etc.

Which leads to only one conclusion: The cars in Europe pollute (at
least) an order of magnitude less than cars in North America. Or as I
originally put it: "certainly far less than 4WD monsters common in
North America".

Don.

 

[Don]

You seem to have forgotten to include that in any of your recent posts.

I don't think so... Starting with my very *first* message in the
thread:

Both VueScan and SilverFast are "point-and-shoot" programs similar to
disposable cameras. OK for casual tourist but not for people who care
for quality or have an inkling of what they are doing. Nothing wrong
with that if that's what they are after, but...

And ending with the last (before today's posting):

As I said many times, there's absolutely nothing wrong with that. If
you're happy with what you're getting - even if SF or VueScan or
whatever "massaged" scanner data to death - all that doesn't matter.
The important thing is that you are getting results which satisfy your
requirements.

Don.

 

[Don]

You have said a lot of things in this thread Don, but those particular
words do not appear in any of your articles prior to today.

It's a summary. For details refer to the messages themselves. They are
all there. I really don't have the time to quote them all individually
like I just did in the previous message.

Don.

 

[Don]

Analogue gain!!

We are not talking about Analog Gain. Analog Gain is on both sides of
the equation. We are talking about scanner software "curves, contrast
and friends..." vs. raw. Again:

I prefer to scan raw and edit in Photoshop.

You lose a lot of quality this way.

Don.

 

[Don]

Hi Toby!

If, and *only* if you will be working with files in PS at a lower bit depth
than your native scanner bit depth, I still contend that you are better off
doing adjustments at the higher bit depth in the scanner before the file
gets interpolated down to 8 bit from 14, for example.

Now, that is something completely different. You are not comparing
like with like.

Even working with the limited set of scanner software tools and the
problems of a small Preview window, odds are the 16-bit image
processed with scanner software may still be better than scanning
8-bit and processing in Photoshop. But it's still the roll of the
dice, I would say, although Kennedy will be around shortly to
contradict me... ;o)

I'd say, if you are concerned with quality and have the time, then by
all means scan at maximum depth and process later, but save the
original and work on a copy at least until you're completely satisfied
with the result. Many people keep these "digital negatives" anyway in
addition to the final edit.

If you set B & W
points intelligently so as not to clip information, and even curves to
change the tonality of some pixels, you are better doing so in 14 bits where
there are more levels to choose from and interpolating down, than taking a
raw scan that you are obviously going to have to make the same adjustments
to in PS and doing so in 256 levels. The gaps will be finer.

That's how I see it, anyway.

This absolutely does not apply to working in PS in a bit depth equal to or
higher than the scanner bit depth.

Right? Wrong?

If I understand correctly, you want to use scanner software to make
adjustments and scan in 14-bits and then scale down to 8-bit
afterwards. Right?

I'm assuming you will be doing this scaling down to 8-bit in Photoshop
and if that is the case you may just as well scan in 14-bit and do all
the editing in Photoshop... I mean, you're not saving any time by
doing the editing in scanner software, and you certainly have more
tools in Photoshop.

Anyway this has been a most useful conversation for me, and has clarified my
rather confused thinking on the issues.

That's good to hear. Nevertheless, take everything with a grain of
salt and run tests for yourself... And then bounce your results and
conclusions off of people to see what they think. That may not work
for all but that's how I learned the most.

But as I keep repeating what counts in the end is what works for you.
If your current workflow produces results you are happy with, that's
all that matters regardless of what anybody else says...

Don.

 

[Don]

I didn't realize that about interpolation.

If you check the thread I have exact instructions on how to (mis)use
;-) interpolation to get smooth histograms in both NikonScan and
Photopshop.

One word of warning, though, when you scale an image you lose quality.
In particular, the image loses sharpness.

That's why people usually suggest applying some Unsharp mask after any
scaling. But then you're already twice removed from the original image
just to avoid a gappy histogram.

And that's before you even started applying curves, contrast, etc.
which will then promptly "corrupt" this smooth histogram in no time...

Don.

 

[Don]

from Melbourne where I live due North to Longreach, Western Queensland.
Neat!

Our theme was "The Last Great Aussie Pub Crawl" because we stopped at
just about every pub.

Good on you mate! ;-)

When I lived in London (many, many, many moons ago) my best friend was
an Australian guy from Melbourne. I still remember his favorite saying
when he would bring the beers (and now use it myself):

Wrap yourself around this!

Of course, I would still be nursing my first beer while he would be on
his 3rd or 4th...

I took an EOS30 with just a 28-105mm and this has kept my scanner busy.

I've pretty much given up analog (an ancient Canon A1) since I got my
digital cameras (first Kodak, now Nikon) but since I started scanning
all my slides and negatives, I'm thinking of picking up my old A1
again.

Don.

 

[Kennedy McEwen]

We are not talking about Analog Gain. Analog Gain is on both sides of
the equation. We are talking about scanner software "curves, contrast
and friends..." vs. raw. Again:

Yes you are talking about analogue gain if you are discussing "changes
in your scanner software (regardless of bit depth)". You might not be
thinking that, but its real hard to work out what you are thinking from
what you are saying!