Film vs. Scanning resolution

[warden]

Well... that's a good question isn't it? I realize, particularly in the film world, the The Print seems to be more important than the Negative-- for example, I've noticed a tendency in some to define characteristics of the negative in terms of the final print, which to me is a bit backwards. Or it's quantum physics, I'm not sure.

An example: Dave Heath used to fearlessly crop the ever lovin' life out of his 35mm negatives, turning some of them into 110 size, basically. This image is from my copy of Multitude, Solitude, which I highly recommend:

So the question is how do we digitize that little speck of a negative and print it the size that we need? The reason the final print size is so important is that it forces all other downstream imaging decisions, starting with the scanning equipment. And that can be very expensive and/or time consuming. If Heath's image needed to be the size of a postcard I'd do it myself with a Cooolscan, but if it needed to be printed 11X17" then I have no scanner that can do the job and would need to hire someone.

(Not that Heath bothered with any of this, because he was a darkroom badass. But we're talking computers and printers here.)

;-)

 

[warden]

For a single 4x5, I need 42 (6x7) separate images to reach maximum resolution (at 1:1).

This is exactly the kind of inspired lunacy I'd tune in to watch. Do it. Report back.

 

[Helge]

Helge... no ruler is required... the scan is made a 5400dpi (hardware), isn't it ?

Then the edge transition takes al lest 5 or 6 pixels (consider both sides of the edge), belive me: it is not worth the half of the scanned resolution... can you see it ?

Click to enlarge...
View attachment 253625




This is what I was teached at college (8/64 aprox), if you want I can search references for you, I've seen them sometimes.


There is also a x1.5 variability, depending on subject:

"Peak cone density varies highly between individuals, such that peak values below 100,000 cones/mm2 and above 324,000 cones/mm2 are not uncommon.[21] Assuming average focal lengths, this suggests that individuals with both high cone densities and perfect optics may resolve pixels with an angular size of 21.2 arc seconds, requiring PPI values at least 1.5 times those shown above in order for images not to appear pixelated."

https://en.wikipedia.org/wiki/Fovea_centralis




It has to be remarked that this is debatable, Ctein (in Postexposure) cites a very high mumber (30, IIRC), but inspect by eye your USAF 1951 glass slide and tell me what you see in Element 2.6 !!! (6.35 lp/mm )

There is some controversy because one thing is Visual Acuity and another one Hyperacuity, and many people don't state what they are talking about when speaking human vision.

Perhaps Hyperacuity is not popular concept, but it's quite interesting:

"Transcending by far the size limits set by the retinal 'pixels', depends on sophisticated information processing in the brain." https://en.wikipedia.org/wiki/Hyperacuity_(scientific_term)


If we speak about Hyperacuity then rating is higher, perhaps what Ctein says, but pictorial situations are not about that, we are more in the Graphic industry ratings: 6 or 7. IMO Hyperacuity vs Acuity ratings is what fuels any controversy.

Again, like the devil reads the Bible...
Now you're just cherrypicking. No sensor/camera/lens combination would be expected to completely reliably make two pixel transitions with contrast like that.
Your carefully chosen example, could be due to a number of factors such as colour bleed from the wall behind, heat haze around the ballister due to the column being hot etc.

There are plenty of other examples in the crop, where there are clear lines that are as thin or even thinner than the 200 dot pair indication at the edge of the frame. Others have edges with much sharper transitions than your example. In these it's clearly the scanner that is the limiting factor. Examples are: The double railing at the top, the horizontal lines in the wall, the wires etc.
These lines have much lower contrast than the 120 line pairs per milimeter @ 1000:1 contrast claimed in the data sheet for Reala.
And yet they still show up clearly.

And again you are conveniently eschewing Tim Parkins results, which are even stronger evidence than what is discussed above.

About your acuity talk: As I thought, it was an arsepull. The eye is a scanner and the Retina is a processor. There is no straight line from a single rod or cone to the brain. Edges, patterns, and colours have to a large degree already been decided upon, when the signal leaves the eyeball.
A good example of how impossible it is to pigeonhole the human eye is deciding the ISO, or more apt the QE of the human eye.
Dark adapted the eye can detect a single photon. In daylight the equivalent. ISO performance is estimated by some researchers to be around 1.

 

[MattKing]

for example, I've noticed a tendency in some to define characteristics of the negative in terms of the final print, which to me is a bit backwards.

It is a results oriented approach. Negatives are inherently useless unless you do something with them. So you grade their value based on what you can do with them, with prints being the most demanding likely use.

 

[Helge]

I would suspect the number of people using either well-calibrated venerable commercial solutions, and the number of people doing stitched DSLR "scans", are both fairly low compared to the people who take a 20 MP snap of their 135 or 120 negative. With a 32MP DSLR, I need two shots of a 6x6 negative to match the 2400 ppi scans I'm doing now.

I view the "archival" version as either the one I can always go back to and start over with, or the one that's an acceptable substitute if the original goes up in flames. Either way-- "archival" and "digital" are difficult to reconcile, given that technology tends to substantially alter every 5-7 years-- a serious problem for archiving the "Digital Age".

The ability to do high end scans is within almost anyones reach. That was my point.

And, this thread is about scanning and resolution, right?
When you went through the trouble of using the best image sensor there is, film, it would be foolish not to use it to it's full potential.
A job done sloppily as a holdover or interim preliminary, has a nasty habit of sticking around, and becoming the default version.

If a popular scanned image is shared among several drives there is emperical evidence that when the negative is either physically lost, or much more likely lost in some cellar or ring binder on a shelf forgotten for years, that the scan is going to be the final popular incarnation of the photo.
So better make it count when you scan. It quickly becomes an overwhelming and even impossible job to rescan.

If you want to print, crop or just examine a photo in detail, 80 MP is actually not excessive or redundant.

 

[138S]

Your carefully chosen example, could be due to a number of factors such as colour bleed from the wall behind, heat haze around the ballister due to the column being hot etc.

Helge, sorry... but I have not choosen that example, that was choosen by another poster (SCHWARZZEIT ) to say it almost had 5400 dpi effective, which is clearly not true by an x2 factor, at least. https://www.photrio.com/forum/threads/film-vs-scanning-resolution.177544/page-5#post-2314547

With the crop you selected it happens the same.

Don't tell nosense, the crop you personally posted (https://www.photrio.com/forum/threads/film-vs-scanning-resolution.177544/page-4#post-2314208)

...is not worth 1/2 of the scanned resolution, LOL

Click it to see that's not worth the half of the scanned 5400, LOL if you say this is more than 2600

______

Parkin deserves all my respect, I fully agree with the figures he posted:

But don't read the bible like the evil. That resolution is at contrast extintion, and with a flat target, real shots usually have less, and some may have to enlarge much less for the quality they want.

I won't more post about that image, it is evident it does not reach the half.

 

[grat]

It is a results oriented approach. Negatives are inherently useless unless you do something with them. So you grade their value based on what you can do with them, with prints being the most demanding likely use.

... and that, I agree with. You can't make a good print from a lousy negative, and they're a fairly important intermediate step, as a rule.

 

[Helge]

Helge, sorry... but I have not choosen that example, that was choosen by another poster (SCHWARZZEIT ) to say it almost had 5400 dpi effective, which is clearly not true by an x2 factor, at least. https://www.photrio.com/forum/threads/film-vs-scanning-resolution.177544/page-5#post-2314547

With the crop you selected it happens the same.

Don't tell nosense, the crop you personally posted (https://www.photrio.com/forum/threads/film-vs-scanning-resolution.177544/page-4#post-2314208)

...is not worth 1/2 of the scanned resolution, LOL

Click it to see that's not worth the half of the scanned 5400, LOL if you say this is more than 2600
View attachment 253701
______

Parkin deserves all my respect, I fully agree with the figures he posted:

View attachment 253698

But don't read the bible like the evil. That resolution is at contrast extintion, and with a flat target, real shots usually have less, and some may have to enlarge much less for the quality they want.

I won't more post about that image, it is evident it does not reach the half.

Exactly.
Again you manage to contradict yourself within your own post.
Show us an example of a real world shot you think qualifies as resolving to approximately the fullest potential of 5600 dpi or approximately 42 megapixels for a Barnack frame.
Any film photo is going to have some grain, but you are not going to find more real detail in a digital shot with similar resolution at that magnification.

Have a look at a somewhat similar photo taken with a D850 in full resolution: https://www.flickr.com/photos/151539916@N07/27775525518/sizes/o/

At the same magnification you'll find the same varying contrast, leading to less sharply defined edges. And the same amount of not resolving to the pixel, which you wouldn't want anyway.
It's how that kind of cropping looks.
Doesn't mean it's worthless or doesn't contribute to how the image looks pulled back.

And "LOL"‽ Really?
The clock faces minute marks clearly have quite low contrast (dirt or colouring?).That’s what makes it so impressive.

Yes, once again, like the devil... You conveniently leave out the context of the chart. Among other things he is not talking about equivalence to digital sensors/cameras of similar megapixel range, but of absolute resolution satisfied.
There is a big difference.
Among other things, a little further down he talks about the real resolution achieved by the digital cameras used in the test.
https://www.sonyalpharumors.com/sony-36-megapixels-vs-6x7-velvia-tim-parkin/
Caveat:
I'm unable to find the exact version of the article from which your crop is from (link please). But in the version I reference, the above is true.

Fact of the matter is there is a marked improvement going from 4000 dpi to 8000 dpi, and yet again further resolution revealed when you go to the microscope.
That will of course translate to real world examples too, even if the situation is less optically ideal (same goes for a direct digital camera capture, except not quite when it comes to colour resolution).

 

[SCHWARZZEIT]

C. No one except for the poster thought the scaling idea was a good way of determining real resolution.
That might be a trick for artificial images, but for anything photo, it's too fraught with multiple obvious problems.

When I suggested the downsizing/upsizing method to check at which point real image detail is lost and then draw conclusions on the effective resolution of the original file I didn’t mean it to be scientific. I reevaluated the method with a digitized negative that included a resolution test chart which allowed me to verify the effective resolution. The contrast at the film’s limiting spatial frequencies is very low. Barely resolved lines on film are more likes strings of grain clumps with uneven spacing which require more pixels to be separated, in this case about 2.28 pixels per line pair which is 14% higher than the 2 pixel per line pair minimum requirement suggested by Shannon-Nyquist. The method is only a rough approximation to determine the level of detail in a file.

That said, the resolution of film at its limit is extremely fuzzy. The randomly distributed grain clumps vary in size. On test charts with long lines the film is more likely to resolve the lines near its limit on enough points that we can detect the line pattern at sufficient magnification. Test charts with very short lines or points patterns are less likely to be resolved at the same spatial frequencies.

But provia, for example, resolves this:

View attachment 253619

You can see that even at 1000:1 extreme high contrast by 50cy/mm only 30% of the contrast remains, pictorial situations are usually under 8:1 contrast.

Fuji just don't show the low contrast MTF, saying only that extintion is at 60. This is before degradation from lenses, vibrations, focus miss, DOF vs perfect focus...

This is important to know what effective scanning resolution we need usually. Of course there are exceptions were an extraordinary high effective dpi is required, but this is not the usual case.

MTF diagrams are not for a specific contrast ratio. The MTF shows the relative modulation over a spatial frequency range of any given contrast ratio. The green line in your edited MTF drawing doesn’t make any sense.

Even discounting the above for whatever reason, there is still the rigorous test shots from Parkin. Where it is clearly shown how all of the films warrants at least 8000 dpi.
And the microscope crops shows that much more is there for most film stocks.

Tim Parkin’s article shows the crops of the converging line target at both 8000 ppi and 4000 ppi scanning resolution as well as a microscope view for 4x5” and 6x7cm formats. It’s quite obvious that the 8000 ppi drum scans barely contain any more detail than the 4000 ppi scans. For information on the effective resolution in each crop Tim provided a reference table. Based on that table the resolution of the Mamiya 7 TMX shot is limited to ~100 lp/mm when evaluated by microscope. Despite the 8000 ppi sampling rate his drum scanner is unable to resolve the limiting resolution of the film because of its low sampling efficiency at this resolution. On the website for his drum scanning service Tim Parkin has additional statements on scanning resolution and the limitations of various scanners: http://www.drumscanning.co.uk/about/resolution/

He mentions that his Screen Cezanne outresolves all drum scanners he has tested for real world film. The Cezanne is limited to 5300 ppi but apparently maintains good sampling efficiency at this resolution.

The scanning resolutions he recommends for various formats (2000 ppi for 8x10”, 2000 – 4000 ppi for 4x5”, 5000 ppi for medium format and 6000 ppi for 35mm) are based on the sampling efficiency of his Heidelberg drum scanner, but not on his tests for actual resolution on film.

In conclusion, films resolution is in the common "web mind", for lack of a better term, vastly underrated.

There are limits to the resolution of film of course, plateaus and diminishing returns.
The bar is just much higher than film has been given credit for, for the last ten to fifteen years. And there are pitfalls, gotchas and much FUD spread out there.

If you really want a number, then it's 8000 dpi. As a starting point.
And that goes for all film. Most film has real image carrying detail to, or above this point.
Even if P3200 might not quite warrant it WRT image detail, it's still important to resolve the grain in good way.

That's the minimum you should aim for if you want an archival, relatively future proof format.
Sure you could do with less, for posting online. But even for downscaling to much lower resolutions you are going to see the benefits of the higher res scan.
And for prints, there is no competition.

One of the reasons for underrating film resolution is the effective film resolution was often evaluated by scanning, and most scanners are unable to capture the full resolution potential of most regular photographic films. Of course, this makes sense when you chose a hybrid workflow with a particular scanner. But then the actual film resolution easily gets mixed up with the system resolution that is measured.

Therefore, Henning Serger provides valuable information when he shares his film resolution test results.

Regarding diminishing returns:
As mentioned above, most scanners have diminishing returns with increased resolution. Before I had the HXY scanner to accommodate even the most extreme resolution requirements, I always let my clients know about the diminishing returns when their request entailed a scanning resolution significantly higher than 8000 ppi.

Because of the long slope of a film’s resolution fall-off, a truly lossless preservation grade film digitization requires very high resolution, beyond the capabilities of most scanners. I have yet to see a system that could reliably digitize film at very high resolutions (more than 8000 ppi) fully automated. Ultra-high-resolution film digitization is costly. In the real world where cost is a critical factor, the diminishing returns have to be considered.

FADGI, who makes the digitization guidelines for US cultural heritage institutions, in their strictest 4*-standard for digitizing transmissive materials requires 4000 ppi with at least 90% sampling efficiency on MTF10% for the digitization of film formats up to 4x5” and 2000 ppi for formats larger than 4x5”.
These standards are more moderate and easier to comply to cost-effectively in the high-volume digitization environment of cultural heritage institutions.

You might have already shared this information but how big are you planning on printing, if you know?

The questions about transferring the maximum possible information from a piece of film to a computer are of interest to me intellectually, but from a practical standpoint I know I have yet to need all the information in any negative I've ever made. I consider the negative the asset and archive, and scans are temporary tools to use with a printer when I don't want to wet print. So my thinking usually goes:

1 - How big is the print?
2 - How much information can the selected printer deliver?
3 - Can I hit that quality target given a) the size of my negative and b) the scanners that I own?

Scanning for the largest intended print size while considering the intended printing method is usually the strategy most of my clients want to follow. Even for very blurry images a higher resolution scan results in a more accurate rendition of the film grain which gives the image a texture and the viewer's eye something to focus on.

-Dominique

 

[SCHWARZZEIT]

Do you have the issue number?
And/or alternatively do you have the address of the webpage the results where on? Archive.org might have a backup.

Issue 17: on films that Zeiss used for lens resolution tests
Issue 19: summary of Zeiss resolution tests for several films
Issue 24: reporting the 400 lp/mm record shot with the 25mm Biogon ZM on SPUR Orthopan UR (same emulsion as Adox CMS 20)

If you cannot find them I can send you 17 and 19 as PDF. As far as I know 24 was never published as PDF.

 

[Helge]

When I suggested the downsizing/upsizing method to check at which point real image detail is lost and then draw conclusions on the effective resolution of the original file I didn’t mean it to be scientific. I reevaluated the method with a digitized negative that included a resolution test chart which allowed me to verify the effective resolution. The contrast at the film’s limiting spatial frequencies is very low. Barely resolved lines on film are more likes strings of grain clumps with uneven spacing which require more pixels to be separated, in this case about 2.28 pixels per line pair which is 14% higher than the 2 pixel per line pair minimum requirement suggested by Shannon-Nyquist. The method is only a rough approximation to determine the level of detail in a file.

Then why even suggest it?
A guess, on a hunch, on a SWAG, on a notion, is just too inaccurate to say anything worthwhile.

That said, the resolution of film at its limit is extremely fuzzy. The randomly distributed grain clumps vary in size. On test charts with long lines the film is more likely to resolve the lines near its limit on enough points that we can detect the line pattern at sufficient magnification. Test charts with very short lines or points patterns are less likely to be resolved at the same spatial frequencies.

But that goes for any optical system and indeed many recording systems in general.
On digital it is somewhat masked by denoising and guesses from edge detection and demosaicing. But the ambiguity is still there, and especially outside of ideal cases.
It is this pre-chewing, differential treatment of different features, that is initially alluring about digital images, but also what becomes tiresome and harsh quickly.

Tim Parkin’s article

shows the crops of the converging line target at both 8000 ppi and 4000 ppi scanning resolution as well as a microscope view for 4x5” and 6x7cm formats. It’s quite obvious that the 8000 ppi drum scans barely contain any more detail than the 4000 ppi scans. For information on the effective resolution in each crop Tim provided a reference table. Based on that table the resolution of the Mamiya 7 TMX shot is limited to ~100 lp/mm when evaluated by microscope. Despite the 8000 ppi sampling rate his drum scanner is unable to resolve the limiting resolution of the film because of its low sampling efficiency at this resolution. On the website for his drum scanning service Tim Parkin has additional statements on scanning resolution and the limitations of various scanners: http://www.drumscanning.co.uk/about/resolution/

Really‽

I'd say that is a pretty significant difference.

At these kinds of magnifications those are the general differences and demising returns you'll generally see regardless of format. Again digital will have an artificially "cleaner" scrubbed look. But that is not an advantage in my book.
Diminishing returns doesn't doesn't necessarily mean it's not worth it though. It depends on where on the curve you are.
There is diminishing returns on going from walking, to a regular bike and then to a very good bike. Still, quality bikes sell to people who care.

You can still clearly see which way the scan head ran, aliasing and other artifacting. That is of course a problem, and a problem in general with these types of actually scanning scanners.

Tim has admitted long ago that camera scanning, as you have found out, is most probably better for smaller formats, and only because of potential veiling and lens flare avoided with the lensless aperture of a drum scanner, would that type of scanner still make sense for large format.

One of the reasons for underrating film resolution is the effective film resolution was often evaluated by scanning, and most scanners are unable to capture the full resolution potential of most regular photographic films. Of course, this makes sense when you chose a hybrid workflow with a particular scanner. But then the actual film resolution easily gets mixed up with the system resolution that is measured.

Exactly! How stupid is that‽ Really inexcusable from most persons who should know better.
Really smells like a conspiracy theory waiting to be coined. ;-)

Because of the long slope of a film’s resolution fall-off, a truly lossless preservation grade film digitization requires very high resolution, beyond the capabilities of most scanners. I have yet to see a system that could reliably digitize film at very high resolutions (more than 8000 ppi) fully automated. Ultra-high-resolution film digitization is costly. In the real world where cost is a critical factor, the diminishing returns have to be considered.

FADGI, who makes the digitization guidelines for US cultural heritage institutions, in their strictest 4*-standard for digitizing transmissive materials requires 4000 ppi with at least 90% sampling efficiency on MTF10% for the digitization of film formats up to 4x5” and 2000 ppi for formats larger than 4x5”.
These standards are more moderate and easier to comply to cost-effectively in the high-volume digitization environment of cultural heritage institutions.

As a museum guy myself, I can attest to the often surprisingly prosaic measures necessary or deemed necessary, by conservators and archivists.
Those are not something to be taken as a golden standard by private users.

Most people will only print a small percentage of their total number of negatives. In most cases you'll be lucky if it's one percent even for people who consider themselves photographers.
I'd say, do a contact print style scan to see what actually warrants being printed and then scan that to it's highest potential.
And then never, ever throw away negatives if you can help it.

 

[Helge]

Issue 17: on films that Zeiss used for lens resolution tests
Issue 19: summary of Zeiss resolution tests for several films
Issue 24: reporting the 400 lp/mm record shot with the 25mm Biogon ZM on SPUR Orthopan UR (same emulsion as Adox CMS 20)

If you cannot find them I can send you 17 and 19 as PDF. As far as I know 24 was never published as PDF.

Thank you so much!
I'd very much like that if it's not too much trouble (I mean the last part . Do I PM you my email address?
And thank you for the link to Tim Parkins original article. It had somehow gotten lost in Googles suggestions.

 

[138S]

MTF diagrams are not for a specific contrast ratio. The MTF shows the relative modulation over a spatial frequency range of any given contrast ratio. The green line in your edited MTF drawing doesn’t make any sense.

For film, MTF graphs are for an specific contrast ratio, this is not the same case than with lenses.

With film, depending on the contrast ratio grains of different sizes take action, so depending on contrast (and base exposure) the MTF graph is totally different.

In the 1000:1 contrast there is a 10 stops difference between lines, so in practice this can be made with a contact copy or perhaps with something like a semiconductor integrated circuit manufacturing lens (projection with reduction). By controlled fogging the film in advance (or after) you adjust contrast of the test and the base exposure of the (projected) black lines.

In that 1000:1 test, if black (projected) lines are in the toe (or not exposed) then white lines will be overexposed around +7, so you are to detect the effect of the ultra small grains having high resolving power over a "fog" from more sensitive crystals... 1000:1 does not describe the film behaviour in pictorial situations as you test crystals that are sensitive at +7 overexposure.

Instead a 1.61:1 tells the truth for common textures, 8:1 for contrasty edges, and 30:1 is good to guess performance for backlighted silhouettes.

If you see the TMX datasheet it says Modulation Transfer extintion at 200 lp/mm for TOC 1000:1 and 65lp/mm for TOC 1.61:1

The posted Provia graph is the one for the unobtanium 1000:1, clearly it has a tendence to extintion at 160lp/mm, well... commercially this MTF graph is nicer than the 1.61:1 would be.

I llustrated in green what it could be the 1.6:1 MTF graph, with extintion at 60lp/mm in that case.

_______

Kodak datasheet cites ISO 6328:2000 as the norm inspiring them for that.

_______

Let me reiterate that testing film capability at 1000:1 with a contact copy of a glass target is a totally different situation than with pictorial situations, you'll never get that situation in real shooting, by very, very far.


QUOTE="Helge, post: 2316490, member: 87684"]Exactly.
Again you manage to contradict yourself within your own post.[/QUOTE]

Sorry, I don't understand what you say or ask...

Show us an example of a real world shot you think qualifies as resolving to approximately the fullest potential of 5600 dpi or approximately 42 megapixels for a Barnack frame.

The Serger's Porsche, for example.

The clock faces minute marks clearly have quite low contrast (dirt or colouring?).That’s what makes it so impressive.

It is impressive because it is MF, MF at 2600 effective does that.

Format rules... If you want I show you how my 8x10" shots resolve, made with a W 360.
https://www.flickr.com/photos/125592977@N05/32535835184/

 

[Team ADOX]

Do you have the issue number?
And/or alternatively do you have the address of the webpage the results where on? Archive.org might have a backup.

Zeiss published it at that time on their home page (www.zeiss.de).
First publication was in their "Camera Lens News" No.17 in September 2002 ("Objektivtests - Welche Filme Zeiss benutzt, und warum").
Next publication was in Camera Lens News No.19 , in March 2003 ("Das Auflösungsvermögen von Fotofilmen").
After that Camera Lens News No. 20, in September 2004 (further test results).
And the last publication in February 2006: test results with Kodak Imagelink HQ and SPUR Orthopan UR (which is identical to ADOX CMS 20, version I ).
They explained their test methods, too. They used their Zeiss lenses, a normal tripod, daylight, normal printed medium contrast test charts. So shooting conditions equivalent to standard shooting situations in photography, and not tests with optical benches and high-contrast backlit test patterns.

Here some of the results of films which are probably most interesting:
Agfaortho 25 : 250 lp/mm
Agfa APX 25: 200 lp/mm
T-Max 100: 180 lp/mm
Fuji Acros 100: 160 lp/mm
Velvia: 160 lp/mm
Velvia 100F: 170 lp/mm
Kodak Imagelink HQ: 320 lp/mm
SPUR Orthopan UR (ADOX CMS 20): 400 lp/mm ( with Zeiss Biogon 2.8/25 at f4; it is the diffraction limit of white light at f4).

ADOX - Innovation in Analog Photography.

 

[Team ADOX]

One of the best things about camera scanning is that you can get ninety percent of the way there with very modest equipment.
An old DSLR body is more than good enough.

Hm, yes and no .
Older DSLR: That would mean a 12MP to 24 MP model. In our experience not enough to surpass the scan quality of for example a current Reflecta RPS 10 M / Pacific Image Primefilm XA or a Coolscan 5000.

ADOX - Innovation in Analog Photography.

 

[Team ADOX]

Let me reiterate that testing film capability at 1000:1 with a contact copy of a glass target is a totally different situation than with pictorial situations, you'll never get that situation in real shooting, by very, very far.

As already explained many times here on photrio:
Resolution's dependency on object contrast ist not completely linear acros the whole range from 1.6: to 1000:1. To get very high resolution with film you do not need 10 stops difference in object contrast. The resolution difference of 5 stop contrast diff. to 10 stop contrast diff. is very small.
Even at a 2 - 3 stop difference in object contrast you get very high resolution values: For example we've reached the diffraction limit at f5.6 of about 250 lp/mm (!!) with several lenses and with ADOX CMS 20 II at such quite low object contrast differences.
And with TMX, HR-50, Delta 100, Copex Rapid, Provia, both Velvias you get resolution figures at that object contrast of more than 120 lp/mm. That is even slightly surpassing 50 MP FF sensors at the same object contrast.

ADOX - Innovation in Analog Photography.

 

[Helge]

Hm, yes and no .
Older DSLR: That would mean a 12MP to 24 MP model. In our experience not enough to surpass the scan quality of for example a current Reflecta RPS 10 M / Pacific Image Primefilm XA or a Coolscan 5000.

ADOX - Innovation in Analog Photography.

Depends on the macro level, doesn’t it?
With bellows and flash backlight, I can get to very high magnifications. I need to clamp down the film between glass and use scanning fluid (using a medium format enlarge holder).
But it is possible to exceed 8000 dpi reliably with that setup.

 

[138S]

Resolution's dependency on object contrast ist not completely linear acros the whole range from 1.6: to 1000:1. To get very high resolution with film you do not need 10 stops difference in object contrast.

Let's leave apart the monodisperse CMS 20 exception, we have seen the Serger's Porsche performing... let's see the regular pictorial films in that list

Those ratings are at 1000:1, a contact copy that has no flare:

T-Max 100: 180 lp/mm
Fuji Acros 100: 160 lp/mm
Velvia: 160 lp/mm
Velvia 100F: 170 lp/mm
Agfa APX 100: ?

But our most interesting contrast is 8:1 , this is a contrasty black painted line on a luminous white wall... most of image will have way less "micro contrast" but let's consider 8:1

What are the real ratings at 8:1 ? What modulation transfer we have at (say) 50lp/mm and 8:1contrast . Think that Provia transfers only 30% at 1000:1 by 50 lp/mm, at 8:1 it has to transfer way less than 30%, How much ? 15%? 5%?


This is the point... how much Modulation ("contrast") transfer at 8:1 by 50 lp/mm? 5%?

 

[Helge]

Zeiss published it at that time on their home page (www.zeiss.de).
First publication was in their "Camera Lens News" No.17 in September 2002 ("Objektivtests - Welche Filme Zeiss benutzt, und warum").
Next publication was in Camera Lens News No.19 , in March 2003 ("Das Auflösungsvermögen von Fotofilmen").
After that Camera Lens News No. 20, in September 2004 (further test results).
And the last publication in February 2006: test results with Kodak Imagelink HQ and SPUR Orthopan UR (which is identical to ADOX CMS 20, version I ).
They explained their test methods, too. They used their Zeiss lenses, a normal tripod, daylight, normal printed medium contrast test charts. So shooting conditions equivalent to standard shooting situations in photography, and not tests with optical benches and high-contrast backlit test patterns.

Here some of the results of films which are probably most interesting:
Agfaortho 25 : 250 lp/mm
Agfa APX 25: 200 lp/mm
T-Max 100: 180 lp/mm
Fuji Acros 100: 160 lp/mm
Velvia: 160 lp/mm
Velvia 100F: 170 lp/mm
Kodak Imagelink HQ: 320 lp/mm
SPUR Orthopan UR (ADOX CMS 20): 400 lp/mm ( with Zeiss Biogon 2.8/25 at f4; it is the diffraction limit of white light at f4).

ADOX - Innovation in Analog Photography.

Thank you so much!
I have access to a technical liberary so there is a chance they have those publications.

 

[138S]

Also think that even a superb lens kills most of the contrast at high cycles.

This is a $2800 Zeiss



See that 200lp/mm are only reached in the image center, with a 70% contrast destruction in the center spot, so a black painted edge on perfect white will go from 8:1 to 2.6:1 , from the 70% contrast destruction.

 

[Helge]

Let's leave apart the monodisperse CMS 20 exception, we have seen the Serger's Porsche performing... let's see the regular pictorial films in that list

Those ratings are at 1000:1, a contact copy that has no flare:

T-Max 100: 180 lp/mm
Fuji Acros 100: 160 lp/mm
Velvia: 160 lp/mm
Velvia 100F: 170 lp/mm
Agfa APX 100: ?

But our most interesting contrast is 8:1 , this is a contrasty black painted line on a luminous white wall... most of image will have way less "micro contrast" but let's consider 8:1

What are the real ratings at 8:1 ? What modulation transfer we have at (say) 50lp/mm and 8:1contrast . Think that Provia transfers only 30% at 1000:1 by 50 lp/mm, at 8:1 it has to transfer way less than 30%, How much ? 15%? 5%?


This is the point... how much Modulation ("contrast") transfer at 8:1 by 50 lp/mm? 5%?

They explained their test methods, too. They used their Zeiss lenses, a normal tripod, daylight, normal printed medium contrast test charts. So shooting conditions equivalent to standard shooting situations in photography, and not tests with optical benches and high-contrast backlit test patterns.

 

[Team ADOX]

Depends on the macro level, doesn’t it?
With bellows and flash backlight, I can get to very high magnifications. I need to clamp down the film between glass and use scanning fluid (using a medium format enlarge holder).
But it is possible to exceed 8000 dpi reliably with that setup.

8000ppi with a 24 MP FF DSLR as a scanner? No, definitely not. We've never seen such results. It is physically impossible, as the Nyquist frequency of these sensors is too low.

ADOX - Innovation in Analog Photography.

 

[Team ADOX]

Thank you so much!
I have access to a technical liberary so there is a chance they have those publications.

You are welcome.

ADOX - Innovation in Analog Photography.

 

[138S]

Those values are 1000:1, not 8:1, and they are at extintion :

T-Max 100: 180 lp/mm
Fuji Acros 100: 160 lp/mm
Velvia: 160 lp/mm
Velvia 100F: 170 lp/mm
Agfa APX 100: ?

__________________

My question remains, What contrast transfer at 8:1 by 50cy/mm ?

For Provia is 30% contrast transfer at 1000:1 by 50 cy/mm. What transfer at 8:1?

 

[Team ADOX]

Let's leave apart the monodisperse CMS 20 exception, we have seen the Serger's Porsche performing... let's see the regular pictorial films in that list

Those ratings are at 1000:1, a contact copy that has no flare:

T-Max 100: 180 lp/mm
Fuji Acros 100: 160 lp/mm
Velvia: 160 lp/mm
Velvia 100F: 170 lp/mm

Completely wrong. That Zeiss ratings are not ratings at object contrast 1000:1 !! And not contact copies !!
Zeiss used charts with contrast in the 4 to 5 stop range.
That are the test results Zeiss has got (and published) with their Zeiss lenses, with normal cameras, on a good, standard tripod. They have made these test shots the same way a good photographer works when doing his landscape, architecture or studio photography. Zeiss used the same equipment.
That was explained by them in the Camera Lens News articles.
Quote:
"Im allgemeinen fanden wir, dass die Angaben der Filmhersteller Agfa, Fuji und Kodak zum Auflösungsvermögen ihrer Filme zuverlässig und genau sind. Um diese Werte zu ermitteln, verwenden Filmhersteller spezielle standardisierte Laborverfahren unter Verwendung hochauflösender Mikroskopoptik. Bei Zeiss benutzen wir dagegen normale fotografische Verfahren. Wir verwenden dabei handelsübliche Kameras mit Zeiss-Objektiven und erreichen damit die gleichen Werte für das Auflösungsvermögen wie die Filmhersteller. Damit ist das - oftmals hohe - Auflösungsvermögen, dass die Filmhersteller unter Laborbedingungen erzielen, nicht bloss ein theoretischer Wert, sondern kann auch in der realen Fotografie erreicht und genutzt werden."

ADOX - Innovation in Analog Photography.