new scan snippets: 4990 vs. V700 vs. vs LS-8000 (same negative)
Raphael Bustin
Epson V700 v. Epson 4990 v. Nikon LS-8000 (Same Negative)
<http://www.terrapinphoto.com/jmdavis/>
About 2/3 of the way down the page.
The original is courtesy of Alan Bridgewater, taken
with Mamiya 645 Super, 80mm, on Fuji Reala.
Alan provided the V700 scan and mailed me the film
which I then scanned on my 4990 and LS-8000.
If you've got a V700 or V750 snippet to share, I'll
be happy to post it. See submission rules at the
top of the page (URL above.)
rafe b
www.terrapinphoto.com
rolan...@hotmail.com
Thanks for webbing this. The V700 scan looked a little sharper than the
4990 scan but the colours looked poorer.
Roger S.
surprisingly grainy for MF Reala, although only the LS 8000 even comes
close to showing the grain. The 8000 doesn't have any autosharpening
in the driver, does it?
Raphael Bustin
No, it does not, as far as I can tell.
This sort of grain isn't unusual for Reala on
the Coolscans. Reala is what I usually shoot,
so I'm fairly familiar with how it looks. Most
reversal materials (eg. E6 chromes) will show
substantially less grain and color noise.
With Reala, the most objectionable grain and
noise usually is found in shadow detail -- ie.,
from the *thinnest* portions of the negative.
Go figure....
rafe b
www.terrapinphoto.com
MLIDDELL
nikon.
win...@yahoo.com
for nothing' relative to the LS-8000.
Bart van der Wolf
<win...@yahoo.com> wrote in message
news:1146434181.3...@g10g2000cwb.googlegroups.com...
> Yes, the V700 still falls in the category of 'you can't get
> something
> for nothing' relative to the LS-8000.
From the sample shown, it appears to be the case. However, I don't
have any idea how well 'focused' by the V700 the film was. From what
I've read about it, it may be a bit difficult to determine/achieve the
optimal focus plane.
Bart
Raphael Bustin
>Yes, the V700 still falls in the category of 'you can't get something
>for nothing' relative to the LS-8000.
In all fairness, you should try sharpening the Epson V700 scan.
I'm not saying that will make up the difference -- I'm just saying
that the Epson flatbed/film scanners really take well to
sharpening -- at levels that the Nikon scans won't tolerate.
rafe b
www.terrapinphoto.com
rolan...@hotmail.com
It's the end product that counts, after all, so if it can be made to
look better through sharpening then so be it. What settings do you
recommend for sharpening at the different scan resolutions? Does the
software do a decent job or is it better done in Photoshop? Also for
resizing, is it better to send people the original scan if they might
resize it or sharpen what you've got at full size?
> rafe b
> www.terrapinphoto.com
Ernst Dinkla
Rafe,
Correct, I prefer Portra 160 - 400 and expose for 320 with the
last. That's the only way I can get acceptable shadows. And
the shadows do not improve with more scan sharpness like the
tweaked wet mount carrier of my Nikon 8000 delivers, in fact
this becomes worse. That's why I think that some films are
better suited for Epson's mix of over-sampling + the
acceptable loss in sharpness. In the end you can decide where
you like to stop in sharpening.
In view of a substantial change in the processing of analogue
film it surprises me that there has been so little R&D done on
color and B&W films meant for scanning. I have mentioned the
orange mask of color negative film that could be removed for a
scan film, another one could be a slide film that compresses
the dynamic range like a colorfilm does + get a better
latitude in exposure. Chromogenic B&W film already scans
better than conventional B&W but that has been pure luck as it
wasn't intended for that purpose. A reversal B&W slide film
with dynamic range compression would be more ideal too. The
last probably exists in some thick emulsion, two bath
development schemes, I have not checked that. With the right
scan films for MF + LF photography "analogue" photography
could compete much better with digital cameras for some time
to come.
Ernst
--
--
Ernst Dinkla
www.pigment-print.com
( unvollendet )
Mike
gap significantly.
Raphael Bustin
There are scan snippets (from the 4990) showing the Epson
scan "raw" and sharpened. Ultimately you have to be the
judge of how much sharpening you can get away with.
Just know that as you crank up the sharpness, the image
degrades in other ways.
I doubt very much if it makes any difference whatsoever
*where* you do the sharpening or resizing, but that
discussion quickly degrades into superstition and
metaphysics, splines, nearest neighbors, and so on.
By all means, download the scan samples and see how
they respond to USM.
rafe b
www.terrapinphoto.com
Alan Bridgewater
<f...@bar.com> writes
I think its time for me to take the plunge in these discussions. It was
I who submitted the v700 scan snippet and negative to Rafe.
For the sake of completeness, as soon as the negative comes back from
Rafe, I'll rescan it using the Epson scan software, apply sharpening and
then submit the new snippet.
Bart van der Wolf referred to focussing elsewhere in this thread. Film
flatness is a problem with the scanner, especially, as I'm finding, with
monochrome film.
The spacers that Epson supply with the film folders increase the
distance from the glass to the film surface in 0.5mm increments. There
was a discernible improvement in sharpness ( grain better defined) on
going from 3mm to 3.5mm, 2.5mm (no spacers) was awful. The snippet was
scanned with the spacers set at 3.5mm. I haven't yet tried increasing
the distance any further.
I'm fairly sure that some of my 120 monochrome negatives bow by more
than 0.5mm and the effect is quite obvious; grain is fairly well defined
at the edges of the affected images and becomes noticeably less obvious
towards the centre.
I successfully used a home-made glass holder with my old CanoScan 8400f
and I'm going to experiment with one for the v700.
I can't afford a Nikon LS-8000.
Alan
--
Alan Bridgewater
Raphael Bustin
<al...@thebridgewaters.co.uk> wrote:
>Hello all,
>
>I think its time for me to take the plunge in these discussions. It was
>I who submitted the v700 scan snippet and negative to Rafe.
>
>For the sake of completeness, as soon as the negative comes back from
>Rafe, I'll rescan it using the Epson scan software, apply sharpening and
>then submit the new snippet.
>
>Bart van der Wolf referred to focussing elsewhere in this thread. Film
>flatness is a problem with the scanner, especially, as I'm finding, with
>monochrome film.
>
>The spacers that Epson supply with the film folders increase the
>distance from the glass to the film surface in 0.5mm increments. There
>was a discernible improvement in sharpness ( grain better defined) on
>going from 3mm to 3.5mm, 2.5mm (no spacers) was awful. The snippet was
>scanned with the spacers set at 3.5mm. I haven't yet tried increasing
>the distance any further.
>
>I'm fairly sure that some of my 120 monochrome negatives bow by more
>than 0.5mm and the effect is quite obvious; grain is fairly well defined
>at the edges of the affected images and becomes noticeably less obvious
>towards the centre.
>
>I successfully used a home-made glass holder with my old CanoScan 8400f
>and I'm going to experiment with one for the v700.
>
>I can't afford a Nikon LS-8000.
Hello Alan,
Thanks again for sending the film. It arrived Friday
evening and I hope to have it back in the mail to you
either later today or tomorrow morning.
It will be curious to see if additional care in focusing
can improve the V700 scan. There's no question that
at high resolutions and MTFs, depth-of-focus and
film flatness are big issues.
On the LS-8000, optimal flatness across the frame
takes great care (and/or a glass carrier that costs
another $240.)
The LS-8000 also has a motorized focus adjustment
and measuring scheme, so even on "wavy" piece of
film, you can still get near-perfect focus on a
particular area of the film.
Which kind of wraps back around to the whole
discussion of wet-mounting.
rafe b
www.terrapinphoto.com
Peter Chant
> You need to apply Unsharp Mask to the flatbed scans. It will close the
> gap significantly.
>
Make me quite chuffed. I definitely could not justify the expense of the
Nikon and I would not expect the 4990 to be as good, given the price
increment. However, its not bad in comparison.
juanster
link:http://www.photo-i.co.uk/Reviews/interactive/Epson%20V700/page_13.htm
The author talks about a height adjustment for focus that could make a
difference in this test.
Using the pull down menu the entire test can be viewed. The V700 looks
good in this test.
juanster
rolan...@hotmail.com
I haven't got a copy of Photoshop but what I have noticed (again it
might be more superstition than science) is that I seem to get better
results with my 8M digital camera having no internal sharpening applied
(supposedly) and leaving it on low contrast and doing the sharpening
after I have resized.
MLIDDELL
place.
Bart van der Wolf
"MLIDDELL" <marko....@gmail.com> wrote in message
news:1146592649....@u72g2000cwu.googlegroups.com...
> Possibly, but you still cannot sharpen what isn't there in the first
> place.
Fortunately there usually is more information hidden in the data than
meets the eye at first glance ...
There are plenty examples demonstrating that, the real question
therefore becomes; "How much information is hidden in plain sight?".
http://www.mathworks.com/products/demos/image/ipexlucy/ipexlucy.html
Bart
Mike
Paul Saunders
> It's the end product that counts, after all, so if it can be made to
> look better through sharpening then so be it.
The edges can be made sharper, but you can't improve fine texture detail.
> What settings do you
> recommend for sharpening at the different scan resolutions?
I wouldn't think in terms of scan resolution if I were you, it's the output
resolution that matters.
> Does the
> software do a decent job or is it better done in Photoshop?
Definitely the latter, and definitely do it last of all. Only sharpen
during the scanning stage if you don't intend to any processing and don't
intend to resize it. Sharpening degrades the image, any further processing
or resizing simply ehances that degradation.
> Also for
> resizing, is it better to send people the original scan if they might
> resize it or sharpen what you've got at full size?
Best to send the unsharpened image. Publishers and photo agencies usually
insist on unsharpened images, they prefer to resize them and use their own
sharpening techniques to suit their own printing requirements.
I do printing for a few people and always tell them not to sharpen. I save
all of my processed "master" images unsharpened. You can always add
sharpening later, but you can never take it away. It's practically
impossible to undo the "damage" of a sharpened image.
Paul
win...@yahoo.com
just has better defined detail, colors etc. I appreciate the
difference in price, but Epson is misleading people when they imply
that their flatbeds can compete with high quality dedicated film
scanners.
I have an Epson 3200, and it was 'spanked' by a Minolta Dual Scan III
when it came to scanning 35mm film. Of course the Epson's are cost
effective especially for large format film. But I would think at any
format they still lack good color reproduction.
David J. Littleboy
"Paul Saunders" <pv...@wildwales.fsnet.co.uk> wrote in message
news:e399cp$796$1...@news.freedom2surf.net...
> rolan...@hotmail.com wrote:
>
>> It's the end product that counts, after all, so if it can be made to
>> look better through sharpening then so be it.
>
> The edges can be made sharper, but you can't improve fine texture detail.
I find that sharpening LS-8000 scans does wonders for fine texture, but that
I don't care for what it does to edges, which are largely already of
adequate contrast. (I.e., I hate halos.)
But I find theis true for almost any digital image, not just LS-8000 scans.
David J. Littleboy
Tokyo, Japan
Paul Saunders
> I find that sharpening LS-8000 scans does wonders for fine texture,
> but that I don't care for what it does to edges, which are largely
> already of adequate contrast. (I.e., I hate halos.)
I wonder, is there some way of using the Stylise - Find Edges filter to
create some kind of mask to inhibit sharpening on the edges?
Paul
Peter Chant
> I wonder, is there some way of using the Stylise - Find Edges filter to
> create some kind of mask to inhibit sharpening on the edges?
But then what does it sharpen? Surely it can only sharpen the edges and it
does this by boosting contrast at them. If you mask it it does not work.
Pete
Scott W
Peter Chant wrote:
> Paul Saunders wrote:
>
>
> > I wonder, is there some way of using the Stylise - Find Edges filter to
> > create some kind of mask to inhibit sharpening on the edges?
>
> But then what does it sharpen? Surely it can only sharpen the edges and it
> does this by boosting contrast at them. If you mask it it does not work.
>
Sharpening does much more then sharpen edges, in fact it is not all the
good with edges. Where sharpening works the best is getting textures
back that are low in contrast, due to soft images. And when you
sharpen grass, sand or gravel as examples you don't have to worry about
halos.
Scott
David J. Littleboy
"Peter Chant" <pe...@petezilla.co.uk> wrote:
> Paul Saunders wrote:
>
>> I wonder, is there some way of using the Stylise - Find Edges filter to
>> create some kind of mask to inhibit sharpening on the edges?
>
> But then what does it sharpen? Surely it can only sharpen the edges and
> it
> does this by boosting contrast at them. If you mask it it does not work.
Edges come in a wide range of contrasts. Sharpening seems to boost the
contrast more for edges that are already high contrast. This is exactly the
opposite of what you want; you want edges that are hard to see to be made a
tad clearer. So people mask out the high contrast edges to protect them from
being given "halos".
Bart van der Wolf
"David J. Littleboy" <dav...@gol.com> wrote in message
news:e3olgm$nlj$1...@nnrp.gol.com...
SNIP
> Edges come in a wide range of contrasts. Sharpening seems to boost
> the contrast more for edges that are already high contrast. This is
> exactly the opposite of what you want; you want edges that are hard
> to see to be made a tad clearer. So people mask out the high
> contrast edges to protect them from being given "halos".
I fully agree with that, one doesn't need sharpening of edges that are
sharp already. It'll only lead to halo artifacts and those artifacts
add nothing to perceived sharpness. It just cries-out "I've been
over-sharpened, image quality has been compromised".
That's exactly why I recommend to blend the image with a luminosity
sharpening layer like this on top:
<http://www.xs4all.nl/~bvdwolf/main/downloads/Non-clipped-sharpening.png>
The small radius sharpening used in the example prevents very
wide/visible halo, but still clips all detail into a featureless halo
when the edge contrast is already high with "regular sharpening". With
a minor adjustment to a sharpening layer, this can be reduced to
enhanced medium to low contrast, and progressively less enhancement as
contrast is higher already, which normally prevents clipping
altogether.
The progressively blended sharpening layer is also highly adjustable
by changing opacity or by local masking, and it avoids color
sharpening artifacts.
--
Bart
Randy Howard
(in article <446096b7$0$31648$e4fe...@news.xs4all.nl>):
What do you think of using a mild high-pass filter on a layer
copy above, then blending with hard or vivid light? For some
images that seems to do a nice job without making it overly
obvious it's been sharpened.
--
Randy Howard (2reply remove FOOBAR)
"The power of accurate observation is called cynicism by those
who have not got it." - George Bernard Shaw
Paul Saunders
>> I wonder, is there some way of using the Stylise - Find Edges filter
>> to create some kind of mask to inhibit sharpening on the edges?
>
> But then what does it sharpen?
It sharpens the texture parts of the image. Well it would if I could figure
out a practical way of doing it.
> Surely it can only sharpen the edges
No, sharpening affects everything, but it's most obvious on the edges.
> and it does this by boosting contrast at them. If you mask it it
> does not work.
That's the whole point, David J. Littleboy commented that his edges are
already sharp enough, but that the textures can still benefit from
sharpening.
The edges are areas of high contrast, where adjacent pixels have large
differences in brightness. What I'm calling "texture" is areas of fine
detail where adjacent pixels have small differences in brightness, for
example blades of grass or the surfaces of rocks (unfortunately this also
applies to film grain in the sky).
When using USM you can increase the threshold to limit sharpening to the
edges only, thus preventing grain from being sharpened, but this also
prevents fine detail like grass being sharpened. But the reverse is not
normally possible, to sharpen the texture without sharpening the edges,
which as David said, may already be sharp enough.
Actually, there is a solution available in the form of the FocalBlade
plug-in. I quote; "FocalBlade sharpens edges and surfaces in an image
independently" (they use the term "surface" instead of "texture").
http://thepluginsite.com/products/photowiz/focalblade/index.htm
It even has the ability to suppress halos, which might please David. This
is a very powerful plug-in, well worth a look. I'm still undecided myself
since I like Focus Magic so much.
While we're on the subject, take a look at their excellent ColorWasher and
LightMachine filters too (click on the Example links), and no, I don't work
for them.
http://thepluginsite.com/products/photowiz/index.htm
Paul
Paul Saunders
>> I wonder, is there some way of using the Stylise - Find Edges filter
>> to create some kind of mask to inhibit sharpening on the edges?
>
Simon Meeds
Generally smooth images may require a little sharpening, and generally
detailed images may require a relatively large amount of sharpening. Images
where there is a mixture of the two styles may require sharpening to the
edges only. To do this, first create a copy of the image. Use "find edges"
and taking the resulting image increase the contrast so that the edges
appear bright and the rest dark, next apply some Gaussian blur and then
apply to resulting image as a mask to the original image. Now apply
sharpening to the image.
If you want to apply sharpening to everything but the edges (though I agree
it is a rather strange thing to want to do) you could apply the inverse of
the mask to the original image.
Simon
"Paul Saunders" <pv...@wildwales.fsnet.co.uk> wrote in message
news:e3qsrf$o1t$1...@news.freedom2surf.net...
Bart van der Wolf
"Randy Howard" <randy...@FOOverizonBAR.net> wrote in message
news:0001HW.C08631A6...@news.verizon.net...
SNIP
> What do you think of using a mild high-pass filter on a layer
> copy above, then blending with hard or vivid light?
A high-pass filter can also produces some halo, unless it is an
exactly defined HP-filter. I occasionally produce HP-filters when I've
got an exact analysis of the blur-function, so I can neutralize the
effect by "inverting" the Point Spread Function (=~blur) kernel, with
the exception of the central kernel pixel.
> For some images that seems to do a nice job without making
> it overly obvious it's been sharpened.
In case of HP-filter layers, I use an overlay (or soft light) blending
mode because it comes closer to a predictable behavior, to me anyway.
But if you can get the desired result by using another blending mode,
then who am I to object ;-).
My goal remains to avoid halo and avoid clipping, because it looks
more natural, and I can do that automatically with the method I linked
to earlier.
Bart
Randy Howard
>
> "Randy Howard" <randy...@FOOverizonBAR.net> wrote in message
> news:0001HW.C08631A6...@news.verizon.net...
> SNIP
>> What do you think of using a mild high-pass filter on a layer
>> copy above, then blending with hard or vivid light?
>
> A high-pass filter can also produces some halo, unless it is an
> exactly defined HP-filter. I occasionally produce HP-filters when I've
> got an exact analysis of the blur-function, so I can neutralize the
> effect by "inverting" the Point Spread Function (=~blur) kernel, with
> the exception of the central kernel pixel.
whoosh. I have no idea what you just said. :-)
>> For some images that seems to do a nice job without making
>> it overly obvious it's been sharpened.
>
> In case of HP-filter layers, I use an overlay (or soft light) blending
> mode because it comes closer to a predictable behavior, to me anyway.
> But if you can get the desired result by using another blending mode,
> then who am I to object ;-).
I seem to find the correct mode varies with the image in
question, no hard and fast rule seems to be evident.
> My goal remains to avoid halo and avoid clipping, because it looks
> more natural, and I can do that automatically with the method I linked
> to earlier.
Ok. I'm definitely going to try that out as well.
Bill Carr
9000), although wonderful, can't handle anything larger than 120.
"Alan Bridgewater" <al...@thebridgewaters.co.uk> wrote in message
news:Izmp1qHJ...@thebridgewaters.co.uk...
RaphaelBustin
>
>Yes, the V700 still falls in the category of 'you can't get
>something
>for nothing' relative to the LS-8000.
>
In all fairness, you should try sharpening the Epson V700 scan.
I'm not saying that will make up the difference -- I'm just saying
that the Epson flatbed/film scanners really take well to
sharpening -- at levels that the Nikon scans won't tolerate.
rafe b
www.terrapinphoto.com
-----------------------------------------
Posted at: http://ForumBreak.com