35mm=1 Billion Pixels?
synchro505
billion pixels on a digital camera. Is that 1,000 megapixels?
I'm glad I'm starting over with SLR cameras. I recently got an old Canon T50
and am collecting old lenses. The shots I'm getting are amazing to me. My
old digital was 1.3 megapixels L
Mike
leah
http://www.shortcourses.com/choosing/sensors/05.htm
Here is an excerpt from that site: "The most expensive professional digital
cameras give you about
6-million pixels. Although impressive, not even these
resolutions match the estimated 20 million
or so pixels in traditional 35 mm film and 120 million in your
eye. "
Colyn
<synch...@nospam.hotmail.com> wrote:
>I read somewhere that an exposure on 35mm film is the equivalent to 1
>billion pixels on a digital camera. Is that 1,000 megapixels?
>
Actually I think it's in the range of 20-40 million..
>
>I'm glad I'm starting over with SLR cameras. I recently got an old Canon T50
>and am collecting old lenses. The shots I'm getting are amazing to me. My
>old digital was 1.3 megapixels L
>
You can't beat traditional film cameras for overall quality. Some will
say digital is better but even the developers say it's not so..
******************************************************
******************************************************
Colyn Goodson
Ft. Worth, Texas
Leica camera/lens information
The greatest undeveloped territory in the world lies under your hat.
William E. Graham
billion, or 10 to the 9th power.
"synchro505" <synch...@nospam.hotmail.com> wrote in message
news:uno2eup...@corp.supernews.com...
Mxsmanic
news: uno2eup...@corp.supernews.com...
> I read somewhere that an exposure on 35mm
> film is the equivalent to 1 billion pixels
> on a digital camera. Is that 1,000 megapixels?
It would be, but 35mm film doesn't come anywhere close to 1 billion pixels.
Technical Pan, under absolutely ideal conditions, will give you about 310
megapixels. Under more realistic circumstances, Provia 100F will give you
about 35 megapixels. The number of usable pixels varies with the film and
lens, of course. With coarse-grained, low-resolution films, 6-8 megapixels
is about the most you can get (equivalent to the best digital SLRs).
Mxsmanic
3D7C2AD1...@apex.net...
> "... and 120 million in your eye. "
The highest usable resolution for human viewing of an image viewed from
"standard" distance (equal to the diagonal of the image) is about 6-8
megapixels. Under ideal conditions with perfect vision, it can be 2-3 times
higher. It's not anywhere close to 120 million, though.
Tony Spadaro
--
http://www.chapelhillnoir.com
Photography ... the hard way
and partial home of
The Camera-ist's Manifesto
a Radical approach to photography.
The Links are at
http://home.nc.rr.com/tspadaro/links.html
"synchro505" <synch...@nospam.hotmail.com> wrote in message
news:uno2eup...@corp.supernews.com...
DM
> billion pixels on a digital camera. Is that 1,000 megapixels?
I seriously doubt that. It also depends on what film you use. For
example if you're using garbage like Kodak max 800/400, you might
make do with a 1MP digicam.
T.P.
>Actually I think it's in the range of 20-40 million..
That's about right. But the 20-40 million pixels will each need to
record full colour. The current "16 million" pixel backs for medium
format cameras are still way, way short of that, because each pixel
records only one colour. It takes a group of four single colour
pixels (red, blue, green, green) to record full colour.
Of course there's always Foveon ...
ROTFL!!!
Gordon Gekko
> I ran across this link that compared 35mm film and digital cameras:
>
> http://www.shortcourses.com/choosing/sensors/05.htm
>
> Here is an excerpt from that site: "The most expensive professional digital
> cameras give you about
> 6-million pixels. Although impressive, not even these
> resolutions match the estimated 20 million
> or so pixels in traditional 35 mm film and 120 million in your
> eye. "
I'd like to see real proof, using side by side comparisions, rather
than blanket statements. It's easy to say that film has 20 million
pixels. If it's true, it should be easy to show a side by side
comparision showing films huge advantage. Someone show me.
Kinon O'Cann
I have seen prints, of the same scene, taken by a Nikon D1x and an F100
loaded with Reala. Same scene, same size enlargement. The traditional
chemical process blew the digital out of the water, no contest. The facility
I work in has a very nice photo lab, and while they do much of their work in
digital, when quality is a must, film rules the day.
Here's the problem with most so-called comparisons: take a look at American
Photo this month. They claim to compare the Canon D60 with medium format.
Right. However, they actually compare the D60 to a scanned slide or neg.
Sorry, this doesn't work, since once you introduce a digital conversion, you
have poisoned the test. What you need to see, and I don't really know where,
is a test that truly compares a full-digital process with full-chemical. The
difference is incredible, and chemical wins every time.
"Gordon Gekko" <gordong...@yahoo.com> wrote in message
news:d07edc6d.02090...@posting.google.com...
Steffen Kluge
>Actually that is about 100% bull.
Only about 50%, I'd say.
First of all, the film is hardly the limiting factor. Good lenses
for the 35mm format give about 80 lines/mm resolution. That's
actually 80 line *pairs* (one white and one black) at a contrast
of 1:1000 (about 10bit). Thus, translating to 160 pixel lines per
millimetre and a format of 36mm x 24mm that yields about 22
megapixels.
This completely disregards the analog nature of film, though.
With film, pixels are randomly arranged (perfectly "dithered"),
of largely differing sizes, and (more importantly): who says the
contrast has to be 1:1000? You can actually make out vastly more
detail if you scale down the contrast requirement.
Anyway, 22 megapixels is of course far short of the capabilities
of film. Velvia does 140 lines/mm I think, and I had read a
resolution test of some slow Ektar that managed 180 lines/mm.
Cheers
Steffen.
Karl Winkler
My rough estimate was that for good film (high quality slide film like
K64, Velvia, Provia, Astia), that you can get 20-30 Mega pixels worth
of data. However, analog is always different than digital... one
reason is because pixels are in a grid and film grain is truly random.
This equates to the ability to capture subtle patterns, shapes and
gradations in ways that digital cannot. Same for analog tape or LP vs.
CD. Same for movie film vs. videotape. There's definitely a
difference, even if strict numerical "resolution" is the same.
Skip
processes. Any posting on the Usenet is diminished by the limitations of
the digital medium itself. Thus a print made from a film original becomes a
digital image, itself. If you were in the San Diego area, I could direct
you to just such an example, but you're not.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Gordon Gekko" <gordong...@yahoo.com> wrote in message
news:d07edc6d.02090...@posting.google.com...
Shmoo
wrote:
>According to Tony Spadaro <tspa...@ncmaps.rr.com>:
>>Actually that is about 100% bull.
>
>Only about 50%, I'd say.
>
>First of all, the film is hardly the limiting factor. Good lenses
>for the 35mm format give about 80 lines/mm resolution. That's
>actually 80 line *pairs* (one white and one black) at a contrast
>of 1:1000 (about 10bit). Thus, translating to 160 pixel lines per
>millimetre and a format of 36mm x 24mm that yields about 22
>megapixels.
This is the number I have heard before as the realistic maximum MP for
the 35mm film/lens system.
>This completely disregards the analog nature of film, though.
>With film, pixels are randomly arranged (perfectly "dithered"),
>of largely differing sizes, and (more importantly): who says the
>contrast has to be 1:1000? You can actually make out vastly more
>detail if you scale down the contrast requirement.
>
>Anyway, 22 megapixels is of course far short of the capabilities
>of film. Velvia does 140 lines/mm I think, and I had read a
>resolution test of some slow Ektar that managed 180 lines/mm.
>
>Cheers
>Steffen.
While I am 100% digital now, I agree that film is still the best
choice for demanding work that reproduced 8x10 or larger. I think that
with the pending release of sensors in the 10-12Mp range, film will
have a much tighter race for its' money. Of course this is only in the
35mm format.
There are people that love film and those that love digital, and I say
leave them alone. Everyone is trying to get others to switch sides. Go
out and enjoy your photography either pixel or grain based.
Shmoo
wrote:
Actually the need to record full color at each pixel is not completely
true. It would be nice, but is far from the necessity people make it
out to be.
If you read up on the subject you will find the human eye is much more
sensitive to changes in brightness than changes in color. And much
more sensitive to green shades than either red or blue, so the mosaic
sensors are designed to take advantage of our optical system.
At least we agree on Foveon.....
William E. Graham
> pixels. If it's true, it should be easy to show a side by side
> comparision showing films huge advantage. Someone show me.
The problem is, how can they show you? - On your computer monitor? -
No way! - You will have to be shown "live".....
Mxsmanic
kplpnu04tbptm0hkh...@4ax.com...
> Actually the need to record full color at
> each pixel is not completely true. It would
> be nice, but is far from the necessity people
> make it out to be.
It's not important for human viewing at normal distances. The matrix filter
on a CCD and the Bayer pattern are designed to mimic the structure of the
human retina, as you have correctly observed. Thus, although color
resolution on a digicam is about three times worst than luminance
resolution, exactly the same is true for the human eye, so it doesn't matter
(you can't see the lack of color resolution).
In enlargements, however, the smearing of the colors eventually becomes
apparently, generally well before the lack of luminance resolution starts to
be a problem. The solution, of course, is to avoid big enlargements, or to
show with more resolution.
Mxsmanic
d07edc6d.02090...@posting.google.com...
> If it's true, it should be easy to show a
> side by side comparision showing films huge
> advantage.
I've examined them side by side, and film wins.
Bruce Murphy
> "Shmoo" <n...@mailplease.com> a écrit dans le message de news:
> kplpnu04tbptm0hkh...@4ax.com...
>
> > Actually the need to record full color at
> > each pixel is not completely true. It would
> > be nice, but is far from the necessity people
> > make it out to be.
>
> It's not important for human viewing at normal distances. The matrix filter
> on a CCD and the Bayer pattern are designed to mimic the structure of the
> human retina, as you have correctly observed. Thus, although color
Designed to mimic structure? bollocks, laddie. You might have meant
'response', but the similar constraints on chrominance/luminance were
both imposed by a planar structure and certainly weren't design
goals. In reality, the chrominance/luminance tradeoff people talk
about isn't actaully as clear cut as all that either. There are no
'intensity only' pixels in camera sensors.
B>
David Eppstein
"Mxsmanic" <mxsm...@hotmail.com> wrote:
> In enlargements, however, the smearing of the colors eventually becomes
> apparently, generally well before the lack of luminance resolution starts to
> be a problem.
I'm not sure about that part. When I've overcropped digital photos,
it's the lack of luminance resolution that looks bad to me.
--
David Eppstein UC Irvine Dept. of Information & Computer Science
epps...@ics.uci.edu http://www.ics.uci.edu/~eppstein/
Mxsmanic
m2u1kyl...@fuscipes.rattus.net...
> Designed to mimic structure?
Yes. The Bayer pattern includes far more green than red or blue pixels
because the human retina includes far more green-sensitive cones than it
does blue- or red-sensitive cones. Thus, the CCD filter mimics the
distribution of cones in the retina, and this is not a coincidence.
Mxsmanic
eppstein-BE5D86...@news.service.uci.edu...
> I'm not sure about that part. When I've
> overcropped digital photos, it's the lack
> of luminance resolution that looks bad to me.
I agree. My mistake.
The smearing of colors is still a problem, too, of course. When you examine
film scans next to digital photos up close, the digital photos always show
smearing of colors, whereas the film scans do not.
Bruce Murphy
Actually, all the figures I've seen show tiny numbers of blue cones
and twice as many red as green. The Bayer map replicates the
*sensitivity* ratio, but has nothing whatsoever to do with the
structure.
Of course, unless virtually all the colour was concentrated in one spot
the CCD has nothing whatsoever to do with the structure of the retina
and what you are *actually* talking about is the human visual field.
B>
Mxsmanic
> Actually, all the figures I've seen show
> tiny numbers of blue cones and twice as
> many red as green. The Bayer map replicates
> the *sensitivity* ratio, but has nothing
> whatsoever to do with the structure.
Correct; my mistake. The principle remains the same, however: green is
favored in the Bayer pattern because it is the color that human beings can
see with the greatest accuracy.
This is also why color television systems typically treat green
preferentially. For example, PAL and SECAM treat blue and red as
low-resolution deviations from green, with green being equated to the
luminance remaining minus the other two colors (giving green the highest
resolution).
Bruce Murphy
Actually, this is more a historial accident since the requirement to make
just a single channel useful for monochrome display. Not unreasonable to
choose green since it's the closest colour response to luminance, but
not quite the involved design you imply either.
B>
Annika1980
>The only way to see this is to look at prints in person made from both
>processes. Any posting on the Usenet is diminished by the limitations of
>the digital medium itself. Thus a print made from a film original becomes a
>digital image, itself.
Is it any more fair to compare prints made from digital images? At that point,
the digital image is now "analog."
Film's quality advantage over digital has gone from "no contest" to "slightly
better" and now is approaching "theoretical."
Soon it will be, "You still use FILM?"
Skip
image, as long as a reasonably decent printer is used, or something like
Fuji Frontier. C'mon, Bret, you're just being argumentative."Annika1980" <annik...@aol.com> wrote in message
news:20020913220709...@mb-ba.aol.com...
Alan Browne
I calculate the "best case" for Velvia (1:1000 Contrast) as 88 Million
pixels. (160 lp/mm)
For the "old Royal Gold 25" it is even better (138 M)
For Tech Pan 25 (B&W) it is 354M.
Alan.
Annika1980
>Demonstrably, there is no loss of quality printing a hard copy of a digital
>image, as long as a reasonably decent printer is used, or something like
>Fuji Frontier. C'mon, Bret, you're just being argumentative.
Who, moi?
If we assume as you suggest that the printing process (Fuji Frontier system)
can capture all the details of both film and digital, then it seems that what
is really at issue here in the big "film vs. digital" debate is whether a print
looks better than a digital image on screen.
I don't think I'm too far off to suggest that if a film image loses something
in the conversion to digital, than a digital image's properties could lose
something in the conversion to printed output (analog).
Anyway, what has been stated by others here is that if you compare prints from
film and digital images side-by-side the film image will look better. Well it
should given that film (print film anyway) is designed to be printed. Would
the same advantage be observed when printing from slides? Probably not.
Annika1980
>I calculate the "best case" for Velvia (1:1000 Contrast) as 88 Million
>pixels. (160 lp/mm)
>
>For the "old Royal Gold 25" it is even better (138 M)
>
>For Tech Pan 25 (B&W) it is 354M.
>
How did you arrive at these numbers?
I'm not disputing you math, but the question arises as to whether these
numerical advantages translate to observable differences. IOW, can you tell
the difference between 88M pixels and 138M pixels under real-world conditions?
And how large do you have to go before these differences become apparent?
Skip
even Ilfochrome, doesn't have the range that, say, K64 has, so you do lose
something. But resolution isn't among the victims. And a film image
displayed digitally won't necessarily look worse than its digital
equivalent, just the same, leading to the misleading conclusion that there
is no difference between film and digital.
And if, for example, Frontier captures all the detail of both digital and
film, then film will necessarily look better, because there is more
information available to the machinery from film. But, I believe, there is
some degradation of image between film and Frontier, because of the digital
nature of the equipment. But I've not seen anything to back that up, not
having put the same image on a Frontier print and on a traditional "wet"
print. The only loss that a digital image should suffer in the transition
from digital file to a hard copy would be due to the limitations of the
printer, and there are printers available widely that would exploit the full
potential of the file.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Annika1980" <annik...@aol.com> wrote in message
news:20020914203147...@mb-cu.aol.com...
Skip
differences could be perceived by the human eye and processed by the brain.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Annika1980" <annik...@aol.com> wrote in message
news:20020914203504...@mb-cu.aol.com...
William E. Graham
film size.....
"Annika1980" <annik...@aol.com> wrote in message
news:20020914203504...@mb-cu.aol.com...
Mxsmanic
20020914203504...@mb-cu.aol.com...
> IOW, can you tell the difference between 88M pixels
> and 138M pixels under real-world conditions?
It all depends on the size of the enlargement. For a snapshot, you cannot
tell. For a wall-sized enlargement examined from inches away, it's pretty
easy to see.
> And how large do you have to go before these
> differences become apparent?
Very large. For an ordinary print viewed from an ordinary distance, 6-8
megapixels is all you'll ever need, because that's about all you can see
under such conditions.
Robert Monaghan
would seem to require at least two pixels (one black set for the black
line, one white set for the white space inbetween lines). So you need to
double your numbers for both axes, or multiply your result by 4, yielding
the indicated figures...
film has tremendous data capabilities; this is why many laser scanners can
yield many hundreds of megabytes of data from a film scan - data, not
grain ;-)
hth bobm
--
* Robert Monaghan POB752182 So. Methodist Univ., Dallas Tx 75275 *
* Third Party 35mm Lenses: http://medfmt.8k.com/third/index.html *
* Medium Format Cameras: http://medfmt.8k.com/mf/index.html *
William E. Graham
will catch up to us in a couple of years.....
(pixelwise, that is....)
"Robert Monaghan" <rmon...@smu.edu> wrote in message
news:am0q6i$uf5$1...@post.cis.smu.edu...
Meryl Arbing
couple of rolls processed and printed so, I found myself forced to go to a
Fuji Frontier lab.
The results were disgustingly bad! I can't imagine why there is all this
"gushing" about Fuji Frontier since it stinks! The comparison to the
traditional prints that I usually have done is laughable. Even on 4x6's the
differences were noticible and, looking through a 5x loupe, they jumped out.
I would NEVER use Fuji Frontier again!
"Skip" <shadow...@cox.net> wrote in message
news:laRg9.16166$S32.1...@news2.west.cox.net...
Skip
Frontier system, and the results used to vary, until she made it clear that
anything but their best effort was unacceptable. Now it works pretty good.
(She's not loud, just persistent. When something came back that wasn't
right, she just has them do it over again. After awhile, they get the idea
that it is much less effort to do it right the first time, rather than do it
over 3 or 4 times.)
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Meryl Arbing" <mar...@sympatico.ca> wrote in message
news:uc%g9.19$_S6....@news20.bellglobal.com...
Meryl Arbing
answer...Forget the Fuji Frontier!! If you can't get away from this horror
then stick with slides where you don't have to worry about some part-time
high-school making a mess of your work!
"Skip" <shadow...@cox.net> wrote in message
news:qg1h9.18856$S32.1...@news2.west.cox.net...
Bruce Murphy
> That makes the whole thing too hit and miss to be useful. The simple
> answer...Forget the Fuji Frontier!! If you can't get away from this horror
> then stick with slides where you don't have to worry about some part-time
> high-school making a mess of your work!
I'd be delighted for you to explain how the Frontier is bad becuase
part-time high school operators can produce bad results with it. Hasn't
that been the issue with minilabs *forever*? I can only presume that
an idiot with a Frontier can't match good prints from a pro lab in
your experience, but I'm not terribly surprised.
I happen to have access to one (of four or five close) Frontier using
places which have a selection of very good staff doing the printing
work *only* (other people do all the other bits and pieces to do with
the store) and a relationship with the place good enough that
important work will be held over until the most experienced person can
deal with it.
No, I'm not getting cibachrome-level results, but looking at the prints
hanging on my wall right now, they're pretty nice nonetheless.
B>
Annika1980
>> IOW, can you tell the difference between 88M pixels
>> and 138M pixels under real-world conditions?
>
>It all depends on the size of the enlargement. For a snapshot, you cannot
>tell. For a wall-sized enlargement examined from inches away, it's pretty
>easy to see.
How often do you view wall-sized enlargements from just inches away?
-----------------------------------
> And how large do you have to go before these
>> differences become apparent?
>
>Very large. For an ordinary print viewed from an ordinary distance, 6-8
>megapixels is all you'll ever need, because that's about all you can see
>under such conditions.
EX-actly!
Gordon Moat
digital. While this is very different to some people than direct digital
SLRs, the capabilities are quite high. Grab most any high quality printed
photo book at your nearest bookstore, and take a look. Lots of detail, lots
of colour quality, and mostly from drum scans.
Resolution is only one issue. When it is still possible to capture more
information with drum and film scanners, than direct digital, then I think
digital will never catch up. The only way that would ever happen is if
development stops on film and drum scanners, which I do not expect for a
few decades . . . if ever.
Now if the manufacturers could meet, or even exceed that level, they would
truly be there. However, the reality is that the average consumer buying
this stuff is quite happy with a plastic zoom lens and "good enough"
results. The consumers will dictate the market, and high end digital gear
(scanners, backs, and digital SLRs) will be used be a much smaller
percentage of the market.
Ciao!
Gordon Moat
Alliance Graphique Studio
<http://www.allgstudio.com/gallery.html>
brian
> I calculate the "best case" for Velvia (1:1000 Contrast) as 88 Million
> pixels. (160 lp/mm)
>
> For the "old Royal Gold 25" it is even better (138 M)
>
> For Tech Pan 25 (B&W) it is 354M.
>
>
> Alan.
>
>
10MP and 53MP images at my site that you can view on your screen or
download and print:
http://www.caldwellphotographic.com/FullSizeMosaicMain.html
I would be very interested if anyone living in or visiting New York
City would go to the same site and shoot the same scene with a 35mm
film camera and then compare images. I doubt that 35mm - even Tech
Pan - would do very well even compared to the 10MP image. Large
format would be required to get to the same quality level as the full
size 53MP image because 35mm and medium format would be inadequate.
Consider it a challenge: if you doubt me, then prove me wrong with
actual images. Statements based on half-baked calculations that
ignore most of the important variables are completely uninteresting to
me.
The image in question can be approximately re-created by shooting a
28mm shift lens on 24x36mm format. The required amount of shift is
beyond the capabilities of any real 28mm shift lens, but you will
still be able to compare resolution near the horizon line if you shoot
with any old 28mm lens. If you want to cheat a little to give film a
fighting chance you can use a 35mm lens instead and only match the
horizontal field of view.
Brian
Gordon Gekko
>
> Here's the problem with most so-called comparisons: take a look at American
> Photo this month. They claim to compare the Canon D60 with medium format.
> Right. However, they actually compare the D60 to a scanned slide or neg.
> Sorry, this doesn't work, since once you introduce a digital conversion, you
> have poisoned the test. What you need to see, and I don't really know where,
> is a test that truly compares a full-digital process with full-chemical. The
> difference is incredible, and chemical wins every time.
I'm not sure that this is so unfair.
A digital file is far more useful than a negative, because the digital
file can be printed on an inkjet printer. Inkjet prints are
inexpensive, yet inkjet prints faithfully reproduce all the resolution
in the digital file. On the other hand, whenver an optical process is
used to copy something, there is a loss of resolution.
Tom Thackrey
On 15-Sep-2002, gordong...@yahoo.com (Gordon Gekko) wrote:
> A digital file is far more useful than a negative, because the digital
> file can be printed on an inkjet printer. Inkjet prints are
> inexpensive, yet inkjet prints faithfully reproduce all the resolution
> in the digital file. On the other hand, whenver an optical process is
> used to copy something, there is a loss of resolution.
Any printing process results in some loss of resolution, optical, inkjet or
whatever. My guess is there's more loss in inkjet than optical, but I have
no objective evidence to support that theory. Further, there's a loss in
translating the colors of one medium to another. Inkjet prints cannot
faithfully reproduce any color outside the color space of the particular
printer/paper combination.
--
Tom Thackrey
www.creative-light.com
-----------== Posted via Newsfeed.Com - Uncensored Usenet News ==----------
http://www.newsfeed.com The #1 Newsgroup Service in the World!
-----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----
Tony Spadaro
I've been doing my stitching in Photoshop but I have only shot lateral
panoramas so far. With sufficient overlap (plus a little adaptor I made to
pivot under the nodal point) it's fairly easy without extra software.
I have a swing lens panoramic (HorizonT) and with that I don't need to
find the node, as the lens rotates about it's own nodal point to make the
picture. It has a 120 degree field of view so I've found that rotating the
tripod head 90 degrees gives me plenty of overlap for resolving problems
with things that move between shots. 4 pictures gives a full 360 degree
panorama.
--
http://www.chapelhillnoir.com
Photography ... the hard way
and partial home of
The Camera-ist's Manifesto
a Radical approach to photography.
The Links are at
http://home.nc.rr.com/tspadaro/links.html
"brian" <brian...@aol.com> wrote in message
news:3c459ba.02091...@posting.google.com...
Tony Spadaro
are the good ones and the bad ones and there are the labs where one
particular operator is better than the rest.
--
http://www.chapelhillnoir.com
Photography ... the hard way
and partial home of
The Camera-ist's Manifesto
a Radical approach to photography.
The Links are at
http://home.nc.rr.com/tspadaro/links.html
"Meryl Arbing" <mar...@sympatico.ca> wrote in message
news:Wy1h9.224$w72....@news20.bellglobal.com...
Gordon Moat
assumptions of comparison are not entirely accurate. While it may be a work practice in your area,
it is quite possible to do greater than a 1000 ppi scan using a drum scanner. Just a look at the
specs of one of the Imacon scanners
<http://www.imacon.dk/usr/imacon/wppImacon.nsf/htmlpages/prectechspec.html> which shows one
scanner capable of 5760 dpi optical. It can also handle up to 5" by 7" transparencies.
Drum scanning a single frame of 35 mm film would not reach the same size. While it is possible to
do better than 6000 ppi scans using some drum scanners, it is about the upper limit for useful
information from 35 mm. This serves as a practical limit, since there is little useful information
capture about that. Moving to medium format film would be one way to capture more information.
Even a 5760 dpi scan of 6 by 4.5 film would be larger than your stitched image, but only by a
little bit.
It appears that your image is created from several digital acquisitions and then stitched. You
have 42 such captures shown. If I am wrong here, please correct me, but it would seem that even
accounting for some frame loss for overlap, there are at least 20 equivalent full (CCD) frames for
the resulting image. I also wonder why you did JPEG, since it is a lossy compression; why not use
TIFF or raw format?
What would keep someone from taking 20 to 42 full frame 35 mm film sections, drum scanning each,
and stitching those together? That would seem to me to be a more accurate comparison, after all,
you did not just do one digital frame. Also, it does not seem there is any limitation to using the
same tripod head as you did.
Your 5761 by 9194 image is created from several to make one complete frame. That is about a 19" by
30" image using high quality 300 dpi printing, (similar to many high quality magazine offset
printing specifications). I can see where this could be better printed than a chemical
photographic print, but is that what your comparison challenge is all about? I would think that
scanned film compared to direct digital would be a better comparison.
Drum scanned medium format would exceed your image on pixels alone. Colour depth, and the amount
of colour information are other possible comparisons. Resolution is only one issue. However, all
this might be moot, since the end printing technology used would be the true comparison, and could
make the differences difficult to discern.
Kodak Pro Photo CD has an upper limit of 4096 by 6144. While this is close to your 5761 by 9194
image, it would only give a 13.5" by 20.5" image at 300 dpi. Obviously a cheap alternative does
not get too close to what you have. Interestingly, cost is another issue. Locally, I can get a
Scitex drum scan up to 650 MB file size for around $100 from 4" by 5" film, or medium format. Not
really too much, unless I have many scans to complete. I have worked on up to 2 GB image files on
my computer gear, though the computer gets noticeably slower when doing these. My normal working
scans are 70 MB to 160 MB in file size as three channel files.
B&H shows the Kaidan QuickPan Spherical at under $500. They also have the Nikon D1x at $4650.
Using that same $4650, and a quick check of scanners, shows the Imacon Flextight for $4900, though
it will only do 3200 dpi from up to 6 cm by 7 cm roll film. That is still good enough for a 7050
by 8750 image, from one frame of film. Using several frames, a smaller format, and/or a lower cost
film scanner, could achieve the same. You have achieved excellent results from your expenditure,
but there are comparable options, some for even less expense.
I wish I was in New York, so I could take your challenge . . . not to prove you wrong, but to show
there is more than one way to get the same results, or better results. I take nothing away from
what you have achieved. It may even be quite possible to do it for less money, even using the same
tripod head you used, by originating on film, and scanning that film. It may even take less than 4
minutes and 21 seconds for the shot, or shots, needed to get the same final image.
Your information has been very informative, and well prepared. I do not think it is fair to
compare stitched images to individual, but I would be interested in your answer as to why you
would not consider stitched and scanned film images. Of course, you could easily dismiss my post
because it does not meet your rules or criteria, though I think are too professional to do so.
Ciao!
Gordon Moat
Alliance Graphique Studio
<http://www.allgstudio.com>
William E. Graham
than flatbed scanners?
"Gordon Moat" <mo...@attglobal.net> wrote in message
news:3D84CB21...@attglobal.net...
Mxsmanic
d07edc6d.02091...@posting.google.com...
> Inkjet prints are inexpensive, yet inkjet
> prints faithfully reproduce all the resolution
> in the digital file.
Inkjet prints cost me several dollars each, and they often do not begin to
exhaust the resolution of the digital file.
Mxsmanic
Xw7h9.489995$UU1.78649@sccrnsc03...
> How do drum scanners work?
The negative or transparency is mounted on a spinning drum using a mounting
fluid or oil and the image is scanned as it spins. The use of
photomultiplier tubes (a special type of light-amplifying vacuum tube)
allows extremly large density ranges to be accurately scanned.
> - Why are they any better (resolution wise)
> than flatbed scanners?
They are not necessarily better resolution-wise (although many drum scanners
provide very high resolution), but they have slightly more ability to handle
a broad density range than do the high-end desktop scanners. The gap is
closing, however.
Mxsmanic
> How often do you view wall-sized enlargements
> from just inches away?
By standing inches in front of them.
Mxsmanic
3c459ba.02091...@posting.google.com...
> I would be very interested if anyone living
> in or visiting New York City would go to the
> same site and shoot the same scene with a 35mm
> film camera and then compare images.
It is okay if they cheat and stitch multiple film images together, as you
had to do with digital?
> I doubt that 35mm - even Tech Pan - would do
> very well even compared to the 10MP image.
Oh, I think they would.
> Large format would be required to get to the
> same quality level as the full size 53MP image
> because 35mm and medium format would be inadequate.
Medium-format is easily 80 megapixels.
As long as you cheat by stitching multiple images together, you can get
whatever quality you want. I could stitch 42 medium-format images together
and blow your example out of the water. Even 42 35mm scans would pretty
much put your example to shame.
Additionally, a camera that requires 5 minutes to take one shot isn't very
practical.
Bruce Murphy
> Additionally, a camera that requires 5 minutes to take one shot isn't very
> practical.
Yeah, Ansel. Get back in your box.
B>
Mxsmanic
m2bs6yn...@fuscipes.rattus.net...
> Yeah, Ansel. Get back in your box.
Ansel was the exception, not the rule. His subjects didn't move (at least
not fast enough to cause a problem for long exposures).
I note in passing that Ansel exposed the entire frame at once; he did not
staple 40 images together.
I've been doing some experiments. The Flatiron building image looks very
sharp and clear indeed--until you compare it to medium-format, at which
point it looks like a blur. I don't have any tripod shots on 35mm handy, so
I can't check that out just yet.
brian
Hi Gordon:
I only wanted to point out that I've got some fairly clean samples of
a single image ranging up to 53megapixels in size. These can be used
to visually compare what you can get with other formats in a
single-shot fashion. When the image is used in this sense it doesn't
matter how it was created.
Calculations of image resolution in terms of number of pixels can be
useful, but I think it is always best to back up calculations with
observation. All I'm trying to do is make the observation part more
interesting.
I'm not trying to convince everyone to go out and stitch bunches of
images together just to get more resolution than you can possibly see
on an 8x10" or 11x14" print. I'll grant you that such a technique
will probably appeal to only a few crazed lunatics like myself.
Nevertheless, by putting together 42 6-megapixel images and then
*downsampling* to 53megapixels I wind up with a pretty decent picture
that can be used as a resolution reference of sorts. The 10megapixel
version in particular has essentially no trace of digital artifacts
aside from minor stuff caused by a level-10 JPEG compression.
More comments below:
> While I will admit that you have pieced together some interesting samples, some of your
> assumptions of comparison are not entirely accurate. While it may be a work practice in your area,
> it is quite possible to do greater than a 1000 ppi scan using a drum scanner. Just a look at the
> specs of one of the Imacon scanners
> <http://www.imacon.dk/usr/imacon/wppImacon.nsf/htmlpages/prectechspec.html> which shows one
> scanner capable of 5760 dpi optical. It can also handle up to 5" by 7" transparencies.
>
> Drum scanning a single frame of 35 mm film would not reach the same size. While it is possible to
> do better than 6000 ppi scans using some drum scanners, it is about the upper limit for useful
> information from 35 mm. This serves as a practical limit, since there is little useful information
> capture about that. Moving to medium format film would be one way to capture more information.
> Even a 5760 dpi scan of 6 by 4.5 film would be larger than your stitched image, but only by a
> little bit.
It would be larger, but it would not have the same image quality in
terms of contrast or resolution. Of course, I'm anxious to see a
comparison!
>
> It appears that your image is created from several digital acquisitions and then stitched. You
> have 42 such captures shown. If I am wrong here, please correct me, but it would seem that even
> accounting for some frame loss for overlap, there are at least 20 equivalent full (CCD) frames for
> the resulting image. I also wonder why you did JPEG, since it is a lossy compression; why not use
> TIFF or raw format?
You're estimate of about 20 equivalent frames is pretty accurate. I
save all of my original stitched images as TIFFs. I only JPEGed them
to make it possible to host a true 53MP image. I've got to *PAY* for
that server space, you know!!
>
> What would keep someone from taking 20 to 42 full frame 35 mm film sections, drum scanning each,
> and stitching those together? That would seem to me to be a more accurate comparison, after all,
> you did not just do one digital frame. Also, it does not seem there is any limitation to using the
> same tripod head as you did.
I've heard of people stitching multiple frames of 6x7cm format and
even 4x5 format. Its just not my cup of tea, though. I've still got
several dozen rolls of exposed film that I just never got around to
processing. Digital is a godsend for semi-disorganized people like me
who hate going to stores to deliver and pick up things. Remember, my
purpose here is not to compare digital with film, but simply to offer
up some large reference images so that people can make their own
comparison based on actual images rather than to believe an
over-simplistic calculation.
>
> Your 5761 by 9194 image is created from several to make one complete frame. That is about a 19" by
> 30" image using high quality 300 dpi printing, (similar to many high quality magazine offset
> printing specifications). I can see where this could be better printed than a chemical
> photographic print, but is that what your comparison challenge is all about? I would think that
> scanned film compared to direct digital would be a better comparison.
>
> Drum scanned medium format would exceed your image on pixels alone. Colour depth, and the amount
> of colour information are other possible comparisons. Resolution is only one issue. However, all
> this might be moot, since the end printing technology used would be the true comparison, and could
> make the differences difficult to discern.
In an ideal world, some kind person would download my image, print it
at 48" x 96", and then do the same with images shot with 35mm, 6x4.5,
6x9, 4x5, 8x10, 11x14, and 20x24. However, I would welcome any sort
of thoughtful comparison based on actual images.
>
> Kodak Pro Photo CD has an upper limit of 4096 by 6144. While this is close to your 5761 by 9194
> image, it would only give a 13.5" by 20.5" image at 300 dpi. Obviously a cheap alternative does
> not get too close to what you have. Interestingly, cost is another issue. Locally, I can get a
> Scitex drum scan up to 650 MB file size for around $100 from 4" by 5" film, or medium format. Not
> really too much, unless I have many scans to complete. I have worked on up to 2 GB image files on
> my computer gear, though the computer gets noticeably slower when doing these. My normal working
> scans are 70 MB to 160 MB in file size as three channel files.
>
I would *love* to see a comparison image from drum-scanned 4x5. I may
even do this myself on my next trip to NYC.
> B&H shows the Kaidan QuickPan Spherical at under $500. They also have the Nikon D1x at $4650.
> Using that same $4650, and a quick check of scanners, shows the Imacon Flextight for $4900, though
> it will only do 3200 dpi from up to 6 cm by 7 cm roll film. That is still good enough for a 7050
> by 8750 image, from one frame of film. Using several frames, a smaller format, and/or a lower cost
> film scanner, could achieve the same. You have achieved excellent results from your expenditure,
> but there are comparable options, some for even less expense.
I've already admitted to being a crazed lunatic. Doing cost-benefit
analyses is not something I would ever do when it comes to
photography.
>
> I wish I was in New York, so I could take your challenge . . . not to prove you wrong, but to show
I also wish I was in New York.
> there is more than one way to get the same results, or better results. I take nothing away from
> what you have achieved. It may even be quite possible to do it for less money, even using the same
> tripod head you used, by originating on film, and scanning that film. It may even take less than 4
> minutes and 21 seconds for the shot, or shots, needed to get the same final image.
>
> Your information has been very informative, and well prepared. I do not think it is fair to
> compare stitched images to individual, but I would be interested in your answer as to why you
> would not consider stitched and scanned film images. Of course, you could easily dismiss my post
> because it does not meet your rules or criteria, though I think are too professional to do so.
Remember, I'm not really trying to compare stitched images to
individual ones: I'm just providing clean high-resolution images that
can be used as a sanity check on statements like "35mm Tech Pan =
354Megapixels!" The only reason(s) I don't stitch scanned film images
are that I already have a nice digital camera that gives me images
about as good as my 2700 dpi slide scanner, I *hate* spending my time
scanning and organizing, and I don't want to spend money on film and
processing.
Also bear in mind that my current limitations on resolution are
computer-based and not camera-based.
brian
I'm eager to actually see your results. Can you post them?
brian
> "brian" <brian...@aol.com> a écrit dans le message de news:
> 3c459ba.02091...@posting.google.com...
>
> > I would be very interested if anyone living
> > in or visiting New York City would go to the
> > same site and shoot the same scene with a 35mm
> > film camera and then compare images.
>
> It is okay if they cheat and stitch multiple film images together, as you
> had to do with digital?
This thread is about how many pixels are in a velvia or tech pan 35mm
frame. All I'm doing is providing fairly clean 10MP and 53MP images
for comparison.
>
> > I doubt that 35mm - even Tech Pan - would do
> > very well even compared to the 10MP image.
>
> Oh, I think they would.
Please show this. I don't like being ignorant!
>
> > Large format would be required to get to the
> > same quality level as the full size 53MP image
> > because 35mm and medium format would be inadequate.
>
> Medium-format is easily 80 megapixels.
Meaning that if you *downsample* to 53megapixels the image will be
sharper/clearer/more contrasty than my larger image. This is exactly
the sort of comparison I would like to see.
>
> As long as you cheat by stitching multiple images together, you can get
> whatever quality you want. I could stitch 42 medium-format images together
> and blow your example out of the water. Even 42 35mm scans would pretty
> much put your example to shame.
Getting high resolution, arbitrary field of view and zero illumination
falloff are the main reasons I bother with such a time-consuming
technique. A non-downsampled version of my 53MP image would have been
about 150MP. I currently have a memory bottleneck that limits me. So
using 42 medium format scans wouldn't have helped at all.
>
> Additionally, a camera that requires 5 minutes to take one shot isn't very
> practical.
Its practical for me because its easier to carry and use than my 4x5
or 8x10 stuff. Plus I can also use the digital camera for ordinary
photography. Attacking my methods is irrelevant anyway, because all
I'm really doing here is providing reference images as a reality check
on some very optimistic statements about how many pixels fit into a
24x36mm image.
Bruce Murphy
> "Bruce Murphy" <pack...@rattus.net> a écrit dans le message de news:
> m2bs6yn...@fuscipes.rattus.net...
>
> > Yeah, Ansel. Get back in your box.
>
> Ansel was the exception, not the rule. His subjects didn't move (at least
> not fast enough to cause a problem for long exposures).
But your statement remains. A lot of people are *very* happy using
cameras that take a long time even 5 minutes to take a picture. I
presume you've heard about some of the terrifying resolution scanning
backs for some medium format digital stuff?
"not practical for people unwilling to use anything less convenient
than an APS point and shoot" perhaps.
B>
brian
> Nice shots Brian. I've added your page and tutorials to my bookmarks.
> I've been doing my stitching in Photoshop but I have only shot lateral
> panoramas so far. With sufficient overlap (plus a little adaptor I made to
> pivot under the nodal point) it's fairly easy without extra software.
> I have a swing lens panoramic (HorizonT) and with that I don't need to
> find the node, as the lens rotates about it's own nodal point to make the
> picture. It has a 120 degree field of view so I've found that rotating the
> tripod head 90 degrees gives me plenty of overlap for resolving problems
> with things that move between shots. 4 pictures gives a full 360 degree
> panorama.
>
Tony:
Thanks. Your site has long been bookmarked.
I'm wondering if you could make several 360-degree swing-lens shots in
different elevations and then stitch them together in Panorama Tools
to form a full 360x180 spherical panorama. I do occasionally make
these with a fisheye lens, but there isn't much resolution when you do
that. Going to a 17mm lens (equivalent to a 25mm lens on a D1x) means
I have to stitch 30 individual frames together for a spherical
panorama.
William E. Graham
"Mxsmanic" <mxsm...@hotmail.com> wrote in message
news:%D9h9.54327$AR1.2...@bin2.nnrp.aus1.giganews.com...
Mxsmanic
> Is the negative/film ruined by the oil/mounting fluid?
Apparently not; it doesn't harm the film at all. Or at least that's what
I've been told. I'm wary of spreading any liquid on my negatives or slides,
but perhaps my concern is unwarranted.
Mxsmanic
3c459ba.02091...@posting.google.com...
> I'm eager to actually see your results. Can you
> post them?
Sure. See http://www.smallevents.com/sunset.jpg (1.1 MB), and compare it to
the full-sized image of the Flatiron building.
Remember that the Flatiron picture is a composite of 42 separate digital
pictures. I reduced the medium-format picture so that it matched what I
would obtain if I used 42 separate medium-format exposures to produce one
photo. You can see that the quality of the medium-format shot very visibly
exceeds that of the digital shot. I didn't bother to unsharp mask the MF
shot (not sure if the Flatiron shot has been unsharp-masked or not).
Mxsmanic
> A lot of people are *very* happy using
> cameras that take a long time even 5 minutes
> to take a picture.
A lot of people? Who?
And if the most touted advantage of digital is speed, what does this imply
when it takes five minutes to take a digital picture? Where's the speed
advantage?
> I presume you've heard about some of the terrifying
> resolution scanning backs for some medium format
> digital stuff?
Sure, but since they are _scanning_ backs, they don't count.
You could devise a camera to expose a strip of film continously by scanning
and whip the performance of scanning backs, easily.
It's important to compare apples to apples.
Similarly, it's highly misleading to compare a 42-image composite of digital
photographs to a single shot on film (although LF would still win easily,
and MF can hold its own). To be fair, you must compare it to a 42-image
composite of film photographs, and when you do that, digital loses.
Mxsmanic
3c459ba.02091...@posting.google.com...
> Please show this. I don't like being ignorant!
Tech Pan can provide more than 200 megapixels on a single frame of 35mm
film. And since you apparently allow five minutes for each photo, why not
use three filters and use Tech Pan to produce a full-color image with 200
megapixel resolution?
> Meaning that if you *downsample* to 53megapixels
> the image will be sharper/clearer/more contrasty
> than my larger image.
Yes, but a good scan will be comparable even at full size.
> This is exactly the sort of comparison I would
> like to see.
Done. See http://www.smallevents.com/sunset.jpg. This is a MF image
reduced in size to compensate for the use of 42 separate images for the
Flatiron building (that is, the MF image is reduced by a factor of 42).
You'll see that it is visible better than the digital photo. And this is a
poor example (but I didn't have many daytime photos taken on a tripod
handy).
> Getting high resolution, arbitrary field of view
> and zero illumination falloff are the main reasons
> I bother with such a time-consuming technique.
Why not shoot film, if time is not a problem?
And by the way, why is there grain in the sky of the digital photo?
> A non-downsampled version of my 53MP image would
> have been about 150MP. I currently have a memory
> bottleneck that limits me. So using 42 medium
> format scans wouldn't have helped at all.
But it is worth nothing that 42 MF scans would have easily provided 3374
megapixels, or about 22 times the resolution of the digital composite.
> Attacking my methods is irrelevant anyway, because
> all I'm really doing here is providing reference
> images as a reality check on some very optimistic
> statements about how many pixels fit into a 24x36mm
> image.
Try doing it with unstitched single digital images, instead of mosaics. You
can get any resolution you want from anything with a mosaic.
Bruce Murphy
> "Bruce Murphy" <pack...@rattus.net> a écrit dans le message de news:
> m2it164...@fuscipes.rattus.net...
>
> > A lot of people are *very* happy using
> > cameras that take a long time even 5 minutes
> > to take a picture.
>
> A lot of people? Who?
People doing product work for glossy magazines and posters.
> And if the most touted advantage of digital is speed, what does this imply
> when it takes five minutes to take a digital picture? Where's the speed
> advantage?
In not having to spend hours developing the image before they're being
incorporated into your catalogues. Once again you've completely missed
the point.
> > I presume you've heard about some of the terrifying
> > resolution scanning backs for some medium format
> > digital stuff?
>
> Sure, but since they are _scanning_ backs, they don't count.
Why not? Precisely how is it that they don't count? They're a
specialised extension of the existing stitch-together digital stuff
that we were already talking about, and to suddenly claim that they're
out of bounds. They're in active use and turned up in high-end
commercial photography far ahead of other digital cameras.
> You could devise a camera to expose a strip of film continously by scanning
> and whip the performance of scanning backs, easily.
Except that you wouldn't immediately have available to you the image
that could be incorporated into your various workflow items. I wonder
also with your 'whip the performance' comment if you're aware that
scanning backs were producing on the order of several tens of
megapixels two years or more ago.
> It's important to compare apples to apples.
images to images. I never claimed than a scanning back on an MF body was
as convenient as an APS point and shoot camera, but you seem to believe
that this is relevant.
> Similarly, it's highly misleading to compare a 42-image composite of digital
> photographs to a single shot on film (although LF would still win easily,
> and MF can hold its own). To be fair, you must compare it to a 42-image
> composite of film photographs, and when you do that, digital loses.
Arguable. LF wins a lot by not having to be enlarged much -- a lot of LF
optics are significantly more primitive than contemporary 35mm stuff.
I'm also a little curious about this 'fair' you're talking about. If
stitching provides a convenience/quality tradeoff then the quality
level that can be achieved is based largely on one's personal convenience
threshold. I also note with amusement that you claim stitching many MF
shots allows you to outperform a digital 35-mm class stitch. Does your
example not count as well?
B>
Mxsmanic
> People doing product work for glossy magazines and posters.
0.01% of the market, you mean?
> In not having to spend hours developing the
> image before they're being incorporated into
> your catalogues.
Developing film requires only 15 minutes. Scanning it requires only two
minutes.
Most film images are likely to spend more time in Photoshop than in the
minilab.
> Why not?
Because 99% of all photographs are not of perfectly stationary objects, so a
scanning back won't work. If that were not the case, scanning digicams
would be commonplace by now.
> Except that you wouldn't immediately have
> available to you the image that could be incorporated
> into your various workflow items.
You'd have it a couple of minutes later.
> I wonder also with your 'whip the performance'
> comment if you're aware that scanning backs were
> producing on the order of several tens of
> megapixels two years or more ago.
Yes, I'm aware of that. And MF film backs were producing 100 megapixels
half a century ago.
> Arguable. LF wins a lot by not having to be enlarged
> much -- a lot of LF optics are significantly more
> primitive than contemporary 35mm stuff.
LF has more pixels, and that's what matters.
> I also note with amusement that you claim
> stitching many MF shots allows you to outperform
> a digital 35-mm class stitch.
I'm comparing apples to apples.
Skip
acceptable for some galleries. No choice. And I don't understand the
reference to slides...
Skip
image. And the resulting print will be more accurate to the original than
an ink jet print. Remember, a scan is produced using an optical element,
too. And there are losses due to the digitizing process.
A digital file is no more useful than a negative. A neg can be printed
too, you know, or have you forgotten the old processes of photography?
Inkjet prints cost me the same, or more, than a "wet" print. The paper I
use for both costs me about $1 per sheet, and ink is more expensive than
chemicals. There is a savings of time with digital, however.
Sometimes, I don't know where you get some of your ideas, GG. Think things
through.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Gordon Gekko" <gordong...@yahoo.com> wrote in message
news:d07edc6d.02091...@posting.google.com...
Ralf C. Kohlrausch
"William E. Graham" <we...@attbi.com> schrieb im Newsbeitrag
news:DNfh9.325591$_91.4...@rwcrnsc51.ops.asp.att.net...
> Is the negative/film ruined by the oil/mounting fluid?
>
No, they are wiped clean afterwards. The oil even fills in minor
scratches so that very good results are possible.
Greetings
Ralf C.
Philip Homburg
Mxsmanic <mxsm...@hotmail.com> wrote:
>"brian" <brian...@aol.com> a écrit dans le message de news:
>3c459ba.02091...@posting.google.com...
>
>> Please show this. I don't like being ignorant!
>
>Tech Pan can provide more than 200 megapixels on a single frame of 35mm
>film. And since you apparently allow five minutes for each photo, why not
>use three filters and use Tech Pan to produce a full-color image with 200
>megapixel resolution?
Please point to an online image that contains 200 Mpixels and comes from a
single frame of 35mm film. (or describe what you means with 'can provide').
Philip Homburg
David J. Littleboy
"Mxsmanic" <mxsm...@hotmail.com> wrote:
>
> Tech Pan can provide more than 200 megapixels on a single frame of 35mm
> film.
That seems extremely overly optimistic.
Since a 35mm frame has 864 sq mm, that's 231,481 pixels/sq mm, or 481 x 481
pixels each mm. So you'd need a lens that delivers in excess of 240 lp/mm
over the whole frame to create a (meaningful) 200 MP image. Since there's no
such thing as a lens that does that (heck, there aren't any lenses that
deliver 120 lp/mm over the whole frame), claiming 200 MP is quite excessive,
silly, ridiculous, <pick your favorite pejorative and insert here>.
If you restrict your images to images that appear in nature, you're talking
about not 1:1000 contrast, but 1:1.6 contrast, at which point, even Tech Pan
isn't a whole lot of megapixels.
For far more realistic estimates, see:
http://www.normankoren.com/Tutorials/MTF7.html
http://www.luminous-landscape.com/tutorials/dq.shtml
These guys have a 6MP D60 being about 70% of a 35mm Provia frame, or a 35mm
frame being about 8.6 MP.
(This sounds about right to me: D60 images look very good at 200 dpi (10 x
15"), and making a "critically sharp" 13x17 from 35 is quite possible if one
is careful, but much larger than that is very hard. (Assuming my print
standards, of course. YMMV.))
David J. Littleboy
Tokyo, Japan
Bruce Murphy
> There is far less loss using an optical enlarger as opposed to scanning an
> image.
Can be, perhaps. Some of the scanning techniques are frighteningly good
though. If we're talking minilab-style though, scanning probably wins.
> And the resulting print will be more accurate to the original than
> an ink jet print. Remember, a scan is produced using an optical element,
> too. And there are losses due to the digitizing process.
Yes.
> A digital file is no more useful than a negative. A neg can be printed
> too, you know, or have you forgotten the old processes of photography?
> Inkjet prints cost me the same, or more, than a "wet" print. The paper I
> use for both costs me about $1 per sheet, and ink is more expensive than
> chemicals. There is a savings of time with digital, however.
> Sometimes, I don't know where you get some of your ideas, GG. Think things
> through.
And that's why people use things likt Frontiers to draw digital images
onto wet prints.
B>
Bruce Murphy
> "Bruce Murphy" <pack...@rattus.net> a écrit dans le message de news:
> m2admim...@fuscipes.rattus.net...
>
> > People doing product work for glossy magazines and posters.
>
> 0.01% of the market, you mean?
OKay. I admit that professionals who care about their images are less
common that idiots using disposable cameras. What was your point again?
> > In not having to spend hours developing the
> > image before they're being incorporated into
> > your catalogues.
>
> Developing film requires only 15 minutes. Scanning it requires only two
> minutes.
Wildly optimistic figures. Particulary in the insanely-high-resolution
medium-format area that could realistically compare to scanning backs.
> > Why not?
>
> Because 99% of all photographs are not of perfectly stationary objects, so a
> scanning back won't work. If that were not the case, scanning digicams
> would be commonplace by now.
They are commonplace, just not in the world of 35mm photography. If you're
going to restrict your comments on photography to random party snapshops
(Which are a large proportion of all photographs taken) then perhaps
you should get out of the megapixel pool?
> You'd have it a couple of minutes later.
MF or LF chromes or colour negs? That's an interesting assertion. You're
the one who claims digital is faster.
> > I wonder also with your 'whip the performance'
> > comment if you're aware that scanning backs were
> > producing on the order of several tens of
> > megapixels two years or more ago.
>
> Yes, I'm aware of that. And MF film backs were producing 100 megapixels
> half a century ago.
Oh? In colour? On what film was this? I'm sure you're going to surprise
quite a few people here.
> > Arguable. LF wins a lot by not having to be enlarged
> > much -- a lot of LF optics are significantly more
> > primitive than contemporary 35mm stuff.
>
> LF has more pixels, and that's what matters.
Bollocks. If your lens doesn't resolve anything usefully onto them, then
the number of raw pixels you've got doesn't matter a damn.
> > I also note with amusement that you claim
> > stitching many MF shots allows you to outperform
> > a digital 35-mm class stitch.
>
> I'm comparing apples to apples.
So _stitching_ is okay, even if you've never done it, yet scanning isn't.
Pretty consistent point of view there.
B>
Skip
The average in home photographer has either a flat bed scanner or a film
scanner, but probably not a drum scanner. My flat bed is good for about
1200 dpi, my Epson 880 printer does 2880x720. a reasonably good lens on a
reasonably good enlarger will probably produce a better, more accurate image
than my digitizing set up. Especially since I am scanning from a print,
made from that enlarger.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Bruce Murphy" <pack...@rattus.net> wrote in message
news:m24rcqm...@fuscipes.rattus.net...
Gordon Gekko
> "Gordon Gekko" <gordong...@yahoo.com> a écrit dans le message de news:
> d07edc6d.02091...@posting.google.com...
>
> > Inkjet prints are inexpensive, yet inkjet
> > prints faithfully reproduce all the resolution
> > in the digital file.
>
> exhaust the resolution of the digital file.
I don't know what size you are printing at. I estimate it costs me $1
of ink to print an 8 x 10". This is a bargain compared to how much it
would cost to get a negative enlarged. And then there's the huge
convenience factor of doing it yourself.
And furthermore, there is the ability to CONTROL the process. It's
funny how on this thread people make fun of point and shoots because
the you have no control over anything. But then the control freak
takes his neg to a processor and has no more control over anything.
When you print from digital, YOU countrol the color balance, YOU
control the contrast, YOU control what you want to crop. You wind up
with the perfect picture that you conceived. Once I've experienced
the control of doing things myself, I can never imagine having to deal
with using a lab.
Bruce Murphy
> I guess I should have stated that what I was referring to is how I do it.
> The average in home photographer has either a flat bed scanner or a film
> scanner, but probably not a drum scanner. My flat bed is good for about
> 1200 dpi, my Epson 880 printer does 2880x720. a reasonably good lens on a
> reasonably good enlarger will probably produce a better, more accurate image
> than my digitizing set up. Especially since I am scanning from a print,
> made from that enlarger.
Ah well, if we comound foibles it certainly will get worse. I think
you can do quite a bit better using a good film scanner
though. Certainly than anything you'll see coming out of a minilab.
B>
brian
You seem to have missed the point entirely and are now going off on
your own tangent. The question to be answered is: "how many pixels
are in a film frame". A *single* film frame. If you think that
medium format is good for 80 megapixels then you should show an 80
megapixel image so that we can examine it. The technique I used to
achieve my 10MP and 53MP images is irrelevant here. The fact that
these images are large, and clean and available *is* relevant. I'm
providing resolution reference images here, nothing more.
I would say that none of your medium format images actually contains
53MP worth of real information (besides grain, of course). To show
this you need to create a 53MP image from one of your medium format
negatives or slides, and then compare your image to my image at 100%
magnification or higher. This should not be a problem for you because
all you would have to do is to downsample from one of your 80
megapixel files. This seems very simple and intuitive to me.
If you can't manage 53 or 80 megapixels, then just try for a true 10
megapixel image. With modest JPEG compression you should be able to
get a 10 megapixel file down to about 2 megabytes or less. The image
you have shown is nice, but it is only 1321x1329 pixels, or less than
2 megapixels. You need to do much better than this to make an
interesting comparison.
Gordon Moat
Thanks for the great response.
brian wrote:
> . . . I'm not trying to convince everyone to go out and stitch bunches of
> images together just to get more resolution than you can possibly see
> on an 8x10" or 11x14" print. I'll grant you that such a technique
> will probably appeal to only a few crazed lunatics like myself.
> Nevertheless, by putting together 42 6-megapixel images and then
> *downsampling* to 53megapixels I wind up with a pretty decent picture
> that can be used as a resolution reference of sorts. The 10megapixel
> version in particular has essentially no trace of digital artifacts
> aside from minor stuff caused by a level-10 JPEG compression.
Actually, one of the things I found the most impressive was the Kaidan Spherical Tripod head.
>
>
> More comments below:
>
>
> >
> > It appears that your image is created from several digital acquisitions and then stitched. You
> > have 42 such captures shown. If I am wrong here, please correct me, but it would seem that even
> > accounting for some frame loss for overlap, there are at least 20 equivalent full (CCD) frames for
> > the resulting image. I also wonder why you did JPEG, since it is a lossy compression; why not use
> > TIFF or raw format?
>
> You're estimate of about 20 equivalent frames is pretty accurate. I
> save all of my original stitched images as TIFFs. I only JPEGed them
> to make it possible to host a true 53MP image. I've got to *PAY* for
> that server space, you know!!
Yeah, and a shame at that. I only have 40 MB available at the moment, though there is no bandwidth
limit.
>
>
> In an ideal world, some kind person would download my image, print it
> at 48" x 96", and then do the same with images shot with 35mm, 6x4.5,
> 6x9, 4x5, 8x10, 11x14, and 20x24. However, I would welcome any sort
> of thoughtful comparison based on actual images.
48" by 96" would drop the resolution a bit. It would also limit your choices on printing locations, and
make for a higher cost. A 20" by 30" would be much easier, and have better edge definition. It would
also be an easier size to match with other formats. I think it would still show a valid comparison.
Maybe one of these days . . . .
>
>
> >
> > Kodak Pro Photo CD has an upper limit of 4096 by 6144. While this is close to your 5761 by 9194
> > image, it would only give a 13.5" by 20.5" image at 300 dpi. Obviously a cheap alternative does
> > not get too close to what you have. Interestingly, cost is another issue. Locally, I can get a
> > Scitex drum scan up to 650 MB file size for around $100 from 4" by 5" film, or medium format. Not
> > really too much, unless I have many scans to complete. I have worked on up to 2 GB image files on
> > my computer gear, though the computer gets noticeably slower when doing these. My normal working
> > scans are 70 MB to 160 MB in file size as three channel files.
> >
>
> I would *love* to see a comparison image from drum-scanned 4x5. I may
> even do this myself on my next trip to NYC.
Drum scanned 4" by 5" is truly amazing. The first one I ever saw was used for a demo of PhotoShop and
Alias Studio running on an SGI O2. The only bad part about these is that you need a fairly large amount
of RAM and scratch space to handle an image this large. Zooming in anywhere is really amazing. One
could almost get lost in the image.
>
> I've already admitted to being a crazed lunatic. Doing cost-benefit
> analyses is not something I would ever do when it comes to
> photography.
>
Yeah . . . unfortunately, that Imacon would look really nice on my desk at my office . . . .
>
> >
> > I wish I was in New York, so I could take your challenge . . . not to prove you wrong, but to show
>
> I also wish I was in New York.
I am looking into MFA programs currently. Some of these are near NYC, so who knows . . . maybe I will
make it there by next year.
>
> Remember, I'm not really trying to compare stitched images to
> individual ones: I'm just providing clean high-resolution images that
> can be used as a sanity check on statements like "35mm Tech Pan =
> 354Megapixels!" The only reason(s) I don't stitch scanned film images
> are that I already have a nice digital camera that gives me images
> about as good as my 2700 dpi slide scanner, I *hate* spending my time
> scanning and organizing, and I don't want to spend money on film and
> processing.
Actually, this is a really great comment. I have recently figured out a technique to greatly exceed the
existing dynamic range of a film scanner. All it takes is two scans, and some tricks in PhotoShop. Your
stitching technique allows smaller frames to digitally reach the capabilities of larger formats (except
colour tonality. . . though closer). I charge for my scanning and adjusting, so for me it is not
troublesome. Drum scans are nice, PhotoCD is okay, but usually I stick to the film scanner at my main
desk. Extending its' capabilities makes it even more convenient. I think for many people, finding out
ways to get more out of your gear can be a better alternative than acquiring other gear.
>
>
> Also bear in mind that my current limitations on resolution are
> computer-based and not camera-based.
Very true . . . and printing is the other limitation. The world of printed publications evolves slowly.
We are still at 300 resolution and 150 to 175 lpi, with very few places doing stochastic screening.
Even worse, many places are still strictly CMYK, with few Hexachrome, or HiFi options. Plateless
printing is just finally catching on, after more than seven years on the market. Because of this, the
path to upgrades can be stretched much farther. I look at the least common denominator of technology,
and try to match it. It seems much easier than being on the bleeding edge.
Ciao!
Gordon Moat
Alliance Graphique Studio
John Eyles
A bit of a red herring - because the enlarger optics must image
the ENTIRE negative at once, whereas the scanner optics need only
image a very small portion of the negative at one time.
John
Robert Monaghan
There is far less loss using an optical enlarger as opposed to scanning
an image.
This is not my understanding; some scanners use technology like lasers
and photomultiplier tubes which avoid some of the problems and losses
associated with intermediate optical lenses (same problems with chemical
based print enlarger lenses etc.). The high end digitizing processes can
produce amazingly large data and color depth info from film IMHO...
If you mean most current pro/consumer flatbed scanners have lousy cheapy
optics, that is true in many cases ;-)
The losses from a very good enlarger lens can be a relatively small part
of the overall system losses, and many pro and lab enlargers have very
good lenses (now operators, maybe not ;-)...
grins bobm
--
* Robert Monaghan POB752182 So. Methodist Univ., Dallas Tx 75275 *
* Third Party 35mm Lenses: http://medfmt.8k.com/third/index.html *
* Medium Format Cameras: http://medfmt.8k.com/mf/index.html *
Robert Monaghan
some 35mm lenses (pentax 50mm, leica..) are rated with aerial resolutions
up to 650+ lpmm and beyond (per Skudrzyk's tests); Some Leica types are
reporting work with B&W films of over 900 lpmm resolution limits, which
should push system resolution much closer to the lens aerial resolution
limits (cf Puts posting at http://medfmt.8k.com/mf/film.html ) Even if the
film is only 650 lpmm (as is the lens), then 325 lpmm system resolution
would be expected - that's a lot of data.
Can all those leica users claiming large format results from 35mm with
tech pan be right? ;-)
but I'll go along with a study published by AFIP scientist:
Consumer Digital Camera 0.365 megapixels
Amateur 1kX1k " " 1.05 megapixels
Pro 4kX4x " " 16.77 megapixels
35mm fast film 22.11 megapixel equiv.
35mm medium speed film 54 megapixel equiv.
35mm slow speed film 124.76 megapixel equiv.
=====
at this point, since MF 645 is 2.9 times area of 35mm that's 364 Mp equiv,
for 6x6, that's 3.6X and 453 MP, and 6x7 at 4.5 is 566 MP etc. And since
color density can be 3 or more bytes per pixel, that's a lot of megabytes
of data! ;-)
In any case, there are lots of folks getting 4000 ppi and even 5000+ ppi
scans, and with colour depth that translates into huge one image per CD
files - presumably paying for this pricey service for image data that's
there on film ;-0) John Stafford's project with a 135mb image shot scan
is one example, there are presumbably lots of other large scans out there.
Brian can simply shoot a second shot with a MF camera of the same scene,
pay $15-100 to get it scanned, and make the comparison if he wishes. It
will confirm that there is a lot more data in a single medium format,
or 35mm - image than many digeratti are willing to admit ;-)
Annika1980
>> How often do you view wall-sized enlargements
>> from just inches away?
>
>By standing inches in front of them.
I know how. The question was "How often?"
Charlie Dilks
already read some of them regarding "cult lenses." Now I realize there's
=much= more there.
Here's one quote from you that pretty well sums up my point:
"By technique, I mean things like using a lens hood or tripod are much more
likely to have a positive impact on your final image than some minor
resolution differences based on a high contrast lens chart that your film
can't record anyway. We have a page of tips on how to get the most out of
your lens and camera. I recently posted a page on mirror lockup benefits
where using MLU was shown to improve resolution by up to 172% and over 50%
on average at slower shutter speeds. Compare that to the 10-25% difference
in lens resolution between pricey OEM and lower cost third party lenses. If
you are shooting hand-held with typical color print film, your technique is
going to be much more limiting than your lens resolution!"
The question is, "How much resolution will the average, or even extremely
careful shooter get in real life.
I always think back on tests I did on my fantastic original Tamron 90mm
f/2.5 SP macro. I shot the test targets at the prescribed distances with the
prescribed lighting on a tripod and used mirror lockup on my Olympus OM-1.
At the best aperture in the center of the field I got lp/mm in the mid to
high 90s. I was using Tech Pan at, I believe ASA 80, developed in some
"magic" developer of the time (XR-1?).
Later on I ran the test with everything the same except that I illuminated
the cards with a flash. I then got best lp/mm readings at up to 120.
It seems that vibration from the shutter release cable, shutter slap, or
natural vibrations of my house degraded the non-flash tests.
--
Charlie Dilks Newark, DE USA
Skip
its job than what ever is in most flat bed scanners.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"John Eyles" <j...@cs.unc.edu> wrote in message
news:am5c2b$34i$1...@capefear.cs.unc.edu...
Skip
assume that he doesn't have a high end drum scanner. I have seen the
difference that a good enlarger lens can make.
--
Skip Middleton
http://www.shadowcatcherimagery.com
"Robert Monaghan" <rmon...@smu.edu> wrote in message
news:am5f1e$tdi$1...@post.cis.smu.edu...
Mxsmanic
%qkh9.25174$S32.1...@news2.west.cox.net...
> There is far less loss using an optical enlarger
> as opposed to scanning an image.
In an ideal universe, yes. But in practice, scanning always produces a
better result.
To get good results with an enlarger, you must be able to image the _entire_
frame at once. But to get good results with a scanner, you only need to
image a pixel at a time. So it is far easier to get good results by
scanning than it is by enlarging, and overall, scanners do much better than
enlargers.
The labs I use, including the pro labs, scan exclusively now. They don't
have any enlargers.
> Remember, a scan is produced using an optical element,
> too.
But it is an optical element optmized for imaging only one pixel (or one
line of pixels) at a time. And that is thousands of times easier to
engineer than an optical system for imaging an entire frame at once.
> And there are losses due to the digitizing process.
No more so than with enlarging.
Mxsmanic
> Doesn't matter, believe me, a Nikkor enlarger
> lens is as good or better at its job than what
> ever is in most flat bed scanners.
But not as good as the lenses that Nikon puts in its film scanners.
Mxsmanic
> I don't know what size you are printing at.
8x10
Each print costs me $4-$6 for paper and ink. The local photo lab can put
the same photo on an 8x12 sheet of real photo paper for just $4, and it
looks a lot better than the inkjet. As a result, nowadays I'm more likely
to give my digital file to the lab than to try to print it myself.
> I estimate it costs me $1 of ink to print an
> 8 x 10". This is a bargain compared to how
> much it would cost to get a negative enlarged.
Are you sure? I've seen 20x30-inch chemical prints from scans for $22, and
they leave inkjet very much in the shade. The Durst Lambda has changed a
lot of things.
> And then there's the huge convenience factor
> of doing it yourself.
It takes longer to print it yourself than to get it printed at a lab. A
nice A3 print takes over an hour on my printer.
> And furthermore, there is the ability to
> CONTROL the process.
Except that printers are very difficult to control. What you see on the
screen is one thing, what the printer produces is quite another.
In contrast, digital minilabs can often print exactly what you saw on the
screen (allowing for gamut). I have many examples proving this.
> But then the control freak takes his neg to
> a processor and has no more control over anything.
Who takes negatives to a lab? I scan myself and take them the finished
digital file, and I tell them to print it exactly as-is. And I get the same
thing on paper that I saw on my screen. It's great!
Mxsmanic
c4i5qte38i16l...@stereo.hq.phicoh.net...
> Please point to an online image that contains
> 200 Mpixels and comes from a single frame of
> 35mm film.
I don't know of any, and my own scanners don't scan at high-enough
resolution to extract that much detail. I do have Tech Pan negatives that
contain that much detail, however (I've examined them under a microscope).
Gordon Gekko
> A digital file is no more useful than a negative. A neg can be printed
> too, you know, or have you forgotten the old processes of photography?
> Inkjet prints cost me the same, or more, than a "wet" print. The paper I
> use for both costs me about $1 per sheet, and ink is more expensive than
> chemicals. There is a savings of time with digital, however.
> Sometimes, I don't know where you get some of your ideas, GG. Think things
> through.
Maybe you have a color lab in your house. I sure don't. A color lab
would cost a lot more to set up than a top of the line inkjet printer.
Digital printing beats chemical printing for the average person.
Mxsmanic
am4om4$dcj$1...@nnrp.gol.com...
> That seems extremely overly optimistic.
It comes right from the specs for the film.
> Since a 35mm frame has 864 sq mm, that's
> 231,481 pixels/sq mm, or 481 x 481 pixels
> each mm. So you'd need a lens that delivers
> in excess of 240 lp/mm over the whole frame
> to create a (meaningful) 200 MP image. Since
> there's no such thing as a lens that does that
> (heck, there aren't any lenses that deliver
> 120 lp/mm over the whole frame), claiming 200
> MP is quite excessive, silly, ridiculous ...
If no lens can do that for film, then no lens can do that for digital,
either.
> If you restrict your images to images that
> appear in nature, you're talking about not 1:1000
> contrast, but 1:1.6 contrast, at which point,
> even Tech Pan isn't a whole lot of megapixels.
According to the specs, which I have right in front of me, Tech Pan delivers
a minimum of 7200x4800 pixels on 135 film, at 50% modulation transfer.
That's 34.56 megapixels.
Mxsmanic
> Wildly optimistic figures.
My local lab can do C-41 or E-6 in 15-20 minutes. A 2700-dpi scan of a 35mm
frame requires about 2 minutes with my scanner.
> Particulary in the insanely-high-resolution
> medium-format area that could realistically
> compare to scanning backs.
It's a lot more practical to scan the image after it has been frozen on film
than to expect your subject to remain immobile while you scan it directly
from real life.
> They are commonplace, just not in the world of
> 35mm photography.
And most photography is 35mm photography. A scanning image capture is
useless for that type of photography.
> MF or LF chromes or colour negs? That's an
> interesting assertion. You're the one who claims
> digital is faster.
It is. You can have one-shot digital images on a file within seconds. But
you can't do that with a scanning back.
> Oh? In colour?
I don't know if color films were available in 120 fifty years ago.
Kodachrome is older than that, of course, but I believe it has always been
available in 135 only.
> Bollocks. If your lens doesn't resolve anything
> usefully onto them, then the number of raw pixels
> you've got doesn't matter a damn.
Oh, I think LF lenses resolve pretty well. I've seen the results. It's
best not to try to compete with LF for pixels.
> So _stitching_ is okay, even if you've never done
> it, yet scanning isn't.
Stitching is fine as long as it's compared to other stitching.
Mxsmanic
3c459ba.02091...@posting.google.com...
> The question to be answered is: "how many pixels
> are in a film frame". A *single* film frame.
About 100 megapixels for 6x6, and about 25 megapixels for 35mm, on Provia
100F.
> If you think that medium format is good for 80
> megapixels then you should show an 80 megapixel
> image so that we can examine it.
It would be a 470 MB download.
> The technique I used to achieve my 10MP and
> 53MP images is irrelevant here.
Sorry, but when you take multiple pictures, stitch them together, and then
try to compare the resulting composite image with a _single_ film frame,
there is a problem.
> I would say that none of your medium format
> images actually contains 53MP worth of real
> information (besides grain, of course).
Then contain more than that.
> To show this you need to create a 53MP image
> from one of your medium format negatives or slides,
> and then compare your image to my image at 100%
> magnification or higher.
I have, and film holds its own, even at 4000 dpi, and even in less than
optimal shots.
> This should not be a problem for you because
> all you would have to do is to downsample from
> one of your 80 megapixel files.
Correct, and it's not a problem for me. I've done it several times out of
curiosity. Digital is no better.
And the important thing that you are overlooking here is that you're
cheating in your digital example. You had to make 42 separate exposures to
get that 53 mp image; I used one exposure to get an 80 mp image. If I
follow your example, and make 42 separate MF exposures, and stitch them into
one giant image, and reduce it to 53 mp to match your stitched digital
image, the digital image is completely, totally blown away.
An easy way to make this comparison is to divide the stitched digital
resolution by 42, reduce the MF shot to that same size, and compare. The
film version is superior, even to the casual observer.
> The image you have shown is nice, but it is only
> 1321x1329 pixels, or less than 2 megapixels.
I know. I was following your example. Since each of your digital images is
only 1/42 of the 53 megapixel whole, I reduced my MF image to 1/42 of it's
normal size, giving the image quality that would result if I had built the
original image from 42 MF shots. And the result is vastly superior to your
digital version.
> You need to do much better than this to make an
> interesting comparison.
Those who understand what I'm doing will already see the significance of
this comparison. Those who don't, won't. I can only explain it so many
different ways.
Mxsmanic
20020916173537...@mb-cg.aol.com...
> I know how. The question was "How often?"
As often as I see wall-sized enlargements. In fact, I was doing that last
night.
David Eppstein
"Mxsmanic" <mxsm...@hotmail.com> wrote:
> > If you think that medium format is good for 80
> > megapixels then you should show an 80 megapixel
> > image so that we can examine it.
>
> It would be a 470 MB download.
So cut out a representative 400x400 pixel patch from the image, show us
that, and tell us how big the original image was that you cut it out of.
--
David Eppstein UC Irvine Dept. of Information & Computer Science
epps...@ics.uci.edu http://www.ics.uci.edu/~eppstein/
Annika1980
>> The image you have shown is nice, but it is only
>> 1321x1329 pixels, or less than 2 megapixels.
>
>I know. I was following your example. Since each of your digital images is
>only 1/42 of the 53 megapixel whole, I reduced my MF image to 1/42 of it's
>normal size, giving the image quality that would result if I had built the
>original image from 42 MF shots. And the result is vastly superior to your
>digital version.
I disagree. Your example didn't impress me at all. Now maybe if you had
simply cropped a 1321x1329 section from your original, that might've shown more
detail. As it is, I can't really tell anything about the image, except that
it's a boring pic.
Annika1980
>> I know how. The question was "How often?"
>
>As often as I see wall-sized enlargements. In fact, I was doing that last
>night.
>
We really don't wanna hear about your lifesize porn pics and what you do with
them.
Bruce Murphy
> "Bruce Murphy" <pack...@rattus.net> a écrit dans le message de news:
> m23csam...@fuscipes.rattus.net...
>
> > Wildly optimistic figures.
>
> My local lab can do C-41 or E-6 in 15-20 minutes. A 2700-dpi scan of a 35mm
> frame requires about 2 minutes with my scanner.
I guess that most product photographers don't feel like having a lab like
that sitting in their studio. Even if they could justify the costs of
setting it up, the wasted chem would kill them.
> > Particulary in the insanely-high-resolution
> > medium-format area that could realistically
> > compare to scanning backs.
>
> It's a lot more practical to scan the image after it has been frozen on film
> than to expect your subject to remain immobile while you scan it directly
> from real life.
Not for product photography. Were you *listening* when I explained
precisely where these things were used?
> And most photography is 35mm photography. A scanning image capture is
> useless for that type of photography.
Most photgraphy is 35mm snapshots with point and shoot cameras or
disposables. What exactly is the point about talking about 'most'
photography.
> > MF or LF chromes or colour negs? That's an
> > interesting assertion. You're the one who claims
> > digital is faster.
>
> It is. You can have one-shot digital images on a file within seconds. But
> you can't do that with a scanning back.
And you couldn't (three years ago) put a 30+ megapixel image into a digital
file within seconds (And in fact still can't) without using a scanning
back. What was that quaint little phrase you used? 'apples to apples'
> > Bollocks. If your lens doesn't resolve anything
> > usefully onto them, then the number of raw pixels
> > you've got doesn't matter a damn.
>
> Oh, I think LF lenses resolve pretty well. I've seen the results. It's
> best not to try to compete with LF for pixels.
Actually they don't resolve all that well at all, what you're
forgetting is the huge multiplcation factor in film size.
> > So _stitching_ is okay, even if you've never done
> > it, yet scanning isn't.
>
> Stitching is fine as long as it's compared to other stitching.
Your assertion was that stitching is evil and bad and impractical
because noone can use a camera that takes 5 minutes to get a single
image. Now I've introduced scanning backs, all of a sudden you're
defending stitching and saying that noone actually uses scanning backs
becuase they're less convenient than a 35mm point and shoot.
Unfortunately, people *do* use them, thus demonstrating that your
original assertion about tradeoffs between image quality and speed of
acquisition was not correct.
I imagine that you're going to continue arguing for the sake of
arguing, though.
B>