Pages: 1 [2] 3 4 ... 7   Go Down

Author Topic: This Test was completely flawed  (Read 94697 times)

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
This Test was completely flawed
« Reply #20 on: May 28, 2006, 10:06:03 am »

Re print performance:  Just for the record, I recently salvaged a frame from a Cibachrome that I had made perhaps 15 years ago.  The print had been mounted with archival materials and displayed in typical room light with no direct sun for approximately 10 years, then pulled down and stowed for eventual recycling of the frame.  It looked fine after those 15 years -- until I removed it from the frame and matte.  I could very clearly see where the portion under the matte retained deeper, more accurate color.  

So in at least one example, the 50-year life of Cibachrome is a myth...
Logged
Jack
[url=http://forum.getdpi.com/forum/

neal shields

  • Newbie
  • *
  • Offline Offline
  • Posts: 4
This Test was completely flawed
« Reply #21 on: May 29, 2006, 10:53:58 am »

No one seemed much impressed with my FBI.gov link:

http://www.fbi.gov/hq/lab/fsc/backissu/apr...swgitfield1.htm

So how about Zeiss.com?

http://www.zeiss.com/C12567A8003B58B9/Cont...1256F2C002B7DBB

They are resolving 170 lp/mm with Velva, and up to 400 with specialized B&W film.

Logicaly it takes at least two rows of pixels to resolve a line pair. (most people say three)  One for the black line and one for the white space.

That means that it takes an inch tall piece of film would need at least a scan of 340 times 24 to resolve all the detail, or about at least 8000 dpi.  That just happens to be the number other testers have found empirically.  So why would anyone "test" film by scanning it at 3200 dpi?
Logged

collum

  • Full Member
  • ***
  • Offline Offline
  • Posts: 189
    • http://www.jcollum.com
This Test was completely flawed
« Reply #22 on: May 29, 2006, 11:39:13 am »

Quote
No one seemed much impressed with my FBI.gov link:


They are resolving 170 lp/mm with Velva, and up to 400 with specialized B&W film.

Logicaly it takes at least two rows of pixels to resolve a line pair. (most people say three)  One for the black line and one for the white space.

That means that it takes an inch tall piece of film would need at least a scan of 340 times 24 to resolve all the detail, or about at least 8000 dpi.  That just happens to be the number other testers have found empirically.  So why would anyone "test" film by scanning it at 3200 dpi?
[a href=\"index.php?act=findpost&pid=66848\"][{POST_SNAPBACK}][/a]

i am impressed that you've beeen able to test film/lens combinations and demonstrate a 170 lp/mm with Velvia. Which lens  did you use (4x5 specific, since that's the film we're talking about)?
« Last Edit: May 29, 2006, 12:25:45 pm by collum »
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
This Test was completely flawed
« Reply #23 on: May 29, 2006, 03:08:18 pm »

Quote
No one seemed much impressed with my FBI.gov link:

http://www.fbi.gov/hq/lab/fsc/backissu/apr...swgitfield1.htm

So how about Zeiss.com?

http://www.zeiss.com/C12567A8003B58B9/Cont...1256F2C002B7DBB

They are resolving 170 lp/mm with Velva, and up to 400 with specialized B&W film.
[a href=\"index.php?act=findpost&pid=66848\"][{POST_SNAPBACK}][/a]
These sources, Zeiss at least, seem to be about "extinction resolution" with extreme high contrast subject matter like resolution test charts. This might be relvant to th FBI when reproducing high contrast black and white text documents, but it is is all in the realm of MTF levels below 10%, which is of no relevance to normal photographic needs. Instead for most photography it is more interesting to look at how fine the details can be and still be resolved with about 50% MTF or better.

When you change to more relevant standards like 50% MTF, film resolution numbers fall much lower, as film has a long slow decline of MTF as lp/mm increases, whereas digital sensors can hold 50% MTF or better almost up to the limit where resolution fails entirely. (The spec sheet for one of the the new 30MP plus higher Dalsa sensors reports an excellent MTF of 70% all the way to the Nyquist limit on one line pair for each two pixels, but than is for a monochrome sensor without Bayer color filter array.)

An example: that Zeiss link reports Velvia resolution of 170lp/mm, and Fuji reports 160lp/mm with an exteme high contrast 1000:1 test pattern. But Fuji also publishes MTF curves for Velvia which show that MTF is already below 50% by 50lp/mm, and those curves for Velvia do not even go beyond about 60lp/mm.

The drop from 160 or 170 lp/mm to below 50 lp/mm reduces the "pixel count equivalent" by a factor of about (170/50)^2 =11.6. That is, knock one zero of some of those those wildly optimistic pixel count equivalents.


P. S. The FBI source merely asserts a range of 40-160 lp/mm wit no details on measuremtn procedre of definitions. It then states that
"Color films used at crime scenes have resolutions at the lower end of this range"
and that
"A single frame of 35 mm ISO 200 color film is 36 mm wide by 24 mm high. With a resolution of 50 line pairs per millimeter, such a frame can resolve ... 8,640,000 pixels."

This fits fairly well with other sources sugesting that Velvia in 35mm format matches about the 8MP of an Olympus E-500 or Canon 350D and probably falls a bit short of the Sony R1 as a choice for high resolution in crime sceme photography.
« Last Edit: May 29, 2006, 03:23:12 pm by BJL »
Logged

Gary Ferguson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 550
    • http://
This Test was completely flawed
« Reply #24 on: May 29, 2006, 07:27:24 pm »

Quote
No one seemed much impressed with my FBI.gov link:
http://www.fbi.gov/hq/lab/fsc/backissu/apr...swgitfield1.htm
So how about Zeiss.com?
http://www.zeiss.com/C12567A8003B58B9/Cont...1256F2C002B7DBB
They are resolving 170 lp/mm with Velva, and up to 400 with specialized B&W film.

I'm intrigued by this one, and as a long standing Zeiss/Hasselblad user I'd love for it to be true! Especially as Zeiss also claim their medium format lenses are every bit as good as their 35mm lenses. Unfortunately I have to put sentiment on one side and be a bit more realistic.

First, Zeiss of late seem to be stretching the credibility envelope. I'll give you just one example. A year or so ago one of Zeiss's suplliers of optical glass announced they'd no longer produce a particular glass containing arsenic and lead. This formulation was only available from them and was critical for the 903 38mm Biogon, which is one of the unique, bedrock products in the Zeiss/Hasselblad range. So they produced a new design, the 905, and trumpeted it as a wonderful advance that, by the way, will also save the planet as well as taking your photography to the next level. Problem is, when you look at the MTF charts for the 903 and 905 side by side you soon reach the conclusion that all that lead and arsenic was in there for a good reason! The moral of the story is to treat Zeiss's pronouncements with a king-sized grain of salt.

Next, these 400 lppm claims met a scathing response from Erwin Puts, the Leica guru. He basically agreed with BJL's conclusions that in real world photography such extreme claims are disingenuous, and for practical purposes photography pretty much hits a wall at about 80-100 lppm. Furthermore, he said even this is only achievable with the most stringent technique and with a few exceptional lenses, for the most part 60-80 lppm is a demanding enough limit.

This sounds like realistic advice born from genuine experience.

I look for medium format lenses that give 40 lppm at 40% MTF out to about 31mm from the centre (appropriate for the 37mm x 49mm digital sensor I use), by choosing selectively I've assembled a reasonable selection of Hasselblad lenses that deliver against this objective. And I've always found this more than adequate for my own photography, allowing crisp enlargements of x10 or occasionally even larger. What is the current insanity that drives photographers to expect anything more?
Logged

michael

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5084
This Test was completely flawed
« Reply #25 on: May 29, 2006, 08:15:30 pm »

Quote
What is the current insanity that drives photographers to expect anything more?
[a href=\"index.php?act=findpost&pid=66879\"][{POST_SNAPBACK}][/a]
A cynic would say, because...

A: They don't understand the science.

or

B: They don't have much real-world experience with these tools.

Michael
Logged

Dave Millier

  • Full Member
  • ***
  • Offline Offline
  • Posts: 146
    • Whispering Cat Photography
This Test was completely flawed
« Reply #26 on: May 30, 2006, 05:40:20 am »

Quote
I reviewed the Web Article from the four Photographers regarding their experiments with different Digital Cameras / Lenses and 4x5 Inch Film Sheet.  Their undertaking had a major fatal flaw.  They didn't produce Optical Prints from the Film.  Instead, they just scanned the Film Sheet into their Computer which means that the original Resolution (86 Trillion) and Colour of the Film is lost.  Film has a Resolution of 6.9 Billion  Molecules of Dye per Square Millimetre, but this is only retained if the Picture Print is made using real Light -- not a Computer Scan.  The Attached File explains the inferiorities of Digital Photography.
[a href=\"index.php?act=findpost&pid=66700\"][{POST_SNAPBACK}][/a]

I note the self-published "theory" and the heavy use of capitialisation in the poster's  writing. Add in the determined iconoclasm, the superior position of knowledge, the apparent learnedness and the resoluteness....

These are often classic symptoms of the species well known as "The Crank".  

I may be completely wrong in which case I apologise but I wouldn't be surprised if he also has unusual opinions on zero-point energy and why Einstein was wrong....

The best example I have of this kind of thing is a massive publication by Michael Pinder called "Time on our Hands - Global Philosophy for the New Age".  This is a splendid self published tome, many copies of which have been sent unsolicited to UK Civil Servants, which espouses the case for the adoption of a decimalised time system.

Not necesarily a crazy idea - afterall in Europe at least we are embracing the decimalisation of most measuring systems - but what really sets it apart is how it claims decimal time will prove to be the cure for just about all economic, political, medical and scientific problems.

The publication is highly literate, thoroughly researched and quite entertaining. Quite  barking, of course.  

There seem to be quite a lot of these sorts of people about - well meaning, determined, obessive. The more intelligent and/or better educated they are the more trouble they cause because it takes a bit more effort to counter their arguments.

The recent introduction of FOI legislation in the UK has given these people a field day as they can demand their legal rights to endless replies to requests for information.

My office recently spent over £50,000 of tax payers' money responding to one crank's demands for information and a public enquiry into a supposed political cover up...
Logged
My website and photo galleries: http://w

Dennis

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 82
This Test was completely flawed
« Reply #27 on: May 30, 2006, 08:07:47 am »

Quote
They are resolving 170 lp/mm with Velva, and up to 400 with specialized B&W film.
Don't confuse film resolution with resolution in a photograph. Be aware, that there is an diffraction limit to resolution, as well. E.g. at f/5.6, this limit is at 246 lp/mm. If you want to resolve 400 lp/mm, you'd need using a f-stop around 1:3. But there, you'll get some trouble with the optical performance of the lens (there's some glass inside, you remeber?). BTW: Film resolution is not determind by photographing some test patterns with a camera and a lens, it's more a process of a contact copy.

Quote
Logicaly it takes at least two rows of pixels to resolve a line pair. (most people say three)  One for the black line and one for the white space.
No, you need four rows of pixel to resolve one pair of lines. Ever heard of Nyquist theorem?

Quote
I note the self-published "theory" and the heavy use of capitialisation in the poster's  writing. Add in the determined iconoclasm, the superior position of knowledge, the apparent learnedness and the resoluteness....

These are often classic symptoms of the species well known as "The Crank".
 No comment on this.

But Michael, you should consider going back to film. As we now know:

In truth, the only type of person who actually needs a Digital Camera is someone who needs to be able to instantly E-Mail a picture (such as a newspaper journalist) or someone taking temporary pictures they don't wish to keep or print out.
(see Terry's linked article)

 

Quote
It is so outrageosly dumb that it would appear to someone knowledable as a joke, but regretably I believe that this person actually believes what he writes.
You just knew it, hm?
Logged
Best Regards

Dennis.

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
This Test was completely flawed
« Reply #28 on: May 30, 2006, 10:21:29 am »

Quote
No, you need four rows of pixel to resolve one pair of lines. Ever heard of Nyquist theorem?
[a href=\"index.php?act=findpost&pid=66894\"][{POST_SNAPBACK}][/a]
Dennis, I agree with most of your post, but as the resident professional mathematician here (who actually knows the proof the Nyquist's theorem, not just its statement) I have to disagree on this one.

Nyquists's theorem says that it takes at least two samples (two pixels) to resolve one cycle, meaning one period of variation. A "dark/light" line pair is a cycle, which is why line pairs rather than lines are traditionally used in describing resolution. Some MTF graphs are labelled in "cycles per mm" instead of lp/mm.

So, Nyquist demands at least two pixels per line pair, or one pixels per line.

For example, the 7.2 micron pixel spacing of Dalsa's new medium format FF CCD sensors, two pixels span 14.4 microns, and lie pairs 14.4 microns wide give a Nyquist frequency of gives 1000/14.4lp/mm, about 70lp/mm. Indeed Dalsa's data sheet for the 33MP monochrome FTF5066M uses this in the resolution specification:
"Resolution (MTF) @ 70lp/mm: minimum 65%".
« Last Edit: May 30, 2006, 10:22:18 am by BJL »
Logged

Graeme Nattress

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 584
    • http://www.nattress.com
This Test was completely flawed
« Reply #29 on: May 30, 2006, 11:36:29 am »

Yes, only a single pair of pixels needed to resolve one line pair. You generally need steeper anti-alias filtering for bayer pattern sensors though, and with a good demosaicing algorithm reckon on about >70% of the actual pixel resolution as RGB equivalent.
Logged

barryfitzgerald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 688
This Test was completely flawed
« Reply #30 on: May 30, 2006, 11:40:04 am »

The argument in itself is pointless anyway. Why do people need to measure things all the time? It never ceases to amaze me...

Did you hear bresson moaning about film quality?

I cannot agree with all the original poster says, but I will say I think the scanning of film for the test makes it somewhat flawed. Scanning adds noise to the image. I dont know anyone who scans film and then prints it at huge sizes.

I use, and will continue to use film and digital. Nobody denies how handy digital is, but neither can you say that film has no use either. I have yet to see digital match the lattitude and character of b&w film, one reason its still popular.

Those who moan about "grain or Noise" would do well to see that certain styles of photography are enhanced with grain...A large number of b&w photographers use high grain films, they like it. So do I sometimes, and sometimes I like low grain. Its a personal choice.

The real issue is that people in general would do far better to worry about their skills behind the camera, than if the "quality" is good enough, and conduct a series of mildly interesting, but ultimately unimportant tests...so as to "hang out to dry" the loser as such. Technology always moves on, and will continue to do so..but that doesnt stop you taking great shots either.

I know of a keen pinhole camera photographer, and he is one of the most gifted people I know of...he knows his shots are not as sharp, or wont print as big..but it doesnt matter. The real meat is the image itself, it matters not how you got there, film or digital..

Regards
« Last Edit: May 30, 2006, 11:41:52 am by barryfitzgerald »
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
This Test was completely flawed
« Reply #31 on: May 30, 2006, 01:01:59 pm »

Quote
Dennis, I agree with most of your post, but as the resident professional mathematician here (who actually knows the proof the Nyquist's theorem, not just its statement) I have to disagree on this one.

Nyquists's theorem says that it takes at least two samples (two pixels) to resolve one cycle, meaning one period of variation. A "dark/light" line pair is a cycle, which is why line pairs rather than lines are traditionally used in describing resolution. Some MTF graphs are labelled in "cycles per mm" instead of lp/mm.

So, Nyquist demands at least two pixels per line pair, or one pixels per line.

For example, the 7.2 micron pixel spacing of Dalsa's new medium format FF CCD sensors, two pixels span 14.4 microns, and lie pairs 14.4 microns wide give a Nyquist frequency of gives 1000/14.4lp/mm, about 70lp/mm. Indeed Dalsa's data sheet for the 33MP monochrome FTF5066M uses this in the resolution specification:
"Resolution (MTF) @ 70lp/mm: minimum 65%".
[{POST_SNAPBACK}][/a]

Four pixels per LP may be a but much, but 2 pixels per cycle will work only if the sampling is done in phase: i.e. the lines on the chart must line up with the rows of pixels. On his web site Roger Clark recommends 3 pixels per cycle. This is consistent  with your statement of at least 2 pixels per cycle.


[a href=\"http://www.clarkvision.com/imagedetail/sampling1.html]http://www.clarkvision.com/imagedetail/sampling1.html[/url]
Logged

Graeme Nattress

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 584
    • http://www.nattress.com
This Test was completely flawed
« Reply #32 on: May 30, 2006, 01:47:01 pm »

The Clark link doesn't look correct to me at all.

He's using a square wave input, the lines are effectively a square wave, hence have practically infinite frequency response way up and above nyquist. For you to see no reconstruction errors on the output, there must be no frequencies > nyquist on the input. The input has not been adequately filtered, hence you see errors on the output.

Isn't sampling theory wonderful?
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
This Test was completely flawed
« Reply #33 on: May 30, 2006, 02:00:06 pm »

Quote
Four pixels per LP may be a but much, but 2 pixels per cycle will work only if the sampling is done in phase ...Roger Clark recommends 3 pixels per cycle. This is consistent  with your statement of at least 2 pixels per cycle.
http://www.clarkvision.com/imagedetail/sampling1.html
[a href=\"index.php?act=findpost&pid=66919\"][{POST_SNAPBACK}][/a]
Agreed: I was only sayng that two pixels per cycle is the theoretical minimum needed according to Nyquists' theorem. And in fact, that theorem says that the sampling rate must be slightly more than two per cycle to avoid aliasing, so a little safety margin is needed even before effects like low-pass (AA) filters and deBayerising inerpolation algorithms come into play.

From various sources, the practical figure seems to be something between 2.5 and 3 pixels per line pair. Norm Koren takes a shot at this question too, and I usually find him more reliable than Roger Clark.
Logged

Dennis

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 82
This Test was completely flawed
« Reply #34 on: May 30, 2006, 06:14:15 pm »

[span style=\'font-size:14pt;line-height:100%\']In signal processing, the Nyquist rate is the minimum sampling rate required to avoid aliasing when sampling a continuous signal.[/span]
(Nyquist rate at Wikipedia)

If you record one cycle or pair of lines with two pixel, you'll end up with aliasing. Or am I wrong here? Sure, if the recording pixel are perfectly aligned with the signal cycles, there is no aliasing, but that would be pure coincidence in a real world situation. Recording a spatial frequency with two pixel, you'll have an aliased signal or no signal. To record one cycle reliably, you need 4 pixel. So, if you complete the above with avoid aliasing reliably, you need 4 pixel, if you read it as avoid aliasing accidentially, it may be 2 pixel.
« Last Edit: May 30, 2006, 06:14:51 pm by Dennis »
Logged
Best Regards

Dennis.

Graeme Nattress

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 584
    • http://www.nattress.com
This Test was completely flawed
« Reply #35 on: May 30, 2006, 07:57:01 pm »

Denis, a two pixels to record a line pair. The problem is that a pair of lines looks like a square wave. A square wave needs an infinite frequency to describe it, and hence has many frequencies > the nyquist limit. That's why you see aliassing because the input has not been adequately filtered. If you'd put in a pure sine wave of the same wavelength of your square wave, those two pixels would reconstruct it perfectly, no matter what the phase.

Graeme
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
This Test was completely flawed
« Reply #36 on: May 30, 2006, 08:53:32 pm »

Whilst I confess to not understanding the higher mathematics of this issue, it would seem logical, in a situation where the number of lines on the target corresponds to the number of pixels on the sensor, along the same dimension, that there will be just 2 positions where the lines and pixels are either in complete registration or complete disregistration. One position will have a pixel completely illuminated by a white line with the adjacent pixel not illuminated by a black line (the best case). At the other extreme we will have a situation where all pixels are half illuminated by a white line and half not illuminated by a black line. That is, all pixels will receive the same amount of light and will produce a continuous tone of grey.

However, the vast majority of positions will be somewhere in between those 2 extremes, so results should be variable. Factor in the unavoidable distortion of all lenses and the lack of perfect uniformity of evenly spaced lines in real world scenes, then 2 pixels per line pair seems about right...... for a Foveon type sensor.

The Bayer type sensor seems to be disadvantaged by other factors such as interpolation and demosaicing, so in practice it seems like closer to 3 pixels per line pair are required for Bayer type sensors.
Logged

Graeme Nattress

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 584
    • http://www.nattress.com
This Test was completely flawed
« Reply #37 on: May 30, 2006, 09:00:32 pm »

Ah, but a line has sharp edges, sharp edges have high frequencies, and hence the nyquist limit much greater than you'd expect.... The figures of >2 pixels make sense for lines with finite precision on their edges, but you can't think of them as a signal to a sensor where you can easily calculate a nyquist value. In a real sensor, with real lines, you should never see any aliassing because the sensor is fitted with an anti-alias filter.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
This Test was completely flawed
« Reply #38 on: May 30, 2006, 09:28:38 pm »

Another limiting factor in the process of transferring high resolution detail from film to print is due to the fact that the detail always has to pass through another lens with its own MTF response (except in the case of a contact print).

However good the enlarger lens or the scanner lens, it will unavoidably degrade the detail on the negative, just as the camera lens has already degraded detail captured on film in the original shot. In both cases (scanning and 'wet darkroom') we are effectively taking a photograph of a photograph.
Logged

Terry Mester

  • Newbie
  • *
  • Offline Offline
  • Posts: 7
This Test was completely flawed
« Reply #39 on: May 31, 2006, 02:50:52 am »

I have Triacetate Negatives and Slides that are 45 -47 Years old, and they are in fantastic condition!  I have Nitrate Negatives that are 60 Years old, and Hollywood has its Nitrate Films from 80 Years ago.  Modern ESTAR Base Film possesses very high durability, and Kodak has stated that Films can last for centuries in storage if kept properly Frozen.  No hi-tech record can enjoy the longevity of a low-technology medium such as Film which never changes.  Having a computer programming background, I know the capricious nature of Software Applications and Files.  Most working people don't have the time to keep updating Computer Picture Files.  As for misinformation, I never stated anything about Optical Photographs not fading away.  That's why you preserve your Negatives.

  The figure of 6.9 Billion Molecules of Dye per Square Millimetre was scrupulously calculated, and can be affirmed by any Physicist or Chemist who understands this science.  That figure is two-dimensional, and doesn't include the actual thickness to the Dye Layers which would amount to about 100 Billion Molecules for each of the Three Dyes.  The matter of Dye Clouds, which varies based upon the amount of Light exposing the Film, is what determines the ultimate Resolution of the specific Picture.  However, the Resolution Quality provided by Molecules and Light far exceeds the ability of any electronic sensors.  When compared to an original Film Photo, I could see the smudging of finite details on a small 4x6 Digital Print made from a scanned Picture File with a Resolution of about 300 MegaPixels -- much larger than 39 MegaPixel Cameras!  Needless to say, such smudging would be more visible on larger size Digital Prints.  When enlarging Film Photographs, extending the exposure time for the Paper does not equate to having bright enough lights for a large blow up.  Extending this time will lead to a grainy Photo.  I have seen very large high quality Photographs -- so it can be done.

Quote
Exactly.

First of all, I'm glad that the preposterous premise of billions of molecules is now behind us, or are you still riding that particular hobby horse?.

Colour transparency film as well as colour negatives have a very short life span. Most will fade to the point of unusability in a couple of decades at best. The only one with staying power is Kodachrome, which is regretably fast on its way to oblivion. (How many Kodachrome labs are left in the world? A half dozen?)

B&W film is better, maybe 100+ years, but only assuming that it's been archivally processed. Otherwise, again, just decades.

The real Achilles heel of film is that it is a single physical object, subject to loss and damage. A digital file though can exists in multiple copies. Lose or destroy one, and the others are unaffected.

Yes, data storage media are subject to deterioration and obsolecense, but by making new copies from time to time they can be made to literally last forever.

As for prints, well again you're wrong. Colour prints using chemistry are fugative. A typical C print will start to seriously fade after just a few decades. Cibachrome prints after about 50-70 years. Even Dye Transfers not much more.

On the other hand Epson inkjet prints using K3 pigment inks are rated by Henry Wilhelm (the industry standard) as 100 years+ on display, and much longer in dark storage. In fact an inkjet print made on cotton rag paper with K3 inks is the longest lived colour photographic reproducttion media ever! (Carbro prints are another story, but one that's only relevent to the 4 people left in the world who know how to make them).

B&W, is also another story. An archivally processed, selenium or gold toned B&W print made on silver gelatin paper (not RC) will last for hundreds of years. But (and it's a big but), with the exception of a limited number of very skilled darkroom workers who still make such prints, you're likely to never see them outside of galleries and museums, and certainly not with your family photographs.

So. We've debunked your mythology about the superiority of optical enlargements. Now we see clearly that chemical prints in fact don't last as long as inkjets, to "ensure that you preserve your important Pictures for the future".

Any more misinformation you'd like to share with us?

Michael
[a href=\"index.php?act=findpost&pid=66775\"][{POST_SNAPBACK}][/a]
Logged
Pages: 1 [2] 3 4 ... 7   Go Up