film is higher res than digital--Puts

[bags27]

Advertisement (gone after registration)

160 lp/mm with Leica M7 and Ilford 100D!

With the new generation of image sensors that offer between 40 and 50 Mp, a separating power of up to between 80 and 90 Lp/mm is possible according to the test chart and test conditions I use.
Under the same conditions I investigated the (old) Ilford 100 Delta. That film was loaded into the Leica M7 and the Apo-Summicron-M 50 mm with an aperture f/5.6 and the distance between camera (of course on tripod) was 2.2 meters. Development in SuperGrain (an improved version of the classic Amaloco AM74). To my surprise, this combination delivers a separation power of 160 Lp/mm, so almost twice as much as is achievable with the most modern sensors.
So instead of spending a lot of money on upgrading to a current image sensor (which is guaranteed to be obsolete in a few years time) you better buy an old M6 and load it with film from the Ilford stable. The editing is a bit more difficult and time consuming, but the results are still much better than what you can achieve with a current and expensive image sensor.

https://photo.imx.nl/blog/

[jaapv]

Erwin contradicts himself in that blog. In his lens compendiums he argues, supported by Zeiss research, that any resolution over 80 LP/mm has no relevance for film.  And now it is the other way around?  🙄  These discussions are utterly pointless. Are those lines going to improve your photography?

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[philipus]

He's made some similar claims in previous blog posts. It would've been good if he could have posted some images to prove this somehow.

[adan]

Playing Devil's Advocate:

1) It would be more correct say that one film (not film in general) in a particular developer is higher-res than certain digital sensors. And frankly, if one wanted to mess around with ISO 20-40 microfilms in special developers, that was already true. The gain here is not so much resolution as ISO speed (now up to 100).

(I was in high school when H&W Control VTE film (repackaged microfilm - bright blue ) and developer came out in 1971. Magazine ad below via flickr. Incidentally, H&W Control's developer was also a hydroquinone/phenidone combination, like Rollei SuperGrain.)

https://www.flickr.com/photos/nesster/4922598585

That doesn't mean that Portra 800 or Portra 160 or Tri-X or HP5 or even Delta 400 are now magically higher-res than digital (absent going to 6x9 or 4x5).

2) Looks like this film/dev combo would resolve on the close order of an 88Mp Monochrom body (and a 50 APO-Summicron). If we use the common measure of human visual acuity as being ~5 lp/mm looking at a print, such an image could be printed to 5 x 7.5 feet or 2,3 x 1,5 meters with "apparently infinite detail". With the usual caveat that the photographer's technique throughout the process would have to be top-notch as well.

Whether that performance has relevance for any particular photographer is up to the photographer. But an achievment is an achievement, regardless of relevance. One can travel around the Earth at 28000 kph. That may not be relevant to anyone but NASA and the CIA and such, but it remains a notable achievement.

3) Erwin was no doubt checking the film with a loupe. To actually use that much resolution would require a wet darkroom with the capability to enlarge to the sizes mentioned above. Horizontal or laser-drum projection.

Using a scanner would require one that can actually scan (and have its own lens that resolves) around 12000 ppi, same as the original lines per mm (double the lp/mm). Even a Flextight would not do the job.

Put simply, if the scanner or enlarger lens is not effectively equal to a $8800 APO-Summicron itself - fugedaboutit!

[darylgo]

Curious, Kodachrome 25 or older Kodachrome II, where would it fit into this discussion?  

[adan]

Kodachrome 25/64 appear to hit MTF50% at about 40 lp/mm and MTF10% at about 80 lp/mm - per Kodak's own charts:

https://125px.com/docs/film/kodak/e55-2009_06.pdf

Back in 2000-2006, the Gigapxl Project "tuned" their whole workflow (custom lens, scanner res) around the 40 lp/mm red-light resolution of the color-neg spy-plane film being used - for 9" x 18" negatives.

Sadly, the project's web presence, and especially their article on the technology, no longer exists. But you can get a feel for the "zoomability" of their images here: http://somethinofnothin.net/blog/wordpress/2007/05/05/megapixel-meet-gigapixel/

[darylgo]

Advertisement (gone after registration)

Thanks Adan, I need your help on those curves, it starts at > 100% at low spatial frequencies, how could this be?   

[adan]

Probably development edge effects - sort of a chemical "unsharp mask".

Relatively fresh developer from relatively silverless shadows drifts across the edges with bright areas and adds development between agitations, enhancing the edge contrast. Meanwhile, the developer that has been working hard in the silver-rich areas poops out due to bromide drag, and drifts into the shadow edges, reducing edge development there. Result: halos of blacker blacks and whiter whites right along the edges.

I don't know about K'chrome specifically, but Fuji promotes their use of DIR (Development-Inhibitor-Releasing) compounds in their emulsions, to produce this effect. It is also supposed to be an acutance benefit of using high-dilution developers like Rodinal (they "die faster" on the side of a tonal boundary with more silver), or using stand-development (no agitation except for Brownian Motion, or other currents in the developer).

This is one reason I take lenses MTF charts with a certain grain of salt - a rather fuzzy lens with very, very good coatings to enhance contrast overall, can produce a higher (or at least equal) MTF than/to a really sharp lens with 1950s coatings (e.g. 50 Summicron "Rigid.")

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

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

https://books.google.com/books?id=G3D3v_S6X7UC&pg=PA52&lpg=PA52&dq=Fuji+velvia+"edge+enhancement"&source=bl&ots=UNAlTWYxvB&sig=ACfU3U1xogAaK1s_Wn4GxV6ssDThuFQDvA&hl=en&sa=X&ved=2ahUKEwiBm6npmtLoAhXLG80KHf-3DPkQ6AEwAHoECAwQKA#v=onepage&q=Fuji velvia "edge enhancement"&f=false

 

[Chuck Albertson]

Not convinced. A few years back, I shot the same scene off my balcony with (1) the 'blad with a 150mm lens on 100 Delta (souped in Rodinal) and (2) the MM1 with a 90mm Summicron AA, and printed both. The Monochrom image looked quite a bit sharper. That didn't cause me to dump the 'blad, though.

[pgk]

On 4/5/2020 at 1:03 PM, darylgo said:

Curious, Kodachrome 25 or older Kodachrome II, where would it fit into this discussion?  

Recently I've been copying Kodachrome 64 slides using a Sony A7II (24MPixel). Grain is apparent in the copies suggesting that the 24MPixel sensor out resolves the film quite easily. The problem with any resolution test involving film is that its using a high contrast target which rarely equates to real world imagery and as Adan says, this can be subject to edge effects which may well provide spurious data in other circumstances. I have a photo science background and am fairly of the opinion that in the real world, most 10MPixel+ sensors provide more information than most 35mm colour films. Notice that I state 'information' which is not necessarily 'resolution'. Suggesting resolution is all that matters is a very simplistic view of an image.

[darylgo]

23 minutes ago, pgk said:

 Suggesting resolution is all that matters is a very simplistic view of an image.

Could you expand on this, perhaps get into the technicals from your background, fascinating stuff (to me).  Thanks adan and pgk.  

[pgk]

Not easily. The problem is that using pure resolution as a yardstick requires specific 'test conditions'. These rarely equate to much that we actually photograph, although they may be relevant to some scientific recording needs. MTF charts are more helpful but require a reasonable amount of interpretation and are complex in that, in order to give an overview of how a lens performs, many graphs need to be viewed. But to give an idea of the problems in achieving a detailed image, consider that diffraction limitation kicks in with many digital cameras at smaller apertures so the resolution limit is not dictated by the sensor but the lens aperture. The higher the MPixels the wider the aperture at which diffraction limitation occurs. However real world photographs may need a small aperture in order to work in terms of depth of field.

So as a real world example, I shoot a lot of macro underwater on Cameras such as A Canon 5DII and Sony A7Ii. By testing, I achieve a balance of high resolution and sufficient depth of field at around f/11-16. At smaller apertures the images soften marginally, but faster than f/11 means that dept of field at say 1:1, is too shallow to show as many features of a natural history subject as I want (I am a marine fish specialist in another photographic life). So increasing resolution would require opening up the lens and in reality the compromise if to use f/11-16 because wider apertures don't usually capture enough, sufficiently in focus detail which I need. Ideally I'd like more resolution and depth of field but physics gets in the way. Increasing sensor Pixels isn't going to change this other than giving more pixels unfortunately.

And for what it is worth, before I shot digital I used to use Fuji Velvia 50. A comparison of images of small fishes (gobies) shot an a 5DII and film camera on Velvia 50 using the same lens has shown me that the detail which I need to see of things like sensory papillae on gobies is not visible on Velvia (taken with flash) but is clearly evident on the Canon digital sensor. FWIW Velvia appears to have about the same information capacity (equivalent) for this sort  of photography as has a Canon 5DmkI, in terms of what is visible in the photograph. So, as far as I am concerned, photography is about trade-offs and compromises, and resolution is just one of the components which needs to be 'optimised' rather than maximised. Does that make sense?

[darylgo]

Yes, it makes perfect sense, thank you.  Perhaps you can clarify one further point, if you photographed on a higher resolution camera (5dMkIV) at f16, could you down res to the equivalent of a MkII and achieve the same result?  Logically this would make sense but physics doesn't always follow my logic (or what I think might be erroneously thought logical 🙂).

[pgk]

3 minutes ago, darylgo said:

Yes, it makes perfect sense, thank you.  Perhaps you can clarify one further point, if you photographed on a higher resolution camera (5dMkIV) at f16, could you down res to the equivalent of a MkII and achieve the same result?  Logically this would make sense but physics doesn't always follow my logic (or what I think might be erroneously thought logical 🙂).

It doesn't work quite like that. You have more recorded data of the same information. Downrezzing is a mathematical process and the resultant is unlikely to look the same as the one from a lower Pixel camera. But the information recorded will be the same despite this - that is to say, there will be no more detail in either image. The easiest way to check on such things is to take identical, real world photos on different gear and see what you can make out in them for yourself. Despite the emphasis placed on technical specification and image theory, photography is a practical subject and you cannot beat testing equipment to decide what does the job you want it to.

FWIW, I showed some prints from my M9 to a graphic designer friend the other day - printed to its 18Mpixel resolution at 200dpi which is something like a 26" wide print. He stated that was impressed by the detail. Why would I upgrade since I don't print bigger if an image conscious viewer is satisfied by the print quality? 

[giannis]

From general experience, I call nonsense. It's very misguided to turn to small format film if resolution is your goal.

You would need an excellent lens, tripod, and a top end enlarging lens (Apo Rodagon etc.) to hit 110-120lp/mm with Delta 100. The moment you scan your film, even with a high end scanner (including drum scanner which virtually no hobbyist can own and operate), you slash that lp/mm value by 1/3rd. Your 120lp/mm easily becomes 80lp/mm with "just" a great scanner like a Nikon Coolscan, or with sub-par enlarging lenses. You need to remember the general equation: 1/(system resolution) = 1/(res of component A) +1/(res of component  and a scanner has a sensor and lens, bringing the system resolution down considerably.

Another major point is at what *contrast* was this resolution achieved. Film has good resolution at high target contrasts (256:1, 128:1), but reolution drops dramatically at lower contrasts (4:1, 2:1) while digital stays the same. The photographically relevant contrasts are around 1.5:1 to 4:1, a test done with a target of contrast 1000:1 is in practice irrelevant.

In practice, with good lenses and technique, and good scanner/enlarger (not top end), in 35mm format you can get around 12-16 megapixels of detail out of "normal" fine grained B&W films, or slides. If you want more, you just shoot a larger format instead chasing after top-end lenses and scanners/enlargers.

Edited April 8, 2020 by jaapv

[eilert anders]

I tested my Leica R lenses with the USBS test charts placed 76 focal lengths from the camera film plane.  With this set-up, I could discern resolution up to 240 l/mm.  The only lens that could achieve this resolution was the 50mm Summicron R (first version) at f4, and only in the center.  THe seperation of the lines was superb, so I suspect the resolution was much higher than the 240 l/mm I could detect.  The film I used was Kodak high contrast copy developed in one of the developers that enabled continuous tones.  I used my Focomat ic enlarger fitted with a 50mm Focotar lens.  In looking at the projected image with the enlarger at its top position (about 10x) with a 10x micro omega grain focuser, I could detect this resolution level.  To see this resolution in a print, one would have to enlarge at least 30x.  At this level, the usual enlarging lens would not be  up to snuff!  Even a resolution of 80 l/mm is not seen in a 10x print.  This is the reality.  Putts stated that Leica designed their lenses to achieve best results at 20 lp/mm (40 l/mm).  They do such a good job that great results are achieved at higher MTF levels such as 40 lp/mm (80 l/mm) 

The Summicron across the field resolved about 200 l/mm.  The 90 Elmarit R (first version) resolved about 180 l/mm.

I miss the much beloved Kodak High Contrast film.  I photographed my first daughter on a swing with my 90 Elmarit R lens at f5.6 with the camera mounted on a sturdy tripod from about 20 feet.  The results were spectacular.   A blow-up of her face showed no grain, and her eyelashes were distinctly resolved.  Certainly a great testimonial to the quality of the HCC film and the Leica 90mm Elmarit R lens, my all time favorite lens.  I could not achieve this quality with tech pan or VTE films and developers.

Kodachrome distinctly resolved 68 l/mm with great acutance and just barely resolved 80 l/mm.  Fuji Velvia just resolved 80 l/mm, but with low acutance.  The visual impression is that Kodachrome is sharper because of the high acutance.  I observed this difference in Kodachrome 25 and velvia 50 slides I took of the Grand Canyon of the Yellowstone with the 100 f2.8 APO Elmarit R lens.  The Kodachromes are super sharp (high acutance) with great color and the Velvia 50 slides are sharp, but with lower sharpness impression.  The color of the Velvia slides is exaggerated but very attractive.

[Guest]

The Focotar lens is the weakest link imo.

[jmahto]

On 4/8/2020 at 5:01 AM, giannis said:

......

In practice, with good lenses and technique, and good scanner/enlarger (not top end), in 35mm format you can get around 12-16 megapixels of detail out of "normal" fine grained B&W films, or slides. .....

Since I started shooting film two years ago (after a hiatus of several decades), I have done several tests of my own to compare film vs digital resolution and I am in agreement with this. Ijust want to add that 12-16 mega pixel is with fine grain film and good lens/technique (lens, focusing etc). For my grainy tri-x I won't assume more than 6-7 mega pixel (but I love that for aesthetic reasons).

I have concluded that now a days digital (of same sensor size) is superior in resolution but may look artificial in details with heavy NR at blown up sizes. Film, on the other hand keeps the organic look with that familiar grain.

[jmahto]

On 4/6/2020 at 1:34 PM, darylgo said:

Yes, it makes perfect sense, thank you.  Perhaps you can clarify one further point, if you photographed on a higher resolution camera (5dMkIV) at f16, could you down res to the equivalent of a MkII and achieve the same result?  Logically this would make sense but physics doesn't always follow my logic (or what I think might be erroneously thought logical 🙂).

For color fidelity higher MP digital photo (even downsized) will be better. Just because bayer algorithm has more data to play with.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Ambro51]

The Stereo Realist Custom came along in 1959 at the tail end of the Stereo craze.  By this time, many serious stereographers were using glass mounted slides in special stereo projectors.  The screens were special and polarized glasses needed.   The Realist Custom was assembled using hand picked parts to a high standard.  The lenses were capable of 300 lpm resolution.  They were 4 element Tessar type made with “rare earth glass” and precision matched.  Though not so indicated Steinheil  Munchen made these.  Focus in the camera build was set by a microscope.  Film...Kodachrome of course.   ••••• I have no doubt the performance here outshines anything digital.

Edited May 3, 2020 by Ambro51