Lightjet Print report
Charles Pezeshki
Thought the ng might be interested in a report on some 16x20 Lightjet prints
I got done recently.
FYI, I've been getting ready for my first show-- 40 pictures and posters in
my local art gallery/community center. All the prints 8x10- 12x15 were done
on my Epson 1270, I had eight printed on a photo poster printer (color
'matted' with text, 2'x3'), and I had 4 printed on the Lightjet (16x20s). I
shoot LF Velvia, Astia, and Provia F exclusively-- no negative film, no B/W.
Nancyscans (www.nancyscans.com) scanned nine of my 6x12 and 4x5
transparencies with their Tango drum scanner. They also did the Lightjet
printing. I scanned the rest using my Microtek 4 scanner, and have a friend
with a 5, which I also used.
Here's my rundown (and yes, I know about the possible color problems with
the Epson 1270).
1. There is no color difference between the Epson and the Lightjet. None.
Neither looks better than the other, as long as one is over 200 dpi on the
scan.
2. The quality of the Tango scans was outstanding-- but not significantly
better than the 5, except for the increased resolution. Only on
difficult-to-scan transparencies was there a noticeable difference. The
great thing about the Tango scans was that Nancyscans did the color
matching, and they did an excellent job. Conclusion? For difficult to scan
transparencies, go with the Tango. For dpi, go for the Tango. Otherwise,
use the 4 or 5 (the 5 is definitely better than the 4, even for contrast).
It's perfectly fine for 80% of the work. The 6x12s definitely did better
with the Tango-- these are hard to hold with the Microtek holders.
Occasionally, the 4 produces artifacts. The Tango produces no artifacts.
These usually only affect one scan line, and are not noticeable.
3. There is no sharpness difference between the Epson and the Lightjet.
4. The two drawbacks to the Epson are-- lack of surety regarding the
archival nature (we're framing and matting ours right now, to take care of
the various chemical problems) and as the printer ages, there seems to be
small defects (unnoticeable except by me) that crop up-- pizza wheels, some
aliasing phenomena, etc. They're not consistent, and not noticeable. There
are NO defects in the Lightjet prints.
5. Color matching between the Tango scans and the Lightjet prints are
exact. Color matching between my profiled monitor and the Lightjet prints
are exact. One of my prints appeared a hair dark, and it had on the
monitor. In the future, I'll use my calibrated monitor with confidence in
adjusting both brightness and color.
6. Nancyscans is competing on price, and quite frankly, I'm having a real
hard time seeing why anyone who understood the technology would go with a
more expensive printer if you want a Lightjet print. $68 for the first, and
$27 for consecutive was what I paid for the prints.
7. I love the latitude the high scan density of the drum scans give you.
You pay, though-- $50/scan. The Microtek won't be going anywhere soon. I
did great with my Epson 1270 and the Microtek 5 on prints to 12"x16",
though. 234 dpi was indistinguishable from any higher resolutions.
8. If I owned a Lightjet, I sure would hope that I'd make my money off it
in the next 2-3 years. The only edge is the assured archival nature and
potential large sizes. The desktop Epson did as well with the smaller
sizes.
Overall, I can't even fathom going back to dealing with traditional chemical
printers. The quality of all this digital stuff is wild.
Just my 2 cents. Hope this helps folks. This isn't intended to be a formal
study-- just my experience.
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
Dan Baumbach
I don't know much about the Microtek scanners. Are they flatbed scanners
with transparency holders? I shoot with 35mm and scan with a Nikon LS 2000.
I've been waiting for them to come out with a higher resolution scanner and
in my impatience, I've been looking at Nancyscans for drum scans.
-- Dan
--------------------------------------------------
Dan Baumbach, Photography
www.timelesslight.com
"Charles Pezeshki" <peze...@mme.wsu.edu> wrote in message
news:B5FB5F91.6502%peze...@mme.wsu.edu...
Todd Caudle
weeks to see the Arizona Highways 75th anniversary exhibit, featuring Ansel
Adams, David Muench and Jack Dykinga (photo gods, all). The Muench and
Dykinga large prints are Lightjets. Can't wait! I'm looking forward to drum
scanning and Lightjetting some of my 6x7cm and 4x5 trannies, too. Your
experience is very enlightening.
Brian Legere
contemplating getting some lightjet prints done.I would like to here from
others.Are the lightjet prints really that good?how are they for b+w prints?
Brian
Mike Long
Great report! It gives us the chance to compare and contrast, using
your reference data as a starting point for our own work.
I'd like to see Dale Lab's "Pro" scans in comparison. They average $22
per scan. You must use the "Pro" instead of the "Master" for medium
format and large format and they suggest it if going over 8x10 for
prints from 35mm (I find 10x14 prints from a full 35mm neg are just
fine).
Since they are less than 1/2 the price, I'd say try them. If I get
comparison results first, I'll post them.
Mike
Gene Anderson
> I recently bought an epson 1270.I am quite pleased with the quality.I have been
> contemplating getting some lightjet prints done.I would like to here from
> others.Are the lightjet prints really that good?how are they for b+w prints?
No idea on the B&W prints, but for color, yes, a drum scan plus a
LightJet print really can be that good. At Galen Rowell's Mountain Light
studio here in the SF Bay Area, he had a display of a 70mm dupe of a
slide on a light box, a traditional print, and a LightJet print of the
same image displayed side-by-side. There was no contest -- the LightJet
print kept so much more detail, especially in the shadows, and looked so
much more like the the transparency than the traditional print that it
was staggering. The LightJet print looked almost like a print-sized
transparency. His prints looked great in the past, but *so much* better
with a drum scan and LightJet print that he'll never go back.
I'm still trying to decide which of my images to splurge on...
gene
--
Gene Anderson
ge...@designsinlight.com
http://www.designsinlight.com/
Keith Clark
Can you describe your monitor calibration process? Are you on a Mac or Wintel
machine?
Keith
Greg Ferguson
Caudle" <tcf...@home.com> wrote:
I just saw the show for the second time.
Muench's works are unbelievable. They explode off the paper.
Dykinga's are um... too rich. His colors almost go neon in some prints.
The show is definitely worth seeing though. Our class (the Friends of
Arizona Highways Sedona trip) was there for several hours last Monday.
It was fun going as a group and talking about the compositions.
Anyone lurking who wants more info can find it at:
http://www.azhighways.com/NewFiles/anniv.html
Greg
Wisbl
>
> Great(!) report, Chuck.
>
It's my understanding that Lightjet prints are outputted on Fuji
Chrystal Archive paper. That paper is estimated to last 60 years tops.
That doesn't sound archival to me.
William
Brian Ellis
I had 10 transparencies and wanted one print from each, it would cost $680
for the ten prints? I don't do color but If that's the case, I sure could
fathom going back to traditional darkroom work in a hurry thank you.
"Wisbl" <isabe...@earthlink.net> wrote in message
news:39D570...@earthlink.net...
geod
Lightjet for B&W looks flat, uninteresting and, unless you like photographs
that look clipped from a glossy magazine, horrible. Hand-made fiber prints
have so much more depth and richness that there's no comparison. The medium
simply isn't suited to black and white.
Mike Long
Nice to have some good weather for a change - it was ALMOST cool in SE
Florida this weekend.
I paid $100 each for 16x20 Ilfochromes from Holland Labs a year ago.
Custom quality color printing is very pricey. Even at home (which I've
given up now), it adds up from all the little hardware and supplies.
The inkjet printers, permanent inks and what's still to come, is a
real bonus for those of us who like color.
If there's demand, B&W will get better also. I haven't heard of anyone
using the Quad inks with the 2000P yet (and I don't know if the combo
even works or not).
Mike
On Sun, 01 Oct 2000 13:39:59 GMT, "Brian Ellis"
<bell...@earthlink.net> wrote:
>Are you saying that the first one of each print cost $68? In other words, if
>I had 10 transparencies and wanted one print from each, it would cost $680
>for the ten prints? I don't do color but If that's the case, I sure could
>fathom going back to traditional darkroom work in a hurry thank you.
>"Wisbl" <isabe...@earthlink.net> wrote in message
>news:39D570...@earthlink.net...
Charles Pezeshki
The Microteks are flat bed transparency scanners. Do remember that I'm
shooting 4x5 and 6x12. They wouldn't work well at all for 35mm, and they
work only marginally for MF work. The quality is fine, but the resolution
doesn't cut it.
I don't know what the deal with drum scanning 35mm is-- they must have to
remove the slide from the mount in order to do that.
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
> From: "Dan Baumbach" <da...@timelesslight.com>
> Organization: EarthLink Inc. -- http://www.EarthLink.net
> Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> Date: Sat, 30 Sep 2000 17:27:19 GMT
> Subject: Re: Lightjet Print report
>
> Thanks for sharing your experiences with us. I find them extremely helpful.
> I don't know much about the Microtek scanners. Are they flatbed scanners
> with transparency holders? I shoot with 35mm and scan with a Nikon LS 2000.
> I've been waiting for them to come out with a higher resolution scanner and
> in my impatience, I've been looking at Nancyscans for drum scans.
>
> -- Dan
> --------------------------------------------------
> Dan Baumbach, Photography
> www.timelesslight.com
>
>
>
> "Charles Pezeshki" <peze...@mme.wsu.edu> wrote in message
> news:B5FB5F91.6502%peze...@mme.wsu.edu...
Charles Pezeshki
> From: ml...@bellsouth.net (Mike Long)
> Chuck,
>
> Great report! It gives us the chance to compare and contrast, using
> your reference data as a starting point for our own work.
Glad ya liked it. It's been pretty educational for me getting all these
results in front of me in one place. All I know is that it's not long until
those of us that do color will have equivalent-to-pro equipment in our homes
for less than $5000 to do 24" wide prints. In fact, the time is pretty much
here.
>
> I'd like to see Dale Lab's "Pro" scans in comparison. They average $22
> per scan. You must use the "Pro" instead of the "Master" for medium
> format and large format and they suggest it if going over 8x10 for
> prints from 35mm (I find 10x14 prints from a full 35mm neg are just
> fine).
>
It's important to remember that I'm shooting LF, which, while being more
expensive in practically everything, saves money cuz I can get by with a
cheaper scanner. The Microteks wouldn't work at all for 35mm for print
work.
> Since they are less than 1/2 the price, I'd say try them. If I get
> comparison results first, I'll post them.
>
> Mike
So you're comparing apples with apples, my Tango scans are 200Mbyte scans.
That's pretty big. Nancyscans' 150 MByte scans are $39.95. I like the 200M
scans cuz they are as big as I would ever need, and lets me do 30"x40" at a
Lightjet resolution.
Chuck
Charles Pezeshki
The reason I gave prices was so that people could compare Lightjet prices
with other Lightjet prices. Some pro shops charge $150+ for these prints.
I offered my experience with a discount house.
Traditional darkroom work might be great if one was doing color constantly,
and one had lots of experience. It's simply impractical for some of us,
though. First off, my house is really small, and I'd have to build a
darkroom--and I'd never recover that $10K cost, not even figuring the cost
of enlarger, chemicals, learning curve, etc..
The great thing about digital processes is that if they're calibrated
correctly, WYSIWYG on the first print round. That's my main point.
There's also a way to prepare the scans so that Nancyscans only charges the
second print price ($27) as opposed to the first print price. When I get
that figured out, I'll post it. For this time, I had to turn these around
immediately.
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
> From: "Brian Ellis" <bell...@earthlink.net>
> Organization: EarthLink Inc. -- http://www.EarthLink.net
> Reply-To: "Brian Ellis" <bell...@earthlink.net>
> Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> Date: Sun, 01 Oct 2000 13:39:59 GMT
> Subject: Re: Lightjet Print report
>
Brian Ellis
familiarity, since it's been about five years since I've done color darkroom
work and have never used a service bureau for my limited about of digital
stuff, but it strikes me that there's a rather huge difference between $2.60
a print and the $68 to $150 per print (plus $50 per scan) that has been
mentioned in this thread. I don't really care much one way or the other
since I only do black and white work. I was just expressing my amazement at
the per print cost and I'm still amazed. I can understand professionals
doing all this, the customer pays in the end, but for serious amateurs who
have no one else to pay the cost, I would think the cost is prohibitive for
most on any kind of quantity basis.
"Gregory Blank" <"g;-)?????????????????? wrote in message
news:g;-)blank-01100...@client-151-196-123-1.bellatlantic.net...
> Not to mention printing in your own darkroom. Lets see if I print RA4
> myself thats $ 70 for 16x20 paper and $60 for chemistry thats $2.60 per
> print in perishable materials.
>
>
> In article <P8HB5.1169$lf5....@newsread1.prod.itd.earthlink.net>, "Brian
> Ellis" <bell...@earthlink.net> wrote:
>
> Are you saying that the first one of each print cost $68? In other words,
if
> I had 10 transparencies and wanted one print from each, it would cost $680
> for the ten prints? I don't do color but If that's the case, I sure could
> fathom going back to traditional darkroom work in a hurry thank you.
> "Wisbl" <isabe...@earthlink.net> wrote in message
> news:39D570...@earthlink.net...
>
> --
> Gregory W.Blank Photography
> P.O. Box 726
> Finksburg, MD. 21048
> Check out my website http://members.bellatlantic.net/~gblank
>
RCoda
...they could output a BW scan (corrected, burned, dodged, etc) to some
large format BW duplicating film in a LightJet, process that film normally,
then take it home and make LARGE BW contact prints.
I've been trying to find a lightjet operator who's willing to try this, but
no one wants to touch it.
Rich
"geod" <ge...@cwo.com> wrote in message news:B5FCC5F9.8890%ge...@cwo.com...
> I've said it before and I'll say it again:
>
> Lightjet for B&W looks flat, uninteresting and, unless you like
photographs
> that look clipped from a glossy magazine, horrible. Hand-made fiber
prints
> have so much more depth and richness that there's no comparison. The
medium
> simply isn't suited to black and white.
>
> > From: Wisbl <isabe...@earthlink.net>
> > Organization: EarthLink Inc. -- http://www.EarthLink.net
> > Reply-To: isabe...@earthlink.net
> > Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> > Date: Sun, 01 Oct 2000 04:47:16 GMT
> > Subject: Re: Lightjet Print report
> >
Keith Clark
> I'm really talking about something with which I don't have a lot of
> familiarity, since it's been about five years since I've done color darkroom
> work and have never used a service bureau for my limited about of digital
> stuff, but it strikes me that there's a rather huge difference between $2.60
> a print and the $68 to $150 per print (plus $50 per scan) that has been
> mentioned in this thread. I don't really care much one way or the other
> since I only do black and white work. I was just expressing my amazement at
> the per print cost and I'm still amazed. I can understand professionals
> doing all this, the customer pays in the end, but for serious amateurs who
> have no one else to pay the cost, I would think the cost is prohibitive for
> most on any kind of quantity basis.
Actually, if you visit the NancyScans site and read the FAQ, you'll find out
that they will give you "second print" prices upfront if the submitted job meets
certain criteria (none of which is difficult).
http://www.nancyscans.com/profiling.html
And $27 a print for 16x20's is hardly "prohibitive" for a print that will last
twice as long as a Cibachrome.
http://www.wilhelm-research.com/4/4.html
The reason why all of this is attractive is that for those of us who are not
chemists can now get exactly what we want the first time out of the lab.
When was the last time you described to a "chem lab" what you wanted in terms
of density and color balance and got back exactly the right print? It doesn't
happen often, does it.
This technology puts the photographer, NOT THE LAB, in the driver's seat. Want
the barn redder- without affecting anything else? Not a problem. Impossible in a
chem lab.
Are drum scans worth the price? Yes. Drum scan some medium format Velvia and
print it on an Epson @ 13x19 and hand it to almost anyone - what you'll
immediately hear is "awesome! How'd you make it so clear and full of details?".
Is it worth it to serious amateurs? Well, over a decade ago I once paid $75 for
an 11x16 Cibachrome of my pet parrot... ;>
Keith
geod
enlargements more precisely than optical, so producing a high quality 8x10
transparency should be possible.
However...
It would be a generation removed from the original and, honestly, would a
contact 8x10 from such a procedure be necessarily better than a 2x
enlargement from 4x5? Would it be worth the trouble and expense? I doubt
it.
Charles Pezeshki
I'd agree with you that you really need to be a pro in order to make this
kind of thing worth one's while. It's my opinion that for me, to shoot
color LF, then I need to be marketing at least some of my work in order to
cover costs. That's what I do. I'm not in the league of some of the real
pros on this NG, but I do sell some stock, and my goal is to make my hobby
pay for itself. It's taken a little effort, but I've been successful.
I've also gotten grants that have paid for my equipment (I'm an
environmental activist heavily involved with forest issues), and I've also
received a grant from the Idaho Commission on the Arts for my work. There
are ways to take the sting out of this obsession, AND shoot the subject
matter that I love. I'd encourage others to look at their local arts
commissions. There's $1K chunks of money out there that can be relatively
easy to get, if you're willing to shoot a coherent set of photos.
The point I was making is that, at least where printing cost is concerned,
one can basically avoid Lightjet prints and print on the 1270 if the
red-shift issue is gotten around (by matting and framing quickly, if that's
a problem). And the next year, I believe, will produce a printer that
doesn't have this problem at all.
There is, of course, the issue of the equipment. But folks routinely go out
and spend more on a 4-wheeler and trailer in this neck of the woods, not
counting the $30K Dodge Ram pickup to pull it.
B/W I understand as something resembling a reasonable hobby. For my income
bracket, color must pay for itself. This show was a hardship because of the
cost of printing and framing everything. I'm hoping that I can sell enough
to cover my costs-- once again.
And if not-- well, hey! I'll have gifts for this Christmas and a bunch of
great stuff for my wall!
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
> From: "Brian Ellis" <bell...@earthlink.net>
> Organization: EarthLink Inc. -- http://www.EarthLink.net
> Reply-To: "Brian Ellis" <bell...@earthlink.net>
> Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> Date: Mon, 02 Oct 2000 10:25:29 GMT
> Subject: Re: Lightjet Print report
>
Keith Clark
Thanks for your email reply to my question on monitor calibration. Too
bad you're on a Mac. I was hoping you were a long suffering pee-cee user
like me (laughing ;>).
Anyway -
Could you comment on the "grant scene"?
I'm sure many of us would like to explore that in detail. I for one, am
clueless as to where to go for more information. Some of those 1K chunks
of change would help...maybe even pay for a 2000P ;>
Cheers,
Keith
> Chuck
> Chuck Pezeshki
> http://users.moscow.com/pezeshki
>
Charles Pezeshki
I've successfully gotten grants to fund part of my hobby. Part of the
reason I've gotten grants is past success as an enviro activist. I have a
very good reputation as a producer (I was one of the people that did the
media work that embarrassed the US Forest Service that led to the
roadbuilding moratorium/ Clinton's roadless initiative on NF ground).
Because of this, I wrote some requests for photo equipment as the place I
defend (primarily), the Clearwater Country in Idaho, has virtually no photo
stock of itself. Well, now it does. No pro (in their right, fiscal mind)
would ever spend the time and effort to document these landscapes-- there is
basically no money in commercial stock sales. But for myself, only looking
to cover my costs, and protect my place, such a project has great appeal.
In all fairness, I can't recommend folks going after that kind of money
unless they have a considerable activist reputation. And as far as hours
are concerned, if I didn't want to do the enviro work, I would have done
better working at Burger King and saving my money for LF equipment. Still,
it was a perk for writing my book, 'Wild to the Last: ENvironmental
Conflict in the Clearwater Country', which is available at lots of libraries
in the Pacific NW, most university libraries, and amazon.com, if yer really
interested.
The arts grants scene, however, is totally different. Every state has an
arts commission. All these commissions give a variety of grants. Some
involve education of schoolchildren (why not give a photo workshop in the
inner-city?) some involve 'fine art', all involve putting together a little
portfolio, visiting the people in charge, and writing a grant. None of it
is hard. Type '(name of state) arts commission' in Yahoo and watch what
pops up. Because my art is very politically relevant (if not commercially
so) in Idaho, and I committed to putting on a show, they gave me $600. I
also get listed as a funded artist (which is nice) and the $600 bought
Quickloads for the summer. I'm sure they'll give me $600 or more next year,
cuz I came through.
The bad news is the show costs around $1500 to put on. So I'm still
losing-- but hopeful that I'll sell enough, once again, to cover costs.
Photography of something other than dogs engaged in obscene acts always has
the potential for local funding, if done well, and is made to be relevant.
I'm sure one could get money for photographing street people in downtown
Portland. You're less likely to get money for photographing National Parks.
Lots of folks do this already, and make money on it. If you're really good,
and you're photographing commercially viable material, the arts commission
folks are gonna figure it out, and expect you to pay your own way. But if
your interest runs to public service, there are ways of covering your costs.
To recap-- your local/state arts commission is a great place to start.
People everywhere are used to trash photography. Just showing up with good
stuff was a big eye-opener for me. I also talked to the folks with a stack
of my 2'x3' posters, which told about the place, and something regarding its
environmental history. Folks loved this--this was clearly different than
some guy shooting solely for his den room wall. My art had a contemporary
relevance (even though it was scenic landscapes) and I also agreed to do
something in the public interest (put on a show about roadless landscapes),
which will be written up in the local papers.
The fact is I probably would have done most of this anyway, so the
necessities behind pursuing the money worked for me. They may not work for
everyone. But you won't get it if ya don't ask!
Chuck
--
> From: Keith Clark <ClarkPho...@hotmail.com>
> Organization: http://www.clarkphoto.com
> Reply-To: ClarkPho...@hotmail.com
> Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> Date: Mon, 02 Oct 2000 11:23:25 -0700
> Subject: Re: Lightjet Print report
>
Roger N Clark
Have you ever tried to burn or dodge a contact print? It's a little trickierhttp://www.reedphoto.com has LVT film output. I don't know cost,
than an enlargement because you can't see what you trying to enhance as the
negative is below you, not above you. This way, you can do all that in
Photoshop ONCE.And, it would be a drum scan of an already large 4x5 or 8x10 negative. So
quality is not an issue.Anyway, I was talking about really big negatives, 16x20 and up. For 8x10 you
can get an LVT or FIRE (LightJet 2000).
but the advertise both color and black and white output up to 11x14.
And the real advantage of digital over
traditional custom dodge and burn is that
you can then make a lot of identical
(i.e perfect) prints compared to
dodging each one by hand (and
therefore each slightly different
and mostly imperfect).
Roger Clark
Roger N Clark
Charles Pezeshki wrote:
Here's my rundown (and yes, I know about the possible color problems with
the Epson 1270).1. There is no color difference between the Epson and the Lightjet. None.
Neither looks better than the other, as long as one is over 200 dpi on the
scan.
I find this hard to believe because
every medium has slightly different
color properties, even between
photographic papers. So when you say
there is no color difference,
do you mean from a general viewpoint considering
the inherent color differences
between media, or do you mean that
every shade/hue of red, green,
blue looked exactly the same?
If it really is the same,
this must be due to calibration
of your monitor/printer/lightjet?
But even a calibrated monitor looks different
to me than any print because a monitor is emitted light
versus reflected light from paper,
which looks different in different lights.
Could you clarify things a little?
I think it would be nice if there were a
small device you could put on your
monitor and on the print so it could
read the color, contrast values and make
the calibration very robust. I've heard
rumors that such a device would soon come
to market. Anyone heard of it?
I've been able to "calibrate" my system
(hp printer, monitor and lightjet output)
by iterative trial. I've found that the photo
papers (both fuji crystal archive and
cibachrome) have very subdued greens
compared to my scanned images
(of velvia). But I've learned how to
match my monitor and inkjet printer
output by boosting saturation ~20% for
the images I send to the lightjet.
2. The quality of the Tango scans was outstanding-- but not significantly
better than the 5, ..snip...The Nancyscans pricing looks pretty good. What is their pricing
for large scans (they say call over 200 MB). I get 650 MB scans
of 4x5 for about $100.
3. There is no sharpness difference between the Epson and the Lightjet.
But the lightjet is more continuous tone, so I think you'll
find that on some colors (and this is getting rarer as inkjets
get better) that the lightjet will be better. The greatest
differences has been in skin tones in my experience. But
I have
no experience with the latest epsons.
4. There
are NO defects in the Lightjet prints.
6. Nancyscans is competing on price, and quite frankly, I'm having a real
hard time seeing why anyone who understood the technology would go with a
more expensive printer if you want a Lightjet print. $68 for the first, and
$27 for consecutive was what I paid for the prints.
Their pricing is good. For example,
Reed Photo in Denver is $253 for a 30x40
($156 2nd, $124 6+); nancyscans is $192, $88 2nd).
I'll have to try nancyscans. But do they do
cibachrome? I've done tests on both
crystal archive and cibachrome, and the colors
are not nearly as good on the fuji (greens in
particular).
7. I love the latitude the high scan density of the drum scans give you.
I strongly agree!
Overall, I can't even fathom going back to dealing with traditional chemical
printers. The quality of all this digital stuff is wild.
I agree. I've put my darkroom on hold now for 4 years.
(I had one--then I moved--now I have a room painted black
and its dark--but I only use it for loading the 4x5 sheets.)
Roger Clark
http://www.users.uswest.net/~rnclark
(Home Page Photography)
Richard Knight
made (on an image setter) for platinum printing. The contrast/curves need
much tweaking for platinum (towards a flat negative) but it's a definite
advantage to hauling around an 8x10 camera. IMO, even a 4x5 enlarged to 8x10
lacks the contact print clarity. Although its certainly easier to dodge and
burn an enlargement, using a digital negative eliminates almost all of this
(if it's right).
Richard
"geod" <ge...@cwo.com> wrote in message news:B5FDF2BC.88C3%ge...@cwo.com...
RCoda
than an enlargement because you can't see what you trying to enhance as the
negative is below you, not above you. This way, you can do all that in
Photoshop ONCE.
And, it would be a drum scan of an already large 4x5 or 8x10 negative. So
quality is not an issue.
Anyway, I was talking about really big negatives, 16x20 and up. For 8x10 you
can get an LVT or FIRE (LightJet 2000).
Just an idea.
"geod" <ge...@cwo.com> wrote in message news:B5FDF2BC.88C3%ge...@cwo.com...
J. Lim
to give it a go.
Perhaps you could try to output your scans into a high resolution
digital lithographic film thru a imagesetter like the AFGA. Your film could be
output
in Positive or negative without any significant loss of resolution/sharpness at
all.
Lenght of your negative is almost unlimited, only the width of the negative
is limited to 52". You can get access thru a service bureau all over states
and I believe the cost is XXX lower than lightjet, further to that, there is no
question
at all comparing defination of resolution of film output by an imagesetter -
your lithographic
film quality will beat the quality of Light jet hands down. Theoreotically, how
digital
lithographic imagesetter works is similar to a lightjet, by means of laser
imaging,
the only difference is you get a real negative (Agfa or Fuji) or Positive.
The best part of imaging your scan to lithographic film, you will get a
oversize neg to
contact print to any print paper. Any good offset lithographic printer in your
area will be able to
put you in touch with a graphic service bureau with imagesetting facilities.
Cost for outputing to lithographic films may be had for only a fraction of what
it cost
to lightjet. Hope this help. If you need any further help or specific info,
email me.
James Lim
Keith Clark
That was great reading. Enlightening, too.
I'd love to see your stuff. How far are you from Moscow?
Keith
Charles Pezeshki wrote:
> > Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> > Date: Mon, 02 Oct 2000 11:23:25 -0700
> > Subject: Re: Lightjet Print report
> >
Sandor Mathe
snip...
> This technology puts the photographer, NOT THE LAB, in the driver's seat.
OK good, but I print my own (which I think is a very important part of calling
myself a photographer), so I already have this control.
> Want
> the barn redder- without affecting anything else? Not a problem. Impossible in a
> chem lab.
Not true at all. There is NO digital manipulation that can not be done using
traditional techniques. NONE. The computer just makes some things orders of
magnitudes easier. Whether that change in ease makes something feasible that
previously was not is certainly important.
For the red barn, simply make a red or cyan filtered mask (possibly hand
removing / bleaching other red things in the scene) of the appropriate
contrast and use that for some or all of the exposure on the print to enhance
the red in the barn. Obvoiusly pin registration careful work and several
hours are required, but certainly not impossible.
--
Sandor Mathe
san...@ca.ibm.com
Keith Clark
OK, I stand corrected. :>
Now, if I were to ask the average chem lab to do this for me, how much
would it cost, and would they get it right the first second or third
time?
Cheers,
Keith
Charles Pezeshki
For all those that are interested, my show is going to be hanging in the
Community Center Gallery on 3rd and Washington (one block E of Main)
8:00-5:00 for the next two months in lovely Moscow, Idaho.
I live in Moscow now, so if folks happen to be driving through, give me a
ring and I'll be happy to show anyone on the ng my transparencies, which
are, of course, the real fun thing!
FWIW, LF has really worked for me and my landscapes. The Clearwater Country
isn't necessarily a land of high peaks (though we have some)-- more forests
and rolling green hills. LF and its expanded range of tonality enable me to
really capture the essence of the Clearwater. I think it would be tough to
shoot this landscape with smaller formats. I've tried, and haven't been
successful.
If you're rolling through, Keith, give me a buzz!
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
--
Chuck Pezeshki
Associate Professor
School of Mechanical and Materials Engineering
WSU-Pullman
Pullman, WA 99164-2920
509-335-7662 (W)
208-883-3001 (H)
peze...@mme.wsu.edu
> From: Keith Clark <clarkpho...@hotmail.com>
> Organization: http://www.clarkphoto.com
> Reply-To: clarkpho...@hotmail.com
> Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> Date: Tue, 03 Oct 2000 04:32:28 GMT
> Subject: Re: Lightjet Print report
>
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> > Neither looks better than the other, as long as one is over 200 dpi on the
> > scan.
This is why the Epson sells for $300 and the Lightjet for $125,000,
because there is no color difference.
;-)
> --------------E3EE9E0C575CC9ABB45B80B4
> Content-Type: text/html; charset=us-ascii
> Content-Transfer-Encoding: 7bit
>
> <!doctype html public "-//w3c//dtd html 4.0 transitional//en">
> <html>...
PLEASE turn off HTML! If for no other reason, you limit your audience. I
just hit "delete" without reading when I see a bunch of HTML in my
text-based newsreader.
--
: Jan Steinman -- Jan AT Bytesmiths DOT com
: Bytesmiths -- digital artistry <http://www.bytesmiths.com/Art_Gallery>
: +1 503 635 3229
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<san...@enzyme.torolab.ibm.com> wrote:
> There is NO digital manipulation that can not be done using
> traditional techniques. NONE.
You haven't been through the KPT filters, or used Genuine Fractals, have
you? :-)
> The computer just makes some things orders of
> magnitudes easier. Whether that change in ease makes something feasible that
> previously was not is certainly important.
Okay. I guess you COULD trace each shape by hand, translate those shapes
into mathematical formulae, then scale the formulae, then somehow get that
back onto film and enlarge them with perfectly sharp edges somehow... :-)
Then there's the random dot stereogram filter. THAT one aught to be
interesting in the darkroom! :-)
Absolutist attitudes can get you in a lot of trouble. I know better than
to try to change the mind of someone who spells NO and NONE in
all-uppercase, but you're gonna get some flak over this one.
Keith Clark
"Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]" wrote:
>
> > > 1. There is no color difference between the Epson and the Lightjet. None.
> > > Neither looks better than the other, as long as one is over 200 dpi on the
> > > scan.
>
> This is why the Epson sells for $300 and the Lightjet for $125,000,
> because there is no color difference.
>
> ;-)
>
Apples and oranges, Jan.
Or rather ink and lasers. ;>
Keith
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> "Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]" wrote:
> >
> > > > 1. There is no color difference between the Epson and the
Lightjet. None.
> > > > Neither looks better than the other, as long as one is over 200
dpi on the
> > > > scan.
> >
> > This is why the Epson sells for $300 and the Lightjet for $125,000,
> > because there is no color difference.
>
> Apples and oranges, Jan.
I agree in general that these two machines are as incomparable as apples
and oranges, but when one makes rash statements like "no color
difference," one is selecting a thin enough slice of the apple and orange
as to claim they're indistinguishable.
When one furthers such rashness with "neither looks better than the
other," one is indeed inferring that an apple is an orange, and
vice-versa. I just wanted to point out that someone, somewhere, must think
there is at least 125,000/300 worth of difference between the two!
I'd be happy to prove (with a densiometer) that the two have very
different color gamuts. I'll even supply the wide-gamut Q-60 drum scan.
Who's going to volunteer the two prints? :-)
At this point, I'm not even claiming the Lightjet is "superior," but I
strongly dispute that "there is no color difference."
c._downs
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>> There is NO digital manipulation that can not be done using
>> traditional techniques. NONE.
>
>You haven't been through the KPT filters, or used Genuine Fractals, have
>you? :-)
>
>> The computer just makes some things orders of
>> magnitudes easier. Whether that change in ease makes something feasible that
>> previously was not is certainly important.
>
>Okay. I guess you COULD trace each shape by hand, translate those shapes
>into mathematical formulae, then scale the formulae, then somehow get that
>back onto film and enlarge them with perfectly sharp edges somehow... :-)
I've got nearly 30 years experience at masking images { silver
processes } as opposed to just three years Photoshop but would watch
absolutist attitudes in either medium. As far as masking to change
colors , remove or add contrast etc it is much of a lost art in
silver. What can be done? almost anything - how undetectable - past
the best efforts of the FBI or any experts I have ever known of.
Expense? Very high in both time and masking film. Easier in Digital -
usually but not always for someone with proper masking equipment and
knowledge. Actually as far as not being able to detect the changes if
you go to the uumpth- detail with masking you can probably out class
all but the largest files { over 650 mgbyts} with digital. . Proper
masking is very rarely done. For the finest work you have to use a
blowback system that produces the masks and negatives or
transparencies at twice the final size and then the - "back" part of
the Blow back system is used to make contact sized negatives from the
finished and masked images.
For fine art I would favor the silver masked methods - for commercial
{ or any price related work} the digital is easily the way to go.
Home darkroom silver masking isn't that difficult but very hard to
find anyone that knows how properly. Home done work can be inexpensive
but if carried to anywhere near the available limits will be an
expensive proposition requiring at least an enlarger twice the size of
the largest film image used. An 8x10 enlarger at least is needed for
medium format or 4x5 . You could possible purchase a used drum scanner
and top of the line computer station for the same price of equipment
that is need for silver masking and blowbacks.
If one looks at the finest examples of movie posters and other
advertizing media it can give an idea of what is possible with silver
masking and printing in a commercial venue. The same quality is rarely
seen in fine art imaging as it takes several days to make the first
image! I once worked on an image of french fries for nearly two weeks
to get the right look. :o) We changed the contrast on the fries to
make them sparkle yet not appear greasy, we changed the color of the
fries to make them more golden ...etc...etc....etc - but 10,000 copies
were made for one of the big fast food places. I would not rule out
anything for someone that knows the business.
In the real world where price and time are factors Digital is close
and may already be king. It will get better as time goes along ....But
in absolute terms I still would not rule out silver printing as able
to produce anything that digital can do. A look at the work of Jerry
Ulesman can give a good idea of what simple B&W techniques can produce
and the rich depth and tones that are still the providence of silver
prints. Just get out your wallet and talk to your banker! :o)
Jim Jonah
www.monacosys.com. The Monaco EZColor software along with a Monaco Sensor
(suction cup device that reads my monitor) costs less than $500. You'll get a
pre-printed "target" to scan on your flatbed scanner. Scanning this target and
a print from your printer allows EZcolor to produce profiles for your scanner
and printer at the same time.
You can also get a 35mm "target" to calibrate your slide scanner.
Little Red Riding Hood Lied.
Keith Clark
types.
For the rest of us, there are profiles for 1/10th that cost at
http://www.inkjetmall.com
Sandor Mathe
a 4x5. You copy it to 8x10, make all the masks and intermediate steps in 8x10
and then either final print from the 8x10 stack or make a final 4x5 from the
8x10 stack. Is that right?
I always new that working in the large and then reducing for the final
could get you good results, I never realized that this was a common
practice. I believe most comic strip artists draw much larger than the
2" square newspaper reproductions.
In rec.photo.equipment.large-format C. wrote:
> On Wed, 04 Oct 2000 00:30:02 GMT, J...@Bytesmiths.com.gov (Jan Steinman
> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>>> There is NO digital manipulation that can not be done using
>>> traditional techniques. NONE.
>>
>>You haven't been through the KPT filters, or used Genuine Fractals, have
>>you? :-)
>>
>>> The computer just makes some things orders of
>>> magnitudes easier. Whether that change in ease makes something feasible that
>>> previously was not is certainly important.
>>
>>Okay. I guess you COULD trace each shape by hand, translate those shapes
>>into mathematical formulae, then scale the formulae, then somehow get that
>>back onto film and enlarge them with perfectly sharp edges somehow... :-)
> I've got nearly 30 years experience at masking images { silver
--
Sandor Mathe
san...@ca.ibm.com
Charles Pezeshki
As I said earlier, I was offering my impressions, having the pieces hang
side-by-side on a wall. I think this information is useful for 90-95% of
folks on the newsgroup, that are in the semi-pro or amateur category like
myself, that are not doing product work where exact color reproduction is
required.
If you get out the loupe, and place it on the photo, you can see a
difference. As far as things that are measured by instrumentation, I am
sure that for a wide range of circumstances, one could find differences.
But for someone like myself, who is lucky if I can sell a framed 8x10 for
$100, all these things are primarily of intellectual interest. I'm trying
to make a decision based on some sort of balance between price and quality.
Considering that what most folks had available to us-- sending to a chemical
lab that may even make an internegative-- and then getting back alternately
a flat or contrasty print that looks butt-ugly-- these new processes are a
revelation.
Then the question comes down to which one to pick. If I had unlimited
funds, I'd get everything I own drum-scanned, then printed with the
Lightjet. But I can't. It's a hardship to get even a handful of
transparencies scanned. And if I can sub in a $6 or $3 print, instead of a
$20 one, that's huge for me. I'll bet it is huge for that same 90% on the
ng.
So, that's the point of my analysis. Amateurs and semi-pros like myself
have had no access to affordable color printing that did any justice to the
work. Now we do. It is imperfect, but so much better than stuff from the
past, it is a paradigm shift-- one that makes doing LF color possible.
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
> From: J...@Bytesmiths.com.gov (Jan Steinman -- jan AT bytesmiths DOT com
> [remove .gov])
> Organization: Bytesmiths
> Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> Date: Wed, 04 Oct 2000 18:32:57 GMT
> Subject: Re: Lightjet Print report
>
> I'd be happy to prove (with a densiometer) that the two have very
> different color gamuts. I'll even supply the wide-gamut Q-60 drum scan.
> Who's going to volunteer the two prints? :-)
>
> At this point, I'm not even claiming the Lightjet is "superior," but I
> strongly dispute that "there is no color difference."
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> On Wed, 04 Oct 2000 00:30:02 GMT, J...@Bytesmiths.com.gov (Jan Steinman
> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>
> >> There is NO digital manipulation that can not be done using
> >> traditional techniques. NONE.
> >
> >You haven't... used Genuine Fractals, have
> >you? :-)
> >
> >Okay. I guess you COULD trace each shape by hand, translate those shapes
> >into mathematical formulae, then scale the formulae, then somehow get that
> >back onto film and enlarge them with perfectly sharp edges somehow... :-)
>
> I've got nearly 30 years experience at masking images...
Charles, you misunderstand. I am not talking about masking here, I'm
disputing the statement that there is "NO digital manipulation that can
not be done using traditional techniques" by describing one such digital
technique that is used for enlarging prints to a degree not possible in
the darkroom.
I admit, Genuine Fractals "synthesizes" detail where there was none in the
original negative, but it IS a technique that cannot be done in the
darkroom.
I'm not making a qualitative judgement about digital vs darkroom; I'm not
even saying Genuine Fractals is "good." I'm just using it as an example of
the sort of thing that one who is deeply ingrained with traditional
techniques might not think of when using words like "NO" and "NONE."
It's a new world out there. There's always going to be room for
traditional techniques, but digital is revolutionary, and it DOES do new
things that were not even concievable using traditional photochemistry.
One might as easily say, "There is NO place one can't go in a car that
can't be gotten to in a horse and buggy. NONE." :-)
Again, I'm not attacking photochemistry -- it has its place, as do horses
and buggies! But while it's a poor craftsman who blames his tools, it's a
poorer one who can't tell them apart!
Keith Clark
At the risk of being off topic - ust looked at your re-designed web
site. The posters are wonderful. My favorite is "Reflected Trees".
Cheers,
Keith
> > From: J...@Bytesmiths.com.gov (Jan Steinman -- jan AT bytesmiths DOT com
> > [remove .gov])
> > Organization: Bytesmiths
> > Newsgroups: rec.photo.equipment.large-format,rec.photo.technique.nature
> > Date: Wed, 04 Oct 2000 18:32:57 GMT
> > Subject: Re: Lightjet Print report
> >
> > I'd be happy to prove (with a densiometer) that the two have very
> > different color gamuts. I'll even supply the wide-gamut Q-60 drum scan.
> > Who's going to volunteer the two prints? :-)
> >
> > At this point, I'm not even claiming the Lightjet is "superior," but I
> > strongly dispute that "there is no color difference."
Sandor Mathe
> In article <39dc80a7...@news.mindspring.com>, C. Downs wrote:
>> On Wed, 04 Oct 2000 00:30:02 GMT, J...@Bytesmiths.com.gov (Jan Steinman
>> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>>
>> >> There is NO digital manipulation that can not be done using
>> >> traditional techniques. NONE.
>> >
>> >You haven't... used Genuine Fractals, have
>> >you? :-)
>> >
>> >Okay. I guess you COULD trace each shape by hand, translate those shapes
>> >into mathematical formulae, then scale the formulae, then somehow get that
>> >back onto film and enlarge them with perfectly sharp edges somehow... :-)
>>
>> I've got nearly 30 years experience at masking images...
> Charles, you misunderstand. I am not talking about masking here, I'm
> disputing the statement that there is "NO digital manipulation that can
> not be done using traditional techniques" by describing one such digital
> technique that is used for enlarging prints to a degree not possible in
> the darkroom.
> I admit, Genuine Fractals "synthesizes" detail where there was none in the
> original negative, but it IS a technique that cannot be done in the
> darkroom.
If the digital process, filter, method or whatever is simulating a physical
process, then the original physical process can be used in a wet darkroom
for the same effect (obviously). If the digital process like fractals is
creating new information, then it is equivalent to hand touching up the
film at the level of individual pixels, so you can do that by hand with a fine
enough brush good enough magnification reduction, and a steady hand.
This is where the quantitative speed difference between doing it with a brush
and doing it digitally is so great that no one would even attempt the manual
method.
> I'm not making a qualitative judgement about digital vs darkroom; I'm not
> even saying Genuine Fractals is "good." I'm just using it as an example of
> the sort of thing that one who is deeply ingrained with traditional
> techniques might not think of when using words like "NO" and "NONE."
> It's a new world out there. There's always going to be room for
> traditional techniques, but digital is revolutionary, and it DOES do new
> things that were not even concievable using traditional photochemistry.
> One might as easily say, "There is NO place one can't go in a car that
> can't be gotten to in a horse and buggy. NONE." :-)
I can't agree with this. One can concieve of anything. Implementing it with
any kind of efficiency and chance of success is another issue.
--
Sandor Mathe
san...@ca.ibm.com
Roger N Clark
Roger Clark
c._downs
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>In article <39dc80a7...@news.mindspring.com>, C. Downs wrote:
>
>> On Wed, 04 Oct 2000 00:30:02 GMT, J...@Bytesmiths.com.gov (Jan Steinman
>> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>>
>> >> There is NO digital manipulation that can not be done using
>> >> traditional techniques. NONE.
>> >
>> >You haven't... used Genuine Fractals, have
>> >you? :-)
>> >
>> >Okay. I guess you COULD trace each shape by hand, translate those shapes
>> >into mathematical formulae, then scale the formulae, then somehow get that
>> >back onto film and enlarge them with perfectly sharp edges somehow... :-)
>>
>> I've got nearly 30 years experience at masking images...
>
>Charles, you misunderstand. I am not talking about masking here, I'm
>disputing the statement that there is "NO digital manipulation that can
>not be done using traditional techniques" by describing one such digital
>technique that is used for enlarging prints to a degree not possible in
>the darkroom.
>
>I admit, Genuine Fractals "synthesizes" detail where there was none in the
>original negative, but it IS a technique that cannot be done in the
>darkroom.
>
>I'm not making a qualitative judgement about digital vs darkroom; I'm not
>even saying Genuine Fractals is "good." I'm just using it as an example of
>the sort of thing that one who is deeply ingrained with traditional
>techniques might not think of when using words like "NO" and "NONE."
>
>It's a new world out there. There's always going to be room for
>traditional techniques, but digital is revolutionary, and it DOES do new
>things that were not even concievable using traditional photochemistry.
>One might as easily say, "There is NO place one can't go in a car that
>can't be gotten to in a horse and buggy. NONE." :-)
>
>Again, I'm not attacking photochemistry -- it has its place, as do horses
>and buggies! But while it's a poor craftsman who blames his tools, it's a
>poorer one who can't tell them apart!
Actually I'm very interested in Genuine Fractals! I have literally
thousands of 2 1/4 negs and slides that while great at 16x20 or
smaller will simply not hold up any larger. My niche in the photo
business is making larger prints so Genuine Fractals would make a huge
difference in my inventory. I shoot 90% 8x10 and split the difference
between 4x5 and 645. It is my understanding that Genuine Fractals
would really help the smaller format work.
I haven't read the facts on the new version of Photoshop but have
heard that something similar may be included...In any case I do see a
great potential for digital. My real concern for digital is the info I
keep getting at the trade shows that there is a need for images with
about the quality of 2 1/4 or less. One of the shows we attended kept
mentioning a file of about 84 mgbyts as the quality line they were
after. This to me in my limited digital knowledge will keep the file
size or resolution size for much of the machinery a little too small
for top quality Fine Art type work. I know that there is plenty of
work that is much better being done but I'm talking about what the
industry medium or standard will be based on. Something similar is
true with video as it just can't compete with film in most cases. NOT
that it couldn't but that the powers that be will not make money by
any greater improvement. There is a thread about Kodak dropping a lot
of top quality products for the much more profitable lower quality
point and shoot products. This is my fear about digital. There may be
a point at which digital products fill the greatest need niche and the
money for higher development will be slim.
c._downs
<san...@enzyme.torolab.ibm.com> wrote:
>Let's see if I understand this "blowback" technique. The original is say
>a 4x5. You copy it to 8x10, make all the masks and intermediate steps in 8x10
>and then either final print from the 8x10 stack or make a final 4x5 from the
>8x10 stack. Is that right?
>
>
>I always new that working in the large and then reducing for the final
>could get you good results, I never realized that this was a common
>practice. I believe most comic strip artists draw much larger than the
>2" square newspaper reproductions.
Yes,
Not really common but still done for best results when price wasn't a
real factor.
You often can make an "enlarged/improved" dupe of the original
image neg, tranny etc. from the masked image you have been working on
- at the larger size and work with it instead of the "stack" or neg
and mask sandwich...... as a clarification lets say the original is
4x5 and you choose to work at 8x10 . All the masks are 8x10 and the
final result is an 8x10.
This is left alone if you have the where with all to work with it and
is generally best to work with at that size . For many service
companies there is either a horizontal through the wall blow back
system or some type of vertical camera that can be fitted with a light
head - or an 8x10 enlarger. The enlarged image is then worked with.
If the image is to be sent back to a photographer that only has 4x5
enlarger then the 8x10 dupe image of the masked image is never made
and a 4x5 image { could be called an "improved/original size" dupe} is
made by the reverse of enlarging from the "stack" or mask and neg
sandwich. This is very high quality and will scan well in something
like a 4x5 home priced scanner. { something like a polaroid sprintscan
45} If you wanted to go digital and had a drum scanner the larger dupe
would be used.
Boy, this is easier to do than write about so sorry if it is
confusing. In practice it is very direct and simple.
A second way of working is by using the equipment to make a contact
sized image to work with- sometimes 40 x120inches or larger. This is
what I used to do many years ago and is best afforded to someone
needing the absolutely finest available at _whatever_ cost. This is
done by formula somewhat. It is usually done one of two ways. First
the blow back way as it was called by the folks I worked with. . I
would make an image of the same generation { positive to positive or
neg to neg. } at the largest size that the equipment would allow - or
at about 4 times the original images diagonal. { this size will
usually contain all of the information in the original and no further
enlargement is necessary..... usually no larger than 20x24 to 30 x40 -
the limit set by whatever equipment the company had that could be
fitted with a light head.{ I worked with a through the wall 30x40 that
had a light head that swung back and forth like a door.} The masking
and other work was done at this size and then this very large image
was enlarged again { or shrunk if necessary} to the size of the final
print. this was then contact printed on silver paper for the final
product instead of the previous way of working with an enlarger.
This can provide very high quality and will sometimes have the
appearance of using a camera to make the original that was at least
double the diagonal dimension of what was really used. Masking
negatives or slides gives so called edge contrast effects that will
make the image appear to have higher accutance when the image is
enlarged. If instead you mask something that is to be contacted the
edge effect is not as evident but the contact printed image will have
a look of sharpness that even when examined very close will appear
sharper and to have better tonality.
I had this explained to me and can only remember that it had
something to do with certain parts of the mask somehow canceling out
the original image as opposed to having the differences in contrast
"highlighted" when doing the final print by enlarging the masked
image..
From here on out I'm way over my head as this was only a three year
job {over 20 years ago}and my mind was on getting behind a camera and
not in 10 hr. days in the dark! :o) This large film processing and
printing left me with a metol allergy as we mixed and used at least
20-50 gallons of developer a day, all by hand. The subjects of my work
were overwhelmingly boring unless you like 8 foot long tennis shoes,
French fries, or wall murals of proposed apartment complexes!
I did get paid well and had the chance to learn at someone other than
my own expense though! :o)
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<san...@enzyme.torolab.ibm.com> wrote:
> In rec.photo.equipment.large-format Jan Steinman -- jan AT bytesmiths
DOT com [remove .gov] <J...@bytesmiths.com.gov> wrote:
> > In article <39dc80a7...@news.mindspring.com>, C. Downs wrote:
>
> >> On Wed, 04 Oct 2000 00:30:02 GMT, J...@Bytesmiths.com.gov (Jan Steinman
> >> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
> >>
> >> >> There is NO digital manipulation that can not be done using
> >> >> traditional techniques. NONE.
> >> >
> >> >You haven't... used Genuine Fractals, have
> >> >you? :-)
> >> >
> >> >Okay. I guess you COULD trace each shape by hand, translate those shapes
> >> >into mathematical formulae, then scale the formulae, then somehow get that
> >> >back onto film and enlarge them with perfectly sharp edges somehow... :-)
> >>
> >> I've got nearly 30 years experience at masking images...
>
> > I'm... describing one such digital
> > technique that is used for enlarging prints to a degree not possible in
> > the darkroom.
>
> ... If the digital process like fractals is
> creating new information, then it is equivalent to hand touching up the
> film at the level of individual pixels, so you can do that by hand with a fine
> enough brush good enough magnification reduction, and a steady hand.
Well, if we're going to be pedantic here, are you claiming that "hand
touching up" is the equivalent of resolution-independent scaling without
optical artifacts? If so, we'll have to agree to disagree.
Pedantic arguments may be intellectually stimulating, but are ultimately
unsatisfying, as each party retreats into their own imagination.
From a pedantic point-of-view, the original question is flawed. Some are
interpreting it to mean "obtain similar results to," which would cover the
hand-retouching argument.
If, however, one interprets it strictly: "can not be done," then I stand
by my argument that reproducing the digital manipulation of MANY digital
techniques -- Genuine Fractals being but an example -- "can not be done
using traditional techniques." QED.
> I can't agree with this. One can concieve of anything. Implementing it with
> any kind of efficiency and chance of success is another issue.
And, if one equates "conception" with "possible implementation," then of
course, the argument holds. But why stop at the darkroom? You can make the
same silly argument for grinding up wood and mining pigments and making a
depiction THAT way! Reductio ad absurdium.
Remember, the original poster said "done," which implies creating an
artifact, which implies a practical implementation. At least, that's the
way I read it.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> Actually I'm very interested in Genuine Fractals! I have literally
> thousands of 2 1/4 negs and slides that while great at 16x20 or
> smaller will simply not hold up any larger. My niche in the photo
> business is making larger prints so Genuine Fractals would make a huge
> difference in my inventory...
I've done some lovely 24"x36" prints using Genuine Fractals from 35mm!
It's not a panacea -- sometimes the results are ugly. It doesn't do well
on faces. The demo on the box is pure bunk -- they should put their
marketing department on the firing squad for such hype.
But in many cases, it does a stupendous job! Think of fractal-like
objects: trees, ferns, moss. These are what it does best at. My most
successful GF work has been on temperate rain forest scenes, with mossy
trees and ferns. Experienced photographers walk up to one of my better
24x36 images and ask what large format equipment I use! ("Linhof Super
Technika," I answer, "but not for this one, which is 35mm.")
> I know that there is plenty of
> work that is much better being done but I'm talking about what the
> industry medium or standard will be based on... NOT
> that it couldn't but that the powers that be will not make money by
> any greater improvement...
> This is my fear about digital. There may be
> a point at which digital products fill the greatest need niche and the
> money for higher development will be slim.
This is always the case, no? There will always be many times more
"consumer" models than "pro" models. In film, that hasn't stopped Leica or
Contax from producing top-notch 35mm stuff, nor has it eliminated large
format.
There are already high-end digitals -- well into five figures -- and the
price will eventually come down. I've been working with output of a
$28,000 Phase One 4x5 back that produces roughly 15,000 x 12,000 pixel
images -- these files are over 500MB, and barely fit on a CDR!
Such products are not going to "go away" because of the proliferation of
consumer models -- they'll just get more affordable!
jw&a
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> I've been working with output of a
> $28,000 Phase One 4x5 back that produces roughly 15,000 x 12,000 pixel
> images -- these files are over 500MB, and barely fit on a CDR!
if you listen really hard, you can probably hear me drooling.
;-)
bob rogers
south carolina
Chris Williams
Thanks for posting.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> <J...@Bytesmiths.com.gov> wrote in message news:Jan-
>
> > I've been working with output of a
> > $28,000 Phase One 4x5 back that produces roughly 15,000 x 12,000 pixel
> > images -- these files are over 500MB, and barely fit on a CDR!
>
> if you listen really hard, you can probably hear me drooling.
Here's something to drool over, Bob...
<http://www.bytesmiths.com/Art_Gallery/1937-2224-02-1024.jpg>
<http://www.bytesmiths.com/Art_Gallery/1937-2224-02-tightcrop-auto.jpg>
The first is a rather famous painting, shot with a Phase One at a rather
famous east coast university's art gallery. The canvas is roughly
110"x60", as I recall. It is, of course, scaled down -- you really don't
want to download the ~500MB original file!
The second is a tight crop of actual pixels from the Phase One.
BTW: the histogram of the original is quite interesting. There is a lot of
detail in the dark areas that doesn't show up on the screen. I'm having a
devil of a time trying to decide how to print a near-full size
reproduction for their gift shop. Do I bring out the detail, or do I keep
the gamma accurate and lose the detail? Oil and canvas has so much more
value range than an Iris, Lightjet, or whatever...
Cyrus Virdeh
I have been working with LightJets for over a year now and if done correctly, I
find them to be the best color print you can get. I do the digital preperation
for an online photo gallery at www.exhibitspace.com and have had the opportunity
to make prints for some of the worlds best photographers. So here are my
thoughts on the LightJet.
1. You have to have a good scan. Table top slide scanners simply will not do.
They just cant get the resolution. Granted, we print big (up to 24x36 for 35mm)
so we need all the resolution we can get, but the real reason to go drum is the
color density. We have tried most tabletops, the Imacon quasi drum, and others
but we use a Tango. If you establish a relationship with a scan house you can
always negotiate a little on price and get a deal. For 35mm work, the biggest
you can go is about 24x36. Bigger than that and it starts to get grainy. File
size is the important thing here. We get 60MB files from a 35mm slide and up to
250MB for 6x17 MF which will give you a print up to 2x6 feet!
2. Go mat. The Fuji gloss paper is so damm delicate you will put a welt in it
just looking at it wrong. The Mat paper also will not reflect like the gloss
which enables it to display in more situations.
3. Black and White. I have tried printing B&W several times on Crystal Archive
with pretty bad results each time. Color casts, flat looking prints, etc dog B&W
LightJet printing. A big problem also is the lack of grey pixel depth on a
LightJet. It is a 24 bit machine which means 8 bits per color. This is fine in
color work, but in B&W, since the R,G,B lasers are all the same intensity to
make a grey, it means only 8 bits grey scale. Thats only 256 shades of grey. Not
too good. I have seen several examples of Kodaks digital B&W paper at this years
Seybold conference and it looked pretty good. The problem is finding a LightJet
house that has B&W chemistery and has this paper. Most don't have either. I have
also heard of the contact print idea being exceuted successfully through our
print house (Calypso of Santa Clara, CA) but have not seen any examples. I am
not sure what they used as the transparent material. It may have been Kodak
Duratrans which can be run through a LJ.
4. It really does help to have a calibrated system. We work off of an Apple G4
with a 21" calibrated studio monitor. We have the profiles from the particular
scanner, the monitor, and our target LJ and everything is on screen exactly as
printed. After a while I have learned to trust this system and I am not even
paranoid anymore about color consistency. Pretty much all labs that have a LJ
will make the ICC profiles for it available as will most scanning houses make
the profile for their scanner available. Some LJ houses even make the CMYK
conversion table available so youcan preview a print on screen with Photoshops
CMYK Preview command. Applying an ICC profile is easy. When done editing a file,
use the Profile to Profile command under the Image menu to apply it. Just make
sure you are done editing the file as the on screen display is then altered from
ground truth. Beign able to edit and profile your own stuff will greatly reduce
cost as you will get the reprint price from the start. Photoshop 5 is a really
important investment for a 21st century photog.
5. Odds and ends. Fuji Crystal Archive used to have a problem with letting it
sit around after exposure but before developing. The reds would shift and go
down hill. We have not really had a problem with this but it has been a problem
with one of our photographers. I think Fuji was trying to fix this, but still
don't let them lay our stuff around between exposure and development.
Hope This Helps
Cyrus Virdeh
VP-Operations
exhibitSpace, Inc.
c._downs
wrote:
You mention the resolution as being a problem with the desk top
scanners. I have a 1200 dpi 42 bit/depth 8x10 scanner that I would
like to use with 8x10 original images. As far as resolution goes I
would think that I would have plenty of file size due to the large
format originals.{ for something like a 20x24 or 24x36 inch print} My
simple math would tell me that this would actually be better than a
2400 dpi scan off of a 4x5 image as I am not imaging any of the "film
artifact" { grain etc }. I'm limited in experience here but have been
told that somewhere above 1600 dpi you are really starting to scan,
what was described to me as, "film artifact" and that a larger
original scanned at below this dpi resolution size would have mostly
image information instead of film artifact in the final file. I was
told { by the scanner salesman- :o) } that in actual image clarity
{resolution of actual objects in the image}that the 1200 DPI 8x10
scans would actually be about equal to drum scans of 4x5 and surpass
the drum scans of 120 film....I have no reference to tell if this is
so but the 300 to 400 mgbyt scans sure look great on my small dye-sub
printer and sections of the image look great on a large monitor.
Any comment here?
The second part of the question I have is about Bit/depth and contrast
- The unit I have is 42 Bit depth and 3.6 contrast range I believe.
I know that the drum scanners have a greater range - What would be my
loss approximately if I scanned on my unit - corrected in photoshop -
and then sent the files to be output on a lightjet? {mostly Velvia and
E100 8x10 original transparencies} I have a service company that will
set up my system WYSIWYG with their color profiles if I do the
printing business with them.
For the time being I would let them do the lightjet prints and when
the next generation of Pigment printers comes out possibly purchasing
a 24 inch web model.
Any comments welcome!
Thanks,
Chuck
Roger N Clark
You mention the resolution as being a problem with the desk top
scanners. I have a 1200 dpi 42 bit/depth 8x10 scanner that I would
like to use with 8x10 original images. As far as resolution goes I
would think that I would have plenty of file size due to the large
format originals.{ for something like a 20x24 or 24x36 inch print} My
simple math would tell me that this would actually be better than a
2400 dpi scan off of a 4x5 image as I am not imaging any of the "film
artifact" { grain etc }. I'm limited in experience here but have been
told that somewhere above 1600 dpi you are really starting to scan,
what was described to me as, "film artifact" and that a larger
original scanned at below this dpi resolution size would have mostly
image information instead of film artifact in the final file. I was
told { by the scanner salesman- :o) } that in actual image clarity
{resolution of actual objects in the image}that the 1200 DPI 8x10
scans would actually be about equal to drum scans of 4x5 and surpass
the drum scans of 120 film....I have no reference to tell if this is
so but the 300 to 400 mgbyt scans sure look great on my small dye-sub
printer and sections of the image look great on a large monitor.
Any comment here?
is correct, then there would be no point in scanning at > 1200dpi
as you would only get film artifacts. This is simply not the
case. See http://www.users.qwest.net/~rnclark/scandetail.htm for my
examples (I do 4x5 at up to 3300 dpi).
Hwever, 1200 dpi on 8x10 will certainly be a nice image size and
may
be adequate for all but really large prints. But you could
get
mouch more detail out of your images. The guy who does my
drum
scans shoots 8x10 and scans 2 GB files! I think he would
go larger
but computers and removable storage media are limiting (he and
I
are both hoping for stable DVD writers--maybe within a year).
The second part of the question I have is about Bit/depth and contrast
- The unit I have is 42 Bit depth and 3.6 contrast range I believe.
I know that the drum scanners have a greater range - What would be my
loss approximately if I scanned on my unit - corrected in photoshop -
and then sent the files to be output on a lightjet? {mostly Velvia and
E100 8x10 original transparencies} I have a service company that will
set up my system WYSIWYG with their color profiles if I do the
printing business with them.
For the time being I would let them do the lightjet prints and when
the next generation of Pigment printers comes out possibly purchasing
a 24 inch web model.
Bit depth is important and too often everyone uses 8-bits/channel.
However, if you scan at more than 8-bits per channel, you must
conver to 8 for output because at present there are no systems
that go higher. There is definitely information in most images
at >8 bits/chan but it is very hard to print. This is the
same with analog traditional printing--that's why we dodge and
burn, etc. Even if there were printers >8b/chan, papers (reflected
media) have more limited dynamic range than transparency (positive
or negative) media, so we'll always be faced with compressing
information down to the output media.
Roger N Clark
(http://www.users.qwest.net/~rnclark
Home Page Photography)
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
In article <39DEB6AC...@earthlink.net>, Cyrus Virdeh
<cyr...@earthlink.net> wrote:
> 3. Black and White. I have tried printing B&W several times on Crystal Archive
> with pretty bad results each time...
Just curious, have you tried duotones or quadtones? It seems to me that
this might work for you, although picking the colors might be a pain.
> 4. It really does help to have a calibrated system.
Amen! I joined the Leben-Epson list for a while, but 80% of it was flames,
and another 15% was people struggling with color. Apple ColorSync Just
Works!
Keith Clark
>
> Hwever, 1200 dpi on 8x10 will certainly be a nice image size and may
> be adequate for all but really large prints. But you could get
> mouch more detail out of your images. The guy who does my drum
> scans shoots 8x10 and scans 2 GB files! I think he would go larger
> but computers and removable storage media are limiting (he and I
"DVD-Multi" promises to address cross-format compatibility. More info
here :
http://www.dvdonet.com/columns.jsp?article=65
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> You mention the resolution as being a problem with the desk top
> scanners. I have a 1200 dpi 42 bit/depth 8x10 scanner that I would
> like to use with 8x10 original images.
The problem with scanning a print is that you've already lost a lot of
density in printing it. The BEST prints have a Dmax no more than about
1.8, whereas MANY chromes can go over 3.8.
Also, unless this is at least a four-figure scanner, I wouldn't trust the
"1200 dpi" spec too much.
> My
> simple math would tell me that this would actually be better than a
> 2400 dpi scan off of a 4x5 image...
Once again, it's the density, NOT the pixels! The main reason for using a
drum is NOT higher resolution -- most of them fake resolution above about
4,000 spi anyway.
> [I've] been
> told that somewhere above 1600 dpi you are really starting to scan,
> what was described to me as, "film artifact"...
It depends entirely on the film. I find I have to scan at about 4,000 spi
to see the grain in Velvia, for example.
> I was
> told { by the scanner salesman- :o) }
Aha! Now that we know your source of information, your confusion is
forgivable! :-)
> that in actual image clarity
> {resolution of actual objects in the image}that the 1200 DPI 8x10
> scans would actually be about equal to drum scans of 4x5 and surpass
> the drum scans of 120 film...
Sure, that's why people are willing to spend $65,000 on a drum scanner,
when they could just go down to CompUSA and buy a supposed "1200 dpi"
scanner! :-)
That salesman has a promising career in politics...
> The second part of the question I have is about Bit/depth and contrast
> - The unit I have is 42 Bit depth and 3.6 contrast range I believe.
I'm a bit dubious. This sounds like you may have the Umax 3000? Not a bad
machine for a flat-bed, but if you look closely at their lit, you'll see
it's really a 36 bit scanner with the extra 4 bits synthesized somehow.
I'm dubious!
Paul Butzi
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>In article <39e00742...@news.mindspring.com>, C. Downs wrote:
>
>> You mention the resolution as being a problem with the desk top
>> scanners. I have a 1200 dpi 42 bit/depth 8x10 scanner that I would
>> like to use with 8x10 original images.
>
>The problem with scanning a print is that you've already lost a lot of
>density in printing it.
It seems to me that you're confusing the optical density range of the
original (in this case a transparency or a negative) with the density
range of some intermediate storage. The fact that a print has a
narrower density range than the transmission medium of the
transparency is of no import.
What matters is how the various optical densities of the original
are mapped into the digital domain. If the Dmin of the transparency
maps onto the minimum value in the digital representation, and the
Dmax maps onto the maximum values, with all intermediate values
of the original mapped onto intermediate values, then there is no
clipping (there might well be quantization).
>The BEST prints have a Dmax no more than about
>1.8, whereas MANY chromes can go over 3.8.
As I said, who cares? The print is an intermediate mapping.
Except to the extent that it introduces noise, and if the tonal
resolution of the print is much smaller than the original (or the
tonal resolution of the reflective scanner is lower than a
transmissive scanner) the actual dmax of the print is of
absolutely no significance at all.
-Paul
--
Newly updated and moved web site at:
http://www.butzi.net
c._downs
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>I'm a bit dubious. This sounds like you may have the Umax 3000? Not a bad
>machine for a flat-bed, but if you look closely at their lit, you'll see
>it's really a 36 bit scanner with the extra 4 bits synthesized somehow.
>I'm dubious!
Yes I have the Powerlook!
And I am scanning negs and transparencies instead of Prints.
If you would I would still like any comments on the usefulness of
this for 24x30 prints.
I have also trouble with the Bit depth figures and exactly what it
means to me after putting it through Photoshop5.5 on a G4. Many of the
features "gray" out and I've been also told that printer bit depth is
only 24 bit usually. Was also told that G Fractiles will only give the
equivalent of 24 bit depth. As I'm 99% a silver printer but interested
in the digital I have not really done much "homework" on these
subjects yet. Any thought on bitdepth and if it really makes any
difference if I have to often default downward anyway?
Is there a good newsgroup for beginners on digital?
Thanks Again
Chuck
Paul Butzi
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>In article <sonutsod1vm07kcfp...@4ax.com>, Paul Butzi
><bu...@halcyon.com> wrote:
>
>> On Sat, 07 Oct 2000 17:24:41 GMT, J...@Bytesmiths.com.gov (Jan Steinman
>> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>>
>> >In article <39e00742...@news.mindspring.com>, C. Downs wrote:
>> >
>> >> You mention the resolution as being a problem with the desk top
>> >> scanners. I have a 1200 dpi 42 bit/depth 8x10 scanner that I would
>> >> like to use with 8x10 original images.
>> >
>> >The problem with scanning a print is that you've already lost a lot of
>> >density in printing it.
>>
>> ...The fact that a print has a
>> narrower density range than the transmission medium of the
>> transparency is of no import.
>
>"Of no import?" But you, yourself even admit:
>
>> (there might well be quantization).
>
>and
>
>> Except to the extent that it introduces noise...
>
>Goethe noted that those who confuse the real with the ideal seldom go
>unpunished. Sure, "in theory" a properly made print "might approach" the
>information from a negative or slide, but those of us who do this daily
>would argue with the practicality of getting optimal results from scanning
>a print if the source neg or chrome is available.
You said "the problem with scanning a print is that you've already
lost a bit of density in printing it."
If you wanted to say 'lost a bit of infomation in printing it" why
didn't you say that?
If you mean "density", say "density". If you mean "information"
say "information". If you mean 'information" but say "density"
expect to get called on it.
Goethe had plenty to say about people who didn't say what they
meant, too.
>
>Analog processes are lossy. A xerox of a xerox is never as good as the original.
>
>> the actual dmax of the print is of
>> absolutely no significance at all.
>
>Sorry, I don't buy it. You lose information, even if only (in the
>theoretically perfect case) the "in between" values that get squished out
>in the lossy medium, then expanded back out during scanning -- the "comb
>histogram" effect that you properly refer to as "quantization."
Yes, but quantization error has to do with the tonal resolution of
the scan, and the possible introduction of noise in making the second
generation of the image. Density has nothing to do with it. Given
the propensity of relatively inexpensive transmission scanners to
have limited ability to read transparencies with high Dmax, it's
perfectly possible that using an inexpensive reflective scanner
to scan a well executed print *intended for scanning* will produce
better results than scanning the original with a similarly inexpensive
transmission scanner, since the inexpensive transmission scanner
will be unable to handle the density range of the original and will
clip, whereas the cheap reflection scanner can probably handle
the entire reflective density range of the print, or else the print
can be made soft enough that the scanner can do this.
Again, if you want to say "information", it's spelled "information",
not "density".
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> On Sat, 07 Oct 2000 17:24:41 GMT, J...@Bytesmiths.com.gov (Jan Steinman
> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>
> >In article <39e00742...@news.mindspring.com>, C. Downs wrote:
> >
> >> You mention the resolution as being a problem with the desk top
> >> scanners. I have a 1200 dpi 42 bit/depth 8x10 scanner that I would
> >> like to use with 8x10 original images.
> >
> >The problem with scanning a print is that you've already lost a lot of
> >density in printing it.
>
> ...The fact that a print has a
> narrower density range than the transmission medium of the
> transparency is of no import.
"Of no import?" But you, yourself even admit:
> (there might well be quantization).
and
> Except to the extent that it introduces noise...
Goethe noted that those who confuse the real with the ideal seldom go
unpunished. Sure, "in theory" a properly made print "might approach" the
information from a negative or slide, but those of us who do this daily
would argue with the practicality of getting optimal results from scanning
a print if the source neg or chrome is available.
Analog processes are lossy. A xerox of a xerox is never as good as the original.
> the actual dmax of the print is of
> absolutely no significance at all.
Sorry, I don't buy it. You lose information, even if only (in the
theoretically perfect case) the "in between" values that get squished out
in the lossy medium, then expanded back out during scanning -- the "comb
histogram" effect that you properly refer to as "quantization."
That was the perfect case. In reality, the scanning glass and lens, the
enlarging lens, the imperfections and resolving limits of the paper,
environmental factors, and the skill of the printmaker all conspire to
make the typical print of MUCH lower information content that its
corresponding source material.
Now these may well be "golden eyes" arguments that I make, since I make my
living making and selling large fine-art prints. The typical home computer
owner with a box of snapshots may well be perfectly happy with scanning
prints, but I believe the original poster's purpose (sorry, I clipped it)
was closer to fine-art than snapshots.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<bu...@halcyon.com> wrote:
> You said "the problem with scanning a print is that you've already
> lost a bit of density in printing it."
That isn't what I wrote. If you're going to paraphrase with your own
interpretation, you should not use quotation marks, which are correctly
used only for direct quotation.
> If you wanted to say 'lost a bit of infomation in printing it" why
> didn't you say that?
Oooh, we're getting testy, aren't we?
> If you mean "density", say "density".
I do mean density. And I'll be happy to prove it to you with my densiometer.
> If you mean "information"
> say "information". If you mean 'information" but say "density"
> expect to get called on it.
I believe those with moderate reading skills understood that I was making
both points, and was careful to separate them by using standard rhetorical
techniques, such as discussing them in separate paragraphs. I'm sorry for
you if you slept through English Composition 101, but that doesn't make me
the bad guy.
> Density has nothing to do with it... it's
> perfectly possible that using an inexpensive reflective scanner
> to scan a well executed print *intended for scanning*...
I see you can dish it out, but you can't take it. You claim "density has
nothing to do with it," then manage to add that you only mean this in a
"well executed print intended for scanning."
So, does density have "nothing to do with it" on ordinary prints, too, or
only on your special-purpose, pristine ones? If you make an absolutist
statement, then start attaching conditions to it, expect to get called on
it! :-)
> Again, if you want to say "information", it's spelled "information",
> not "density".
Thank you for the spelling lesson. Your unfriendly demeanor has earned you
a place right next to Foto Fool in my kill-file.
Like Oscar Wilde, I find it unsporting to be drawn into a battle of wits
with an unarmed person. :-)
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
> On Sat, 07 Oct 2000 17:24:41 GMT, J...@Bytesmiths.com.gov (Jan Steinman
> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>
> >I'm a bit dubious. This sounds like you may have the Umax 3000? Not a bad
> >machine for a flat-bed, but if you look closely at their lit, you'll see
> >it's really a 36 bit scanner with the extra 4 bits synthesized somehow.
>
> Yes I have the Powerlook!
> And I am scanning negs and transparencies instead of Prints.
> If you would I would still like any comments on the usefulness of
> this for 24x30 prints.
I researched this machine a bit, and just couldn't justify it. It was
about $4k, no? You can get a nice used drum for only a bit more!
I ended up getting an Optronics Falcon for $7,500. A "real" Dmax of 4.0,
and at least 4,064 "real" samples per inch. (It claims 5,419 spi
upgradable to 8,000, but the minimum sampling aperture is only 6.25
microns.)
> I have also trouble with the Bit depth figures and exactly what it
> means to me after putting it through Photoshop5.5 on a G4.
What greater dynamic range buys you is more flexibility in manipulation
--- especially in initial contrast, brightness, curves, etc.
> I've been also told that printer bit depth is
> only 24 bit...
Actually, the usable dynamic range of printers is quite a bit less than
that -- more like 12 or 13 bits!
> Was also told that G Fractiles will only give the
> equivalent of 24 bit depth.
Most digital manipulation only happens in 24 bits. The reason for more
bits in the scanner is to select the best 24!
If you work in Photoshop with a high-bit file (48 bits, but some are
generally unused), you'll notice that most things are grayed-out. Adjust
your curves to get the highlights and shadows where you want them, then
convert to 24 bit. THAT'S when you archive in GF for
resolution-independent re-sampling.
> Any thought on bitdepth and if it really makes any
> difference if I have to often default downward anyway?
If you need to suck detail out of shadows, or protect delicate highlights,
then do your curves in high-bit, THEN go down to 24.
If you have a perfectly exposed, flat-contrast image, there's no reason to
use high-bit files.
My empirical rule-of-thumb: a print has about 4-5 stops, a tranny has
about 6-7 stops, a normal digital file has 8, and a neg has about 9-11.
Scanning chromes (and prints), you're generally spreading out value, and
with negs you're generally compressing value. A high-bit file has the
capacity to express 16 stops of dynamic range, which no film can do, but
it can be handy to avoid cumulative errors, such as applying curves more
than once.
To summarize: high-bit files are useful for selecting or synthesizing the
"best" 24 bits to work with, but most other things you do -- especially
printing -- doesn't benefit a bit from high-bit data.
(Now I'm sure there will be lots of people frothing at the mouth over
this, but it works for me. Be polite, and I might listen to you when you
froth... :-)
> Is there a good newsgroup for beginners on digital?
This one ain't bad, once you figure out how to filter out the flamers,
trolls, and otherwise unfriendly people!
Roger N Clark
Excellent post! I agree on everything you said except one
point below:
"Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]" wrote:
> > Any thought on bitdepth and if it really makes any
> > difference if I have to often default downward anyway?
>
> If you need to suck detail out of shadows, or protect delicate highlights,
> then do your curves in high-bit, THEN go down to 24.
>
> If you have a perfectly exposed, flat-contrast image, there's no reason to
> use high-bit files.
A transparency may only have 7 stops of dynamic range, but it takes
more than 7 bits to capture the tonal gradations, much more. You certainly
want to have more than 1 bit per stop (with 8-bits/channel you would
get 1 1/7 bits (1.14 bits) per stop. That's not enough. 12 bits/channel only
approaches what is on the film. Unfortunately, most programs, even
photoshop is pretty limited in >8 bits/channel (is photoshop 6 any
better?). I've had many scans, including drum scans at 8-bits/channel
on well exposed slides that I couldn't bring out the shadow and highlight
details that I can see on the film.
Roger Clark
Roger N Clark
That's good news in a way, but my fear is that after the format war
settles out,
we would see no more multi-format drives. So again, if you write in the
format that loses, you can lose all your stuff. My message to the
manufacturers
if you're listening: settle your format war or I won't buy, period! You
promised DVD writers over a year ago, and we, the consumers, are still
waiting.
Roger Clark
Keith Clark wrote:
> Roger N Clark wrote:
>
> >
> > Hwever, 1200 dpi on 8x10 will certainly be a nice image size and may
> > be adequate for all but really large prints. But you could get
> > mouch more detail out of your images. The guy who does my drum
> > scans shoots 8x10 and scans 2 GB files! I think he would go larger
> > but computers and removable storage media are limiting (he and I
>
Paul Butzi
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>In article <taivtss1v19cled37...@4ax.com>, Paul Butzi
><bu...@halcyon.com> wrote:
>
>> You said "the problem with scanning a print is that you've already
>> lost a bit of density in printing it."
>
>That isn't what I wrote. If you're going to paraphrase with your own
>interpretation, you should not use quotation marks, which are correctly
>used only for direct quotation.
Ok, I stand corrected. The direct quote, suitable for use within
quotation marks, is:
"The problem with scanning a print is that you've already lost
a lot of density in printing it. The BEST prints have a Dmax no more
than about 1.8, whereas MANY chromes can go over 3.8."
I'm very sorry that I messed up and substituted
"a bit" for "a lot".
I confess, I don't see that this alters the argument to any degree.
>> If you mean "density", say "density".
>
>I do mean density. And I'll be happy to prove it to you with my densiometer.
I'd be very happy to learn how the density range is the issue, but I'm
still waiting.
Perhaps a concrete example of a reflective scan and a transmission
scan can advance the discussion.
Let's do a reflective scan first. We have a scanner which scans
a reflective medium and produces a digitized output. At each sample
point, the reflection density of the sampled point is represented in
the digitized output as an 8 bit value (e.g. a value between 0 and 255
inclusive). The width of the digitized output can be extended without
loss of generality in my argument.
To do this, the scanner will map the domain of reflection densities
within it's capabilities into that range of values (0 through 255
inclusive). However, there are two scaling values applied so
that you can pick a white point and a black point - two reflection
densities which will correspond to output values of 0 and 255.
In other words, you can set the scaling so that a reflection density
of 0.15 (or whatever the Dmin of your print is) corresponds to
0, and 2.0 (or whatever the Dmax of your print is)corresponds to 255.
Reflection densities below 0.15 would all map to 0, those above
2.0 map to 255. Note that these low and high points are set to avoid
clipping, *according to the object being scanned*. So, if you have
a print with a Dmin of .65 and a Dmax of 1.85, you'd set the scanner
so that .65 corresponds to 0 and 1.85 corresponds to 255. In any
case, reflection densities between the white point and black point
are assigned a value between 0 and 255, according to a formula
like outputvalue = 255 * ((sample - whitepoint) / (blackpoint -
whitepoint)), appropriately rounded to an integer.
This gives us the ability to produce scans which range in output value
from 0 to 255 when scanning objects with a broad range of minimum
and maximum reflection densities. It's possible to produce a full
range of outputs for any scanned object which doesn't exceed the
measurment limits of the scanner. Internal quantization error and
relative noise levels are an issue for scans of very short scale
objects but not for any object where the Dmin and Dmax approach
the capabilities of the scanner. The net result is that for any
reasonable print and any normal scanner, you get a full scale
ouput.
Transmission scanners, of course, do exactly the same thing. They,
also, will map (using exactly the same mathematics) the domain of
transmission densities onto the range of output values. In this case,
the transmission density of the object being scanned is longer, but
the range of output values is the same, with the whitepoint and
blackpoint values being set accordingly.
So I'm at a loss as to why you insist that density range is the
critical issue for scanning, and that a transmission scan is
superior to a reflection scan BECAUSE THE REFLECTION
DENSITY RANGE OF THE TRANSPARENCY IS GREATER
THAN THE TRANSMISSION DENSITY OF THE PRINT.
I'm perfectly willing to admit that transmission scans may be
superior for a host of reasons. I just don't believe that the reason
is that the density range of the transparency is greater than that
of the print.
>
>> If you mean "information"
>> say "information". If you mean 'information" but say "density"
>> expect to get called on it.
>
>I believe those with moderate reading skills understood that I was making
>both points, and was careful to separate them by using standard rhetorical
>techniques, such as discussing them in separate paragraphs. I'm sorry for
>you if you slept through English Composition 101, but that doesn't make me
>the bad guy.
Look, Jan - I'd be happy to learn from you how density is involved
here. But right now, I don't see it and you're not offering up much
evidence, just insults.
>
>So, does density have "nothing to do with it" on ordinary prints, too, or
>only on your special-purpose, pristine ones? If you make an absolutist
>statement, then start attaching conditions to it, expect to get called on
>it! :-)
As I said, if the density range of the an ordinary print is within the
capabilities of the scanner (very likely), and the print is not so
soft that the scanner cannot scale the scan properly and thus massive
quanitzation occurs, then the maximum density is of no import. Since
your average print has a Dmax of about 2.0, which is well within the
range of most inexpensive desktop scanners, the probability of
exceeding the range of the scanner is very small. Since virtually
all scanners can scale as they map into the digital domain, the
print would need to have a very short density range before
quantization occurs.
Contrast this to the world of transparency scanners, where it's
only the very high end scanners which can handle the density
range of dense transparencies (so clipping is a real problem)
and where you are still faced with the quantization problem if
the scan is not scaled properly
In any case, the problems with scanning prints have nothing
at all to do with the fact that the reflective Dmax of a print is
something like 2.0 or 2.1, whereas the transmission density
range of a color transparency can be far larger. Which is
exactly what I said in the beginning.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<rnc...@uswest.net> wrote:
> Excellent post!
(Blush :-)
> I agree on everything you said except one
> point below:
>
> "Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]" wrote:
>
> > If you have a perfectly exposed, flat-contrast image, there's no reason to
> > use high-bit files.
>
> A transparency may only have 7 stops of dynamic range, but it takes
> more than 7 bits to capture the tonal gradations, much more. You certainly
> want to have more than 1 bit per stop (with 8-bits/channel you would
> get 1 1/7 bits (1.14 bits) per stop.
Okay, I agree, with qualifications.
I'd say 1 bit per stop is theoretically exact for linear gamma (which is
actually logarithmic), with black & white images. After all, a stop is a
doubling of light, and a bit is a doubling of quantity.
But obviously, film is not always linear (actually, not perfectly
logarithmic), and computer monitors and printers are certainly not linear.
To further complicate things, the RGB color space is very unlike the human
perceptual color space (as best modeled by the HSV color space), which
itself is not perfectly logarithmic. Human vision is much stronger in dim
green light and bright red light, for example. Doubling an RGB value DOES
NOT double the perceived brightness -- RGB turns out to be a pretty awful
way to work on anything that humans must look at!
Furthermore, most CCD scanners use linear analog amplifiers -- "real"
linear, not even linear gamma, which is logarithmic. (Drum scanners
typically feed the PMT signal into a log amp.)
So I think the most fair thing to say is that it is almost always useful
to start with possibly too much data, rather than take the risk of
clipping some.
On the other hand, I tried to present my case as a "rule of thumb," and
given the vast number of complicating factors that make a direct
comparison impossible, I find "1 bit equals 1 stop" works fairly well in
practice. I always start with high-bit files for "problem" images, but
still find that many flat, properly exposed images don't lose much by
adjusting in the scanner for the best 8 bits out.
> I've had many scans, including drum scans at 8-bits/channel
> on well exposed slides that I couldn't bring out the shadow and highlight
> details that I can see on the film.
Sounds to me like operator error on the drum scanner! I'd think a proper
drum scan should compress the value range of the media into the 8-bit
range of the computer. Also, newer software allows drum scanners to
produce high-bit images -- perhaps you can ask your provider if they have
that capability.
Nicholas O. Lindan
transparency/negative" can be demonstrated by taking a print (unmounted,
natch) and viewing it by transmitted light.
Unless the print is exceedingly well made there will be a wealth of shadow
detail that shows up only in transmitted light.
Every time an image goes through a silver process the ends of the image get
thrown away in the toe and shoulder - especially true in the final print
where the full dynamic range of the paper is desired and so the image has
to overlap the toe and shoulder.
When the print is imaged this toe/shoulder data is lost in the scan - the
differences of gradation are less than 1 bit. A print made from the scan
of a print will show no extra shadow detail when viewed by transmitted light.
--
Nicholas O. Lindan, Cleveland, Ohio noli...@ix.netcom.com
Technical Management Consulting & Engineering Services:
New Product Development; Electrical Engineering;
Software, System and Circuit Design. Oh, & Photography
Nicholas O. Lindan
>
> A transparency may only have 7 stops of dynamic range, but it takes
> more than 7 bits to capture the tonal gradations, much more. You certainly
> want to have more than 1 bit per stop (with 8-bits/channel you would
> get 1 1/7 bits (1.14 bits) per stop.
It is a bit different than that.
If we take an 8 bit scan, and we adjust the high and low points to say
0.1 and 2.65 OD then each count in the resulting image will be 0.01 OD.
A stop is 0.3 OD. And a 2.55 OD range is 8 stops.
So there are 255/8 = 32 counts per stop. To count to 32 requires 5 bits -
and the result is there are 5 bits per stop.
What is happening is one is taking a double logarithm with the notation
of OD - a base 10 logarithm - and of 'bits' - a base 2 logarithm.
One can then adjust the curve of the scan to introduce further non linearity
and give more tonal resolution to midtones than to highlights and
shadows.
Peter De Smidt
or a print, but I do have a comment on the practical question. Doesn't
Lenswork make their prints by scanning a completed print, adjusting in
PhotoShop, outputing to a digital linescreen negative, and then contact
printing on Ilford FB paper? Their results are outstanding! Notice,
though, that they are scanning and printing at 100%. Perhaps if they were
enlarging the final print, you might not be able to get as good a scan from
a print, but I don't know. Basically, if you can get what you want, then
why worry about theoretical differences?
Regards,
Peter
c._downs
First a statement then a question:
Since I have only a mid-line quality scanner I've found that by making
a very good B&W print [at 100% final size] the scanner does a better
job of things. I've been loosing a bit at the top and bottom when
scanning the original neg but am able to make a very good silver print
first and make it slightly flat {maybe 1/2 to one paper grade} and the
scanner will pick it all up off the print. A slight tweak in photoshop
restores the contrast . As I have said before I'm not really
experienced in digital but it does seem that my scanner does a better
job this way. If I were to need anything larger I'm sure the scan off
the neg would be better.
Now the question - Is this normal - do I need some software that would
even out things? My thoughts are that I have the ability to make very
high quality prints and maybe the paper is closer to my scanners Dmin
and Dmax limits than the original neg is. Does this sound correct?
One more variable I'm not sure of is that the scanner will do much
better with a color transparency and a masked Ciba print is far less
necessary. Does this say anything about the scanner or the software it
uses.
{Standard Umax Powerlook 3000 and software}
Paul Butzi
<noli...@ix.netcom.com> wrote:
>I think some of the "you loose information when scanning a print Vs a
>transparency/negative" can be demonstrated by taking a print (unmounted,
>natch) and viewing it by transmitted light.
>
>Unless the print is exceedingly well made there will be a wealth of shadow
>detail that shows up only in transmitted light.
Sure. I agree. On the other hand, I've had good success with
actually digging much of that shadow detail out when scanning a print.
The problem, apparently, is not that the detail doesn't appear in the
print (as you pointed out) but that our eyes have difficulty picking
it out. You can see much of the same 'submerged' shadow detail
by viewing the print in very bright reflected light. Scanners
apparently are better at this than our eyes, although the 'auto'
mode for most scanner software seems to be such that it contrives to
throw away shadow detail willy-nilly.
>
>Every time an image goes through a silver process the ends of the image get
>thrown away in the toe and shoulder - especially true in the final print
>where the full dynamic range of the paper is desired and so the image has
>to overlap the toe and shoulder.
Yes, that's an issue, and a difficult one at that. But this is not an
issue of the difference in the transmission Dmax of the original
versus the reflection Dmax of the print - it's a matter of the
unfortunate mapping of the data in the original into the intermediate
form, where two values in the original are unfortunately mapped so
closely into the intermediate that they cannot be separated by
the scanner. This will happen regardless of the Dmax of the
print, although one technique to prevent this problem is to make
prints for scanning very soft and somewhat dark in the highlights, so
that the print for scanning avoids using the toe and shoulder of
the paper. Such a print is worthless for normal viewing, of course,
and if you make it too soft you run into the same problem in the
middle of the paper curve because the contrast is too low.
>
>When the print is imaged this toe/shoulder data is lost in the scan - the
>differences of gradation are less than 1 bit. A print made from the scan
>of a print will show no extra shadow detail when viewed by transmitted light.
Yep. The same is true if you adjust the scan from the original so
that the tonal rendering is identical to the way the print looks to
your eye, of course.
The interesting point from my point of view is that it's possible, in
the digital world, to make prints which have very good shadow
contrast, right down to Dmax. That's very hard to do in the gelatin
silver world.
Nicholas O. Lindan
>
> I have no comment on the theoretical differences between scanning a negative
> or a print, but I do have a comment on the practical question. Doesn't
> Lenswork make their prints by scanning a completed print, adjusting in
> PhotoShop, outputing to a digital linescreen negative, and then contact
> printing on Ilford FB paper? Their results are outstanding!
If you look at the original and the LJ print under very bright illumination
do you get the same level of detail in the dark shadows or does the LJ
print go flat?
To the prints look the same under transillumination?
Roger N Clark
> > I've had many scans, including drum scans at 8-bits/channel
> > on well exposed slides that I couldn't bring out the shadow and highlight
> > details that I can see on the film.
>
> Sounds to me like operator error on the drum scanner! I'd think a proper
> drum scan should compress the value range of the media into the 8-bit
> range of the computer. Also, newer software allows drum scanners to
> produce high-bit images -- perhaps you can ask your provider if they have
> that capability.
I disagree. If you have a transparency with a DMAX of 3, that's a range
of 1:1,000. You digitize that into 256 levels of 8-bits/channel, and you
do not have enough detail between each level--meaning film has more
detail than can be scanned at 8-bits/pixel. The operator can be creative,
but you can't get 1000 intensitiy levels into 256 levels--you must
compromise somewhere. Usually operators compromise
in the highlights or shadows or the main image will look
really bad. And a compromise must be made if 8 bits/channel.
Ever notice the spectacular tonality in Ansel Adams clouds?
You need more than 8-bits to digitize that kind of range and
tonality.
Roger Clark
Nicholas O. Lindan
>
> The interesting point from my point of view is that it's possible, in
> the digital world, to make prints which have very good shadow
> contrast, right down to Dmax. That's very hard to do in the gelatin
> silver world.
Ain't that the truth!
Peter De Smidt
Nicholas O. Lindan <noli...@ix.netcom.com> wrote in message
news:39E13311...@ix.netcom.com...
> Peter De Smidt wrote:
> >
> > I have no comment on the theoretical differences between scanning a
negative
> > or a print, but I do have a comment on the practical question. Doesn't
> > Lenswork make their prints by scanning a completed print, adjusting in
> > PhotoShop, outputing to a digital linescreen negative, and then contact
> > printing on Ilford FB paper? Their results are outstanding!
>
> If you look at the original and the LJ print under very bright
illumination
> do you get the same level of detail in the dark shadows or does the LJ
> print go flat?
>
> To the prints look the same under transillumination?
>
LJ print? You mean the lenswork prints? I don't notice any loss of detail
in the shadows, but then I don't have one of the 'original' prints to
compare it to. If I could afford the 'originals', I'd certainly buy them.
The artist (David Best) does claim that the Lenswork print matches his
original print.
Or perhaps you're talking about the George DeWolfe's ink jet prints? The
shadows don't look as rich as my prints on FB glossy, but that might be due
to the fact that the inkjet prints are on matte paper. I don't have any
glossy inkjet prints, and so I can't make a fair judgment.
Um...why would it matter if the prints look different with transmitted
light? I never look at prints that way.
Sorry that I couldn't be of more help.
Peter
Nicholas O. Lindan
>
> LJ print? You mean the lenswork prints?
Ooops - LJ for Lightjet
> Um...why would it matter if the prints look different with transmitted
> light? I never look at prints that way.
Try it, the difference can be pretty amazing.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<rnc...@uswest.net> wrote:
> "Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]" wrote:
>
> > > I've had many scans, including drum scans at 8-bits/channel
> > > on well exposed slides that I couldn't bring out the shadow and highlight
> > > details that I can see on the film.
> >
> > Sounds to me like operator error on the drum scanner! I'd think a proper
> > drum scan should compress the value range of the media into the 8-bit
> > range of the computer...
>
> I disagree. If you have a transparency with a DMAX of 3, that's a range
> of 1:1,000. You digitize that into 256 levels of 8-bits/channel...
Whoa, there! A value range of 1-1000 does not neatly map into 1-1000
ranges (10 bits) for each color!
The RGB space is non-monotonic -- you can get the same perceived color,
let alone value or luminosity, for different combinations of red, green,
and blue. Likewise, you can get many more graduations in perceived
luminosity than you would by changing RGB values in lock-step.
> you
> do not have enough detail between each level--meaning film has more
> detail than can be scanned at 8-bits/pixel. The operator can be creative,
> but you can't get 1000 intensitiy levels into 256 levels...
But you have more than 256 levels at your disposal! Otherwise, there'd be
no reason to do duotones, quadtones, etc. (I've even been toying with
CMYKOG hextones on my Roland...)
You can get MUCH more than 256 levels of gray out of a 24 bit color space!
> Ever notice the spectacular tonality in Ansel Adams clouds?
> You need more than 8-bits to digitize that kind of range and
> tonality.
Actually, the human eye can only discriminate about 100 levels of pure
gray -- another reason multi-toned B&W printing is useful. (I don't know
how to reconcile this with your perception of Ansel's prints. I think I
read this in "Light and Perception" by George Murray? Or some other
required reading at Tektronix Laboratories when I worked there...)
I think we agree on more than we disagree, Roger. I think the safest thing
to say is that comparison between photochemistry and digital is neither
easy nor straightforward!
> > Also, newer software allows drum scanners to
> > produce high-bit images -- perhaps you can ask your provider if they have
> > that capability.
Have you checked this out? ColorRight Pro 2.0 (for Optronics scanners) can
send high-bit files -- real 16-bits per color, not zero-filled 10 or 12
bits -- to the computer.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<bu...@halcyon.com> wrote:
> So I'm at a loss as to why you insist that density range is the
> critical issue for scanning, and that a transmission scan is
> superior to a reflection scan BECAUSE THE REFLECTION
> DENSITY RANGE OF THE TRANSPARENCY IS GREATER
> THAN THE TRANSMISSION DENSITY OF THE PRINT.
You're forgetting the gray (actually, non-gray :-) area between 0 and1.
With a medium of lower absolute density range, the least expressible and
most expressible densities are lost. You might answer "that's information,
not density," to which we'll just have to agree to disagree.
The "quantization effect" you dismiss as an information loss, rather than
a density loss, also affects the beginning and end of the total range.
Mere semantics? Perhaps.
> But right now, I don't see it and you're not offering up much
> evidence, just insults.
But ma, Paul started it! :-)
When you only look at things one way, it's not difficult to not see the
other guy's argument.
> >So, does density have "nothing to do with it" on ordinary prints, too, or
> >only on your special-purpose, pristine ones? If you make an absolutist
> >statement, then start attaching conditions to it, expect to get called on
> >it! :-)
>
> Since
> your average print has a Dmax of about 2.0, which is well within the
> range of most inexpensive desktop scanners...
>
> Contrast this to the world of transparency scanners, where it's
> only the very high end scanners which can handle the density
> range of dense transparencies...
Okay, I get it now -- you're comparing apples with oranges. Fine. We
understand each other. You claim that "density has nothing to do with it"
for a certain set of constraints. Okay. I agree. But I do insist that
density can have everything to do with it for a different set of
constraints.
Getting back to what I believe was the original issue: I agree that
Charlie may be better off scanning his hand-crafted prints on his Umax
3000 than by using its transparency adaptor on his negatives. But I'm not
willing to admit he'll get better results scanning prints than by using a
film scanner of similar price (~$4,000).
Are we all happy now? :-)
Paul Butzi
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>> So I'm at a loss as to why you insist that density range is the
>> critical issue for scanning, and that a transmission scan is
>> superior to a reflection scan BECAUSE THE REFLECTION
>> DENSITY RANGE OF THE TRANSPARENCY IS GREATER
>> THAN THE TRANSMISSION DENSITY OF THE PRINT.
>
>You're forgetting the gray (actually, non-gray :-) area between 0 and1.
No, I don't think so. If you could be more specific about what region
you are discussing it would make your argument easier for me to
follow. Perhaps you could describe this as a specific range of
densities in either the original negative/transparency or else a
specific range of densities in the print.
>
>With a medium of lower absolute density range, the least expressible and
>most expressible densities are lost.
I don't understand what you mean by 'least expressible' and 'most
expressible' so I don't understand your argument. Can you give
an example of a 'least expressible' density, and demonstrate how
it is lost in a medium with a lower absolute density range?
Perhaps you could explain it in the context of the example which
I gave, which you have carefully deleted from your response.
>The "quantization effect" you dismiss as an information loss, rather than
>a density loss,
I don't dismiss quantization at all; it's a very important problem.
But since it is unrelated to any 'density loss', I prefer to use the
correct term and call it information loss. After all, it's
information that's being lost. The reason I don't call it 'density
loss' is because the quantization is unrelated to the fact that
the density range of the print is smaller than the density range of
the transparency or negative. (although actually, if it's a normal
B&W negative, that might not be true).
> also affects the beginning and end of the total range.
>Mere semantics? Perhaps.
If you're discussing the compression of values due to the shoulder and
toe of the print medium, then it would be information loss, and it
would occur regardless of the absolute range of reflective densities
of the print, so yes, I'd call it information loss that had nothing to
do with the density range of the print.
>> >So, does density have "nothing to do with it" on ordinary prints, too, or
>> >only on your special-purpose, pristine ones? If you make an absolutist
>> >statement, then start attaching conditions to it, expect to get called on
>> >it! :-)
>>
>> Since
>> your average print has a Dmax of about 2.0, which is well within the
>> range of most inexpensive desktop scanners...
>>
>> Contrast this to the world of transparency scanners, where it's
>> only the very high end scanners which can handle the density
>> range of dense transparencies...
>
>Okay, I get it now -- you're comparing apples with oranges.
No, I'm addressing the original subject of discussion, which is that
you stated that the reason scanning prints would produce inferior
results was that the reflective density range of a print was smaller
than the transmission density range of a transparency.
If you want to insist that discussing this in the context of prints
being scanned on reflection scanners and transparencies
scanned on transmisison scanners is tantamount to discussing
apples and oranges, then I guess I don't understand the rules
you use for rational discourse.
> Fine. We
>understand each other. You claim that "density has nothing to do with it"
>for a certain set of constraints. Okay. I agree. But I do insist that
>density can have everything to do with it for a different set of
>constraints.
Well, if you ever give some indication of a situation where density
has everything to do with it, perhaps I'll comment on that.
>Getting back to what I believe was the original issue: I agree that
>Charlie may be better off scanning his hand-crafted prints on his Umax
>3000 than by using its transparency adaptor on his negatives. But I'm not
>willing to admit he'll get better results scanning prints than by using a
>film scanner of similar price (~$4,000).
The original issue was that you stated that the problem with scanning
prints (rather than the original transparency or negative) was that
the density range was smaller. I certainly agree that there are a
host of problems with scanning prints rather than original negatives
or transparencies. I just disagree that the smaller density range of
a print is among those problems.
Chuck is trying to scan 8"x10" transparencies and monochrome
negatives.
If you know of a $4000 dollar alternative to the flatbed transparency
scanners he's considering (like the Umax 3000) I suspect that
a considerable number of people in the large format newsgroup
(including me) would love to hear about it.
Paul Butzi
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>
>Quantization effects density at its end-points.
>
>Lets make an unreal example for illustration. These numbers are all off by
>varying amounts, and I'm not going to argue specifics.
>
>You have a computer capable of working with 1000 different values.
>You have a scanner capable of quantizing to any of 10,000 values.
>You have a print that occupies 100 different values, in the range 4,950 to
>5,050.
>It's original negative occupies 1000 different values, in the range 4,500
>to 5,500.
I'd argue that I have a negative with information encoded in the
density range from, say, 0.2 logD to perhaps 2.1 logD transmission
densities. Color transparencies have a higher Dmax.
I also have a print, with information encoded in the density range
from, say, 0.15 LogD to about 2.0 logD reflection densities.
I don't see any reason why the number of 'values' in the print
will be smaller than the number of 'values' contained in the
negative.
Those are fairly representative for modern B&W films with normal
development intended for printing on gelatin silver paper of normal
contrast, and a print on gelatin silver paper from that negative.
>
>When you scan the negative, the scanner software allows you to select a
>density range from which the scanner will extract samples and map them to
>your computer's value range. The scanner sees 1000 different values, and
>sends them to the computer, 1:1.
Ok.
>
>When you scan the print, again, the scanner software allows you to select
>a density range from which the scanner will extract samples and map them
>to your computer's value range. The scanner sees 100 different values,
>scales them to the computer's value range, and sends them to the computer,
>1:10.
No, actually, the scanner will see 1000 values, scale them to the
computers range (in your parlance) and send them to the computer.
Modulo the effect of the H&D curve of the paper not matching the
curve used by the transmission scanner, the effect should be pretty
similar to scanning the negative.
>
>If you look at the histogram of values on the computer for the print, you
>will see the familiar comb-shaped curve, with data clustered at 10, 20,
>30... 970, 980, 990.
No, you'll see a curve that looks much like the curve that you get
when you scan the negative and invert.
At least, that's the way it's looked with every print *I've* scanned.
I seem to have very little trouble making, say, a 16 bit per pel scan
from a print and ending up with a histogram with does not display
the 'comb' quantization artifact that you claim will appear.
> You have lost the density values between 0 and 10 and
>between 990 and 1000! Now you could offset the samples either way, but
>you've still lost 1/2 a digit of density on either end by scanning the low
>density value media rather than the high density value media. This isn't
>rocket science; it's Shannon & Nyquist 101!
Actually, it's neither Shannon nor Nyquist. It's the simple
mathematics of a function mapping a domain of integers onto another
range of integers. The problem with your example is that
you assume that the domain is smaller than the range.
There are a lot more grey values in a print than 100.
In fact, if you are scanning any reasonable print
and getting an 8 bit per pel result, the domain will be larger
than the range - that is, you can easily pick more than
256 grey values off the print, so you'll have to throw data
away when you do the mapping.
>You simply cannot chop up a print's density range into as many pieces as
>you can a negative's.
Why? The negative encodes the information as a bunch of silver
particles suspended in gelatin. The print encodes the information
in the same way. I don't see any reason why the tonal resolution of
a print should be worse than that of a negative. And there's no
particular reason why a reflective scanner should be able to
extract only 100 values from a print but a transmission scanner
can extra 1000 from a negative. None.
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<bu...@halcyon.com> wrote:
> On Mon, 09 Oct 2000 19:19:55 GMT, J...@Bytesmiths.com.gov (Jan Steinman
> -- jan AT bytesmiths DOT com [remove .gov]) wrote:
>
> >With a medium of lower absolute density range, the least expressible and
> >most expressible densities are lost.
>
> I don't understand what you mean by 'least expressible' and 'most
> expressible' so I don't understand your argument.
...
> >The "quantization effect" you dismiss as an information loss, rather than
> >a density loss,
>
> I don't dismiss quantization at all; it's a very important problem.
> But since it is unrelated to any 'density loss'...
Quantization effects density at its end-points.
Lets make an unreal example for illustration. These numbers are all off by
varying amounts, and I'm not going to argue specifics.
You have a computer capable of working with 1000 different values.
You have a scanner capable of quantizing to any of 10,000 values.
You have a print that occupies 100 different values, in the range 4,950 to
5,050.
It's original negative occupies 1000 different values, in the range 4,500
to 5,500.
When you scan the negative, the scanner software allows you to select a
density range from which the scanner will extract samples and map them to
your computer's value range. The scanner sees 1000 different values, and
sends them to the computer, 1:1.
When you scan the print, again, the scanner software allows you to select
a density range from which the scanner will extract samples and map them
to your computer's value range. The scanner sees 100 different values,
scales them to the computer's value range, and sends them to the computer,
1:10.
If you look at the histogram of values on the computer for the print, you
will see the familiar comb-shaped curve, with data clustered at 10, 20,
30... 970, 980, 990. You have lost the density values between 0 and 10 and
between 990 and 1000! Now you could offset the samples either way, but
you've still lost 1/2 a digit of density on either end by scanning the low
density value media rather than the high density value media. This isn't
rocket science; it's Shannon & Nyquist 101!
If you don't like unreal examples, substitute your favorite numbers.
You simply cannot chop up a print's density range into as many pieces as
you can a negative's. This is the "quantization problem," BUT because of
integer math, it is ultimately a total density limiting issue, as well!
Remember, it's a log scale, and a half a digit at each end-point is equal
to 0.1 difference in total density Dmin-max!
> If you know of a $4000 dollar alternative to the flatbed transparency
> scanners he's considering (like the Umax 3000) I suspect that
> a considerable number of people in the large format newsgroup
> (including me) would love to hear about it.
There are used drum scanners on the market at or near that price, and a
whole lot more within double that price. Granted, it's not as simple as
calling up your local dealer and whipping out your credit card, but the
results are MUCH better than any CCD-based machine -- even the "near drum"
Imacon, which only does 1,800 spi over 4"x5".
I'm not going to do your homework for you. It took me a year to locate and
purchase the used drum I wanted. You can start at
<http://www.printusa.com/equip/prepress.htm>, educate yourself about
various models, do web searches. I also recommend used dealer/broker Jeff
Weber <mailto:jeff...@jeffweber.net>, where I got my ColorGetter Falcon.
He offers a 90 warranty for those wary of used machines.
In fact, I see Jeff listed a Scanmate 5000 at PrintUSA on October 3rd for
$4,000, which is $2,000 less than I was considering paying for one. This
unit will do 5,000 spi over about 10"x12", and is exceptionally small and
light for a drum, but a couple people I corresponded with when I was
considering it mentioned expensive maintenance problems. But Jeff's priced
it to move -- it won't sit around long at $4k...
(And no, I have no connection with Weber Systems, other than as a
satisfied customer.)
Charles Pezeshki
I understand your point mathematically.
How 'bout pragmatically? A greater range of light intensity would make it
easier to design a receptor that would measure that range more exactly. To
me, it would be similar to boosting the signal/noise ratio. That's why I
think it would be easier.
But I'm no expert on this stuff.
Chuck
Chuck Pezeshki
http://users.moscow.com/pezeshki
Paul Butzi
<peze...@moscow.com> wrote:
>Hi Paul,
>
>I understand your point mathematically.
>
>How 'bout pragmatically? A greater range of light intensity would make it
>easier to design a receptor that would measure that range more exactly. To
>me, it would be similar to boosting the signal/noise ratio. That's why I
>think it would be easier.
Sure. A wider range *would* seem to imply that with the same absolute
resolution you'd get better *relative* resolution because of the
larger domain. The problem is that as you increase the domain over
which you will measure, it becomes more and more difficult to
build a sensor which can cover that entire domain and maintain the
same absolute resolution. That's why you find fever thermometers
which are accurate to .1F over a range of perhaps 95F-110F but you'll
have to search for a long time to find a thermometer that will be
accurate to .1F over -100F to +1500F.
It's harder to build a scanner with increased resolution over the
smaller range, of course, as you point out. Noise will take its toll
and eventually you can't improve resolution much. So there's a
tradeoff going on, clearly. One the one hand you have the
difficulty of increasing resolution. On the other hand, you have the
difficulty of maintaining resolution as you increase the size of the
domain. Both of these are clearly not linear curves of difficulty
versus payback, so I'd expect that there's some range in the
middle where you're close to the optimum, and to improve
results you are inclined to improve both at the same time.
I'd argue that current scanners are probably in that optimized
range, since scanners seem to be going from 36 bit to 48 bit
color as they seek to drive the Dmax over 4.0 or so. Part of
this is also driven by software support (there's no point in
offering 96 bit color since there's no software to manipulate a
96 bpp color image) and part is probably driven by fundamental
limits of the CCD sensors used in most scanners.
We might infer something about the relative difficulties of these two
tasks by examining the costs and relative availability of solutions to
both problems. I'd note that cheap scanners seemed to make
an early jump from 8 bit monochrome/24 bit color to 12 or 16 bit
monochrome, 36 or 48 bit color. Then there was a lag while
the Dmax of cheap scanners was improved.
I don't seem to hear a lot of clamor for scanners which do more than
48 bit color/16 bit greyscale, which suggests that the tonal
resolution problem has been solved adequately. I do still hear
clamoring for scanners which have the capability to scan objects
with higher Dmax, which suggests that even the current 3.6-4-2 logD
capability might be limiting the quality of scans from transparencies.
So it might appear that the Dmax problem is harder/more costly to
solve than that tonal resolution over a limited domain problem.
I suspect that in fact, the 48 bit tonal resolution and the 3.6 LogD
Dmax are probably the ends of the range of utility and now the
effort is mostly a matter of cost-effectiveness rather than capability
improvement.
I would say that the issue of whether the Dmax of a print is about 2.1
logD and that of a transparency is about (I'm in the dark, here) 4.0
logD means it's better to scan a print or a transparency boils down to
two questions:
1. Can current reflective scanners scan a domain of 0.3 logD to
2.2 logD with sufficient absolute tonal resolution to
provide a full relatively noiseless sample for each pixel
(at whatever pel depth you think is sufficient)?
2. Can current transparency scanners scan a domain of .3 to 4.2 logD
(or whatever you think the useful scale of a transparency is)
with sufficient resolution to provide a full depth noiseless
sample for each pixel (again, at whatever pixel depth
you think is sufficient)?
If you think that the answer to (1) is "yes", as I do, then there's no
handicap to scanning a print THAT IS CAUSED BY THE SMALLER
RANGE OF DENSITIES.
If you think that the answer to (2) is "yes", and I am not sure that
everyone does (I am uninterested, since I don't work in color)
then you would conclude that there is no handicap to scanning
transparencies that is caused by the increased range of densities.
If your print were REALLY soft, of course, the tonal values would be
so compressed that you'd find it increasingly difficult to separate
them. But I've, as an experiment, scanned some quite faded prints
that had aged badly (probably underfixed) and didn't seem to run
into the 'comb histogram' quantization problem or a very noisy scan,
so I'd say that it's not the limiting problem. For normal prints
with a Dmin of say .3 and a Dmax of 2.1 or so, it doesn't seem to be
an issue at all. If you were at the limit of tonal resolution where
the signal to noise ratio was a problem, you'd expect to see
noise in regions of smooth tonality when scanning prints, and
I don't see that in the flatbed 16bit monochrome scans of
gelatin silver prints that I've done, and I have actually looked.
If you want to know what *I* think the reason why scanning prints
is likely to be worse than scanning original transparencies/negatives,
I'd say that often to get a good looking print, the desire to get a
snappy looking print means that people print so that shadow detail
is lost in the shoulder of the paper. The result is that there's
shadow detail in the negative/tranny that isn't there at all in the
print. This sort of limits your options for printing from the scan
from the print. I suspect this is a real difficulty if you make
prints from transparencies, especially if silver masking is not
used to control the contrast and fit the entire range of the tranny
onto the print. That's why you see so many Cibachromes/Ilfochromes
which have huge black suppurating holes where you'd expect
luminous shadow detail.
The other problem I see is that when you make a print from
a transparency, you get all sorts of weird color crossover
problems (and I'm thinking of the ilfochrome cyan highlight
thingie here). I suppose you can sort those out after scanning
but it would seem simpler to just scan the original. There are
probably other issues involving the spectral characteristics
of the dyes used in the transparency and print material and
whether they match the spectral sensitivity of the color sensors
in the scanner but I'm a monochrome guy and honestly don't
know much about it beyond it being something *I'd* worry about
if I did color work.
But neither of those is a is caused by the fact that the Dmax
of the print is smaller than the Dmax of a color transparency,
which is what all the discussion is about.
Bottom line: if I were to work in color and with color transparency
film, I'd sure as hell scan the transparencies rather than make
prints and scan the prints. But I wouldn't claim that the reason was
that print has a smaller range of densities than that tranny.
And that about taps out both my interest in the subject and
the time I have available to discuss it. If there's more discussion
on the subject, I'm afraid I have neither the time nor the inclination
to either read or contribute.
Roy Harrington
Charles Pezeshki wrote:
>
> Hi Paul,
>
> I understand your point mathematically.
>
> How 'bout pragmatically? A greater range of light intensity would make it
> easier to design a receptor that would measure that range more exactly. To
> me, it would be similar to boosting the signal/noise ratio. That's why I
> think it would be easier.
>
> But I'm no expert on this stuff.
>
> Chuck
> Chuck Pezeshki
> http://users.moscow.com/pezeshki
>
Hi Chuck and Paul,
Now that we've come down to the pragmatic issues, I'd like to share
my experiences in the way its worked for me. For quite a few years
I've printed 8x10 prints to be scanned on an old HP ScanJet IIcx.
If you look at my website, just about all the images have been done
this way. Certainly the resolution is very low but the tonal range
if the same as anything. The scanner is just 8 bits of grayscale
(I'm not using color) and just reflective. I won't claim perfect
results but I can get very good results almost automatically i.e.
no extensive photoshop efforts to make it look good.
Recently, I decided I like more resolution and transparency
capability (actually negatives) so I purchased the lastest Epson
1600 Scanner. It does 12 bits per grayscale and a 3.3 Dmax.
I've just started with it and have scanned several negatives. I'm
not an expert with the new software yet, but just using the
simple, automatic settings -- 8 bits and auto everything -- the
results were very poor. This kind of makes some sense because
the reflective scanner is simply calibrated to give 256 levels
between about 0% and 100% reflectivity. If you have a print with
max white to max black you'll get pretty close to 256 different
levels (just look at the histogram). On the transmission
scanner you have a very different situation. Potentially you
have densities from 0 to 3.3, but in actuality the negative is
more likely in 0.2 to 1.5 range for instance. If you don't
compensate for the actual range, an 8 bit scan could give less
than half the number of different levels. And you'd still have
to adjust the end points to get a reasonable looking print.
In practice, I have gotten very good results by scanning at
12bits, putting it into photoshop, and doing a fair amount of
tonal manipulation -- i.e. setting endpoints and changing the
curves. After all this I change it to 8 bit grayscale and get
full set of 256 different levels. I think overall I can get
better results with the fancier negative scanner, but it really
seems to require much more work.
Roy
--
Roy Harrington
r...@harrington.com
Black & White Photography Gallery
http://www.harrington.com
Roger N Clark
> I think we agree on more than we disagree, Roger. I think the safest thing
> to say is that comparison between photochemistry and digital is neither
> easy nor straightforward!
Yes, I agree. And I think you did a great and more detailed job of
explaining quantization error and limitations in your other threads
than I did above.
Roger Clark
Chris Kelly
You may be knowledgable in this area. Has there been empirical tests
on the resolution of lenses on film? In other words, if you take a lens
with a MTF rating of 80 or so, and a film with a rating of 80 or so,
what resolution do you actually get? Is it a straight multiple? Does film
act in this way at all? I am sure there are plenty of theories floating
around,
but has anyone done some real research on this, and tested it? If so,
do you know where the results can be found?
"Roger N Clark" <rnc...@uswest.net> wrote in message
news:39E544A4...@uswest.net...
Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]
<ct...@columbia.edu> wrote:
> ... if you take a lens
> with a MTF rating of 80 or so, and a film with a rating of 80 or so,
> what resolution do you actually get?
I believe it follows the familiar "inverse of the sum of the inverses" rule:
1/Final_Resolution = 1/Res_1 + 1/Res_2 + 1/Res_3 ... + 1/Res_n
In the trivial example of two identical resolutions, the final resolution
is half of either -- 40 lpmm in your example.
--
: Watching George and Al makes me want to Ralph!
Chris Kelly
I am not srue I buy it. That is why I am asking for empirical evidence.
Who
has actually tested this?
Chris
"Jan Steinman -- jan AT bytesmiths DOT com [remove .gov]"
<J...@Bytesmiths.com.gov> wrote in message
news:Jan-121000...@c248527-c.potlnd1.or.home.com...
Roger N Clark
> Jan,
> I am not srue I buy it. That is why I am asking for empirical evidence.
> Who
> has actually tested this?
Chris,
I beleive Jan is correct. In theory
it must be something close to
this. But film responds differently,
with grain adjancy effects, multiple
color layers, etc, so it may deviate a
little. I believe some optics/film books
can confirm this relation.
But, like I've said many times (mainly
in discussions in rec.photo.digital),
that it makes little sense to claim a
specific MTF without also specifing
the MTF level and wavelength. For
example, a complete spec might be:
MTF = 80 lpm at 50% response, 5000 nm.
Most lenses have 80% MTF but at what response level?
Roger Clark
Mark McGilvray
very decent prints, and is an invaluable proofing device, although it is not
in the same league with the Lightjet for color prints. The inkjet color
gamut is more restricted, but more capable of colors that really POP. The
Lightjet has a color gamut that puts the Epson to shame, but has real
problems reproducing a true red - although the hideously expensive gizmo has
red, green, and blue lasers. Some of my red roses have come back from
Lightjet land looking like garbage, but for anything else color, give me a
lightjet any day. All my other Lightjet colors are truly astounding - IMHO,
if I do my part on the camera end, in scanning, and in digital file
preparation. The color depth and detail I get in my Lightjet prints still
leaves me amazed. My friendly local Lightjet equipped lab tells me this red
defecit is a problem with the photopaper used (as in available today). The
red problem is apparent in other Lightjet prints done at other labs as well.
Allegedly, Kodak has a new 5 layer paper that is just phenomenal and will
better reproduce red. This paper is supposed to be available next spring. I
can hardly wait.
"Charles Pezeshki" <peze...@mme.wsu.edu> wrote in message
news:B5FB5F91.6502%peze...@mme.wsu.edu...
> Hi folks,
>
> Thought the ng might be interested in a report on some 16x20 Lightjet
prints
> I got done recently.
>
> FYI, I've been getting ready for my first show-- 40 pictures and posters
in
> my local art gallery/community center. All the prints 8x10- 12x15 were
done
> on my Epson 1270, I had eight printed on a photo poster printer (color
> 'matted' with text, 2'x3'), and I had 4 printed on the Lightjet (16x20s).
I
> shoot LF Velvia, Astia, and Provia F exclusively-- no negative film, no
B/W.
>
> Nancyscans (www.nancyscans.com) scanned nine of my 6x12 and 4x5
> transparencies with their Tango drum scanner. They also did the Lightjet
> printing. I scanned the rest using my Microtek 4 scanner, and have a
friend
> with a 5, which I also used.
>
> Here's my rundown (and yes, I know about the possible color problems with
> the Epson 1270).
>
> 1. There is no color difference between the Epson and the Lightjet. None.
> Neither looks better than the other, as long as one is over 200 dpi on the
> scan.
>
> 2. The quality of the Tango scans was outstanding-- but not significantly
> better than the 5, except for the increased resolution. Only on
> difficult-to-scan transparencies was there a noticeable difference. The
> great thing about the Tango scans was that Nancyscans did the color
> matching, and they did an excellent job. Conclusion? For difficult to
scan
> transparencies, go with the Tango. For dpi, go for the Tango. Otherwise,
> use the 4 or 5 (the 5 is definitely better than the 4, even for contrast).
> It's perfectly fine for 80% of the work. The 6x12s definitely did better
> with the Tango-- these are hard to hold with the Microtek holders.
> Occasionally, the 4 produces artifacts. The Tango produces no artifacts.
> These usually only affect one scan line, and are not noticeable.
>
> 3. There is no sharpness difference between the Epson and the Lightjet.
>
> 4. The two drawbacks to the Epson are-- lack of surety regarding the
> archival nature (we're framing and matting ours right now, to take care of
> the various chemical problems) and as the printer ages, there seems to be
> small defects (unnoticeable except by me) that crop up-- pizza wheels,
some
> aliasing phenomena, etc. They're not consistent, and not noticeable.
There
> are NO defects in the Lightjet prints.
>
> 5. Color matching between the Tango scans and the Lightjet prints are
> exact. Color matching between my profiled monitor and the Lightjet prints
> are exact. One of my prints appeared a hair dark, and it had on the
> monitor. In the future, I'll use my calibrated monitor with confidence in
> adjusting both brightness and color.
>
> 6. Nancyscans is competing on price, and quite frankly, I'm having a real
> hard time seeing why anyone who understood the technology would go with a
> more expensive printer if you want a Lightjet print. $68 for the first,
and
> $27 for consecutive was what I paid for the prints.
>
> 7. I love the latitude the high scan density of the drum scans give you.
> You pay, though-- $50/scan. The Microtek won't be going anywhere soon. I
> did great with my Epson 1270 and the Microtek 5 on prints to 12"x16",
> though. 234 dpi was indistinguishable from any higher resolutions.
>
> 8. If I owned a Lightjet, I sure would hope that I'd make my money off it
> in the next 2-3 years. The only edge is the assured archival nature and
> potential large sizes. The desktop Epson did as well with the smaller
> sizes.
>
> Overall, I can't even fathom going back to dealing with traditional
chemical
> printers. The quality of all this digital stuff is wild.
>
> Just my 2 cents. Hope this helps folks. This isn't intended to be a
formal
> study-- just my experience.
Mark McGilvray
OUTPUT SIZE. If I recall, film recorders operate at about 2000dpi.
"RCoda" <rc...@nospam.com> wrote in message
news:I30C5.56729$4d.87...@news02.optonline.net...
> Ah, but what if...
>
> ...they could output a BW scan (corrected, burned, dodged, etc) to some
> large format BW duplicating film in a LightJet, process that film
normally,
> then take it home and make LARGE BW contact prints.
>
> I've been trying to find a lightjet operator who's willing to try this,
but
> no one wants to touch it.
>
> Rich
>
> www.pctype.com/rmcphoto
>
>
>
>
> "geod" <ge...@cwo.com> wrote in message news:B5FCC5F9.8890%ge...@cwo.com...
> > I've said it before and I'll say it again:
> >
> > Lightjet for B&W looks flat, uninteresting and, unless you like
> photographs
> > that look clipped from a glossy magazine, horrible. Hand-made fiber
> prints
> > have so much more depth and richness that there's no comparison. The
> medium
> > simply isn't suited to black and white.
> >
> > > From: Wisbl <isabe...@earthlink.net>
> > > Organization: EarthLink Inc. -- http://www.EarthLink.net
> > > Reply-To: isabe...@earthlink.net
> > > Newsgroups:
rec.photo.equipment.large-format,rec.photo.technique.nature
> > > Date: Sun, 01 Oct 2000 04:47:16 GMT
> > > Subject: Re: Lightjet Print report
> > >
> > >>
> > >> Great(!) report, Chuck.
> > >>
> > > It's my understanding that Lightjet prints are outputted on Fuji
> > > Chrystal Archive paper. That paper is estimated to last 60 years tops.
> > > That doesn't sound archival to me.
> > >
> > > William
> > >
> > >
> > >> Can you describe your monitor calibration process? Are you on a Mac
or
> Wintel
> > >> machine?
> > >>
> > >> Keith
Mark McGilvray
Hi Chuck. Nice report.
Charles Pezeshki wrote:
Here's my rundown (and yes, I know about the possible color problems with
the Epson 1270).1. There is no color difference between the Epson and the Lightjet. None.
Neither looks better than the other, as long as one is over 200 dpi on the
scan.
I find this hard to believe because
every medium has slightly different
color properties, even between
photographic papers. So when you say
there is no color difference,
do you mean from a general viewpoint considering
the inherent color differences
between media, or do you mean that
every shade/hue of red, green,
blue looked exactly the same?
If it really is the same,
this must be due to calibration
of your monitor/printer/lightjet?
But even a calibrated monitor looks different
to me than any print because a monitor is emitted light
versus reflected light from paper,
which looks different in different lights.
Could you clarify things a little?I think it would be nice if there were a
small device you could put on your
monitor and on the print so it could
read the color, contrast values and make
the calibration very robust. I've heard
rumors that such a device would soon come
to market. Anyone heard of it?I've been able to "calibrate" my system
(hp printer, monitor and lightjet output)
by iterative trial. I've found that the photo
papers (both fuji crystal archive and
cibachrome) have very subdued greens
compared to my scanned images
(of velvia). But I've learned how to
match my monitor and inkjet printer
output by boosting saturation ~20% for
the images I send to the lightjet.
2. The quality of the Tango scans was outstanding-- but not significantly
better than the 5, ..snip...The Nancyscans pricing looks pretty good. What is their pricing
for large scans (they say call over 200 MB). I get 650 MB scans
of 4x5 for about $100.
3. There is no sharpness difference between the Epson and the Lightjet.
But the lightjet is more continuous tone, so I think you'll
find that on some colors (and this is getting rarer as inkjets
get better) that the lightjet will be better. The greatest
differences has been in skin tones in my experience. But I have
no experience with the latest epsons.4. There
are NO defects in the Lightjet prints.
My experience also. They are awesome!
6. Nancyscans is competing on price, and quite frankly, I'm having a real
hard time seeing why anyone who understood the technology would go with a
more expensive printer if you want a Lightjet print. $68 for the first, and
$27 for consecutive was what I paid for the prints.
Their pricing is good. For example,
Reed Photo in Denver is $253 for a 30x40
($156 2nd, $124 6+); nancyscans is $192, $88 2nd).
I'll have to try nancyscans. But do they do
cibachrome? I've done tests on both
crystal archive and cibachrome, and the colors
are not nearly as good on the fuji (greens in
particular).
7. I love the latitude the high scan density of the drum scans give you.
I strongly agree!
Overall, I can't even fathom going back to dealing with traditional chemical
printers. The quality of all this digital stuff is wild.
I agree. I've put my darkroom on hold now for 4 years.
(I had one--then I moved--now I have a room painted black
and its dark--but I only use it for loading the 4x5 sheets.)Roger Clark
http://www.users.uswest.net/~rnclark (Home Page Photography)
Howard A. Cubell
Mark McGilvray wrote:
> Your post is interesting......,but,.... I have an Epson 1270, which does
> very decent prints, and is an invaluable proofing device, although it is not
> in the same league with the Lightjet for color prints. The inkjet color
> gamut is more restricted, but more capable of colors that really POP. The
> Lightjet has a color gamut that puts the Epson to shame, but has real
> problems reproducing a true red - although the hideously expensive gizmo has
> red, green, and blue lasers. Some of my red roses have come back from
> Lightjet land looking like garbage, but for anything else color, give me a
> lightjet any day. All my other Lightjet colors are truly astounding - IMHO,
> if I do my part on the camera end, in scanning, and in digital file
> preparation. The color depth and detail I get in my Lightjet prints still
> leaves me amazed. My friendly local Lightjet equipped lab tells me this red
> defecit is a problem with the photopaper used (as in available today). The
> red problem is apparent in other Lightjet prints done at other labs as well.
> Allegedly, Kodak has a new 5 layer paper that is just phenomenal and will
> better reproduce red. This paper is supposed to be available next spring. I
> can hardly wait.
There are in fact some color labs around the country that do Lightjet prints
using Ilfochrome paper rather than Fuji Crystal Archive paper. While the
Ilfochrome paper may introduce undesirable levels of contrast and block up the
shadows a bit(which you may be able to fix digitally using a digital version of
a contrast reduction mask), I can assure you that you will be very pleased with
the reds you get with the Ilfochrome material.
Mark McGilvray
for the information.
"Howard A. Cubell" <hcu...@mediaone.net> wrote in message
news:39EE5FCB...@mediaone.net...
Mark McGilvray
media?
"Howard A. Cubell" <hcu...@mediaone.net> wrote in message
news:39EE5FCB...@mediaone.net...
>
>