Research and Development of film

[ath]

+1 - I wonder when it will be available?

It was said that samples will be handed out at this photokina. You can order them at fotoimpex as well. Film and developer are marked as being on stock.

 

[Henning Serger]

I am doing lens, film, developer (and lately sensor) tests for more than 20 years now, on a scientific basis.
Running a private, non-profit optic lab and a network of experienced photographers working on these subjects. We have more than 7000 test shots here and have tested almost all films of the market during the years.
So I certainly can give you exact answers.

The pixel game can become meaningless as lens resolution becomes a factor. A very good general purpose lens will only resolve 120 lines per mm for a high contrast subject. Most resolve less. For a 24X36 sensor, that's only 2880X4320, or 12.4 megapixels. The combination of film and lens usually only gets you about 80 lines per mm, maybe 95 with the best equipment and common film.

That is not true. It is an internet myth mostly spread by people who have never did proper tests by themselves.
So let’s have a look at the test results of those who have done very detailed, scientific tests:

Carl Zeiss:
Zeiss published system resolution (Zeiss lens + film) values in their camera lens news 17, 19, 20, 24 and 30. Object contrast 1:32 (five stops). Some examples:
Velvia 50: 160 lp/mm
Velvia 100F: 170 lp/mm
Acros 100: 160 lp/mm
T-Max 100: 180 lp/mm
Agfa APX 25 : 200 lp/mm
Agfaortho 25 : 250 lp/mm
Spur Orthopan UR: 400 lp/mm (with 25mm ZM Biogon at f4; 400 lp/mm is the diffraction limit of white light at f4).

Under the same test conditions with the same object contrast and the same Zeiss lenses a Nikon D800E resolve about 105 lp/mm (that is the Nyquist frequency of its sensor, which is the physical resolution limit of a 35 MP 24x36 sensor).

160 lp/mm with Velvia 50 is the value Fuji has published for this film for an object contrast of 1:1000 (10 stops). The resolution values Fuji has published are very conservative, especially the ones of their color films. I've seen results from a collegue who achieved even higher results at medium contrast than Fuji has published for high contrast.
That is all right and not a contradiction because:

Resolution is dependant on object contrast, but it is not a linear relation, but following "the law of diminishing returns". It is an asymptotic curve.
It doesn't matter much whether you have 6 or 10 stops object contrast, the resolution is almost the same.
But it does matter whether you have 0,5 or 2 stops object contrast. In this range there is an almost linear relation (see camera lens news no. 30 for further details). In this low contrast range up to 1,5 stops the modern 24 MP and 35 MP 24x36 sensors have a very good resolution performance and are better than most films. But from 1,5 stops on and higher, in the medium and higher contrast range, modern films (especially color reversal films, tabular grain BW films and of course microfilms), show a significantly higher resolution. Because they are not limited by the Nyquist frequency like digital sensors.

Tests from our team:
We have tested all the films with an relative low object contrast of 1:4 (two stops). Lenses were Nikkor AI-S 1,8/50 (long barrel version) and Zeiss ZF 2/50 at f5,6 (and some others as well).
Both lenses have the same performance in the center at f4 and f5,6, but the Zeiss is generally better at the corners and at f2.
Some test results from our resolution tests (Nikon F6, MLU, MC-30, 1/250s, focus bracketing, Zeiss ZF 2/50, f5,6, Nikkor AI-S 1,8/50, object contrast 1:4; Berlebach Report 3032; the first resolution value represents the number of clearly separated lines, the second one the resolution limit where still a contrast difference can be seen):

Adox CMS 20 / Spur Orthopan UR developed in: Spur Nano Edge, Spur Nanospeed UR, Spur Modular UR, Adotech: 240 - 260 Lp/mm
That is the physical diffraction limit for white light at f5,6 !

Agfa Copex Rapid; ISO 40/17°; Spur Modular UR New: 165 – 180 Lp/mm
TMX, developed in HRX-3: 135 - 150 Lp/mm
Retro 80S, HRX-3, ISO 25/15°: 135 – 145 lp/mm
Delta 100, HRX-3: 130 – 140 lp/mm
Acros 100, HRX-3: 115 – 130 lp/mm
Pan F+, HRX-3: 110 – 130 lp/mm

Fuji Velvia 50: 110 – 125 Lp/mm
Fuji Sensia 100: 120 – 135 Lp/mm
Fuji Provia 100F: 120 – 135 Lp/mm
Fuji Astia 100F: 120 – 135 Lp/mm
Fuji Velvia 100: 125 – 140 Lp/mm
Fuji Velvia 100F: 125 – 140 Lp/mm
Kodak E100G: 120 – 135 Lp/mm
Kodak Elitechrome 100: 120 – 135 Lp/mm
Fuji Provia 400X: 105 – 115 Lp/mm

Fuji Superia Reala 100: 105 – 115 Lp/mm
Fuji Pro 160 C: 100 – 115 Lp/mm
Kodak Ektar: 90 – 105 Lp/mm

We’ve further done all the relevant tests in the following imaging chain: Optical printing, slide projection, scanning (both with 4000ppi semi-professional scanners and high-end drum scanners).
Optical printing and slide projection delivers by far the the highest resolution. With excellent lenses you can transfer the above listed resolution values with only a minimal (not relevant) loss onto paper and on the projection screen.
The resolution loss with high-end drum scanners (we’ve tested Imacon X5 and ICG 370) is significant, the resolution performance is worse compared to optical printing with APO enlarging lenses and slide projection.
The biggest resolution loss and worst performance delivered the Nikon Coolscan 5000 scanner.
It’s effective resolution limit is 65-70 lp/mm.

Tests from Tim Parkin:
http://www.onlandscape.co.uk/2011/12/big-camera-comparison/
http://static.timparkin.co.uk/static/tmp/cms20-vs.jpg

Tests from Antora et.al and Heuer et.al:
http://www.aphog.de/index.php?option=com_content&task=view&id=401&Itemid=1

Best regards,
Henning

 

[Diapositivo]

Wow, Henning, thanks for that. A truckload of very interesting information indeed. I did save it in a file for future reference.

Frankly, I have some questions about the test method used for the resolution of the Nikon Coolscan 5000. I wonder e.g. if the test was performed with 16x oversampling and with 16bit per channel. The cleanliness of the lens element in the scanner should also be checked. I bought my scanner new in 2007 and when not in use it is always covered with a cloth + plastic bag. Tabletop scanners who are kept on the desktop for years do develop a layer of thin dust on the optical path which certainly degrades performance.

This test seems to be more in line with my experience:

http://www.filmscanner.info/en/NikonSuperCoolscan5000ED.html

browse down to "Image quality", the author arrives to a resolution of 4100 dpi (should be ppi) horizontal and 3650 vertical, averaging to a value of around 3900 ppi. That should be above 70 lp/mm and closer to 80 lp/mm.

In any case it is very interesting to see that the scanner is the weak link in this chain. Future desktop scanners might be able to extract even more details from our slides.

Fabrizio

 

[Photo Engineer]

Henning, that Ektar comparison bothers me. I know that this film was designed to maximize sharpness and grain and the fact that it shows poorly in your example just does not match up with other information.

Any thoughts?

PE

 

[Henning Serger]

Hello Fabrizio,

This test seems to be more in line with my experience:

http://www.filmscanner.info/en/NikonSuperCoolscan5000ED.html

browse down to "Image quality", the author arrives to a resolution of 4100 dpi (should be ppi) horizontal and 3650 vertical, averaging to a value of around 3900 ppi. That should be above 70 lp/mm and closer to 80 lp/mm.

I know the filmscanner -info test. They used a printed higher contrast test target for their test, whereas we've used our real photos, our slides and negatives with the lower 1:4 object contrast testchart we have photographed. Our test therefore is much more relevant for real life shooting (that is always our aim).
With a higher object contrast, e.g. 1:16, you can surpass the 70 lp/mm with a Coolscan 5000, that is right. But the absolut physical limit is in 75 - 77 lp/mm range with this scanner.

But with a higher object contrast (in the subject you photograph) of maybe 1:16 you will also have a higher resolution in your original slide or negative. Then the resolution in our test would not be 120 clearly separated lines with the Zeiss and Nikon for example with Provia 100F, but 130 or higher.
So the huge resolution loss with the scanner is still there.

In any case it is very interesting to see that the scanner is the weak link in this chain. Future desktop scanners might be able to extract even more details from our slides.

Fabrizio

Don't waste your time waiting for better scanners for slides.
The method for highest performance with slides is already existent: Slide projection!
We've tested several projection lenses, e.g. the Leica Super-Colorplan P2, Zeiss P-Sonnar T and the Doctor-Optics / Kindermann MC-B.
With all three lenses we could transfer the 120 lp/mm of E100G, Elitechrome 100, Provia 100F, Sensia 100, Astia 100F onto the screen!
Same with 165 lp/mm of Agfa Copex Rapid as BW-Slide (developed in Scala reversal process)!
And even better: We projected a CMS 20 negative with 240-260 lp/mm and got 230 lp/mm transferred onto the screen!!
As now the recent tests with CMS 20 as BW slide have been successful, it is planned to do another test with the CMS 20 slide in projection. The resolution then will probably
even higher because of the higher contrast of CMS 20 slide compared to the CMS 20 negative.

For big enlargements slide projection delivers an unsurpassed quality. And at extremely low costs ( a projected slide on a 1m x 1,50m screen cost me less than a buck, a print of the same size cost me more than hundred bucks).

Best regards,
Henning

 

[Henning Serger]

Henning, that Ektar comparison bothers me. I know that this film was designed to maximize sharpness and grain and the fact that it shows poorly in your example just does not match up with other information.

Any thoughts?

PE

Hello Ron,

all published scientific resolution tests of Ektar so far had the same results and confirm our test results. E.g. the tests from Antora et.al. and Heuer et.al. (please see the link in my first post), or the tests from D. Ventzke.
We have a long history testing Ektar: We did the first tests with samples given by Kodak at Photokina 2008, when the film was introduced. These films show the 90-105 lp/mm resolution mentioned above. Honestly, we've been a bit disappointed about that.
As good scientists we've checked our test system several times, also with other films, and saw that our system worked perfectly. After that we did test Ektar in the following three years again three times, all films from different batches.
We've got always the same results.

Ektar has a bit finer grain than the other color negative films (but not quite as fine as the ISO 100 slide films like E100G, Provia, Astia etc.). But the resolution and sharpness at the higher spatial frequencies (above 80 lp/mm) is not as good.
At a second view it is indeed not so surprising: This film is very consequently optimised by Kodak for the workflow probably more than 95% of todays color negative film shooters have:
Scanning with max. 4000ppi.
For such a workflow you don't need resolution above 80 lp/mm. Because of the resolution limits of the scanner.
You need fine grain, especially because almost all of these scanners do enhance grain by scanner noise.

In this workflow, for which Ektar is designed, this film does a very good job.
But in absolut terms, it is not the film with the highest resolution and best sharpness (at higher spatial frequencies).

Do you have a microscope? Then I can send you the Ektar films and other Kodak and Fuji films. You will see the differences between the films at once.

So, now dear friends, I will be offline for the next three days. I am going to the Photokina . Sometimes it is an advantage to live in Germany.....

So long and best regards,
Henning

 

[DREW WILEY]

In the REAL world - mine at least, all this is pretty meaningless. Lens resolution cannot evolve at the
same pace as sensors, and all those DLSR's etc cannot compete with basic ancient movements provided by LF cameras. Tilt/shift lenses might be OK for some things, but overall, are a pretty poor
substitute. If you want to venture into LF quality, then figure out all the above, esp a lens with fifty
times more resolution, not two times more! Even in med format, having movements means you can
distribute focus and not rely solely on f-stops or wide-angle lenses to achieve depth of field. Charts
might be useful for comparing similar lenses or film grain per se, but otherwise only mean something if you plan to shoot charts at optimum apertures! Size matters, and that's why in most real world
scenarios, film still rules! And yeah, looking at a damn computer screen sure doesn't carry the impact of an old-time slide show! But then an old 3-lantern system from nearly a century ago, with
large black and white tricolor negs would pretty much blow that away. So maybe we're talking about
R&D de-evolution!

 

[Prof_Pixel]

This is a guy who has been doing large format digital for years and his work is fantastic! http://www.sjphoto.com/

 

[Newt_on_Swings]

That was a really good read, lots of great info from Henning! Thanks

 

[Diapositivo]

Don't waste your time waiting for better scanners for slides.
The method for highest performance with slides is already existent: Slide projection!

That was one of the most interesting parts of your post. When analogue is compared to digital the comparison is always performed on the digital playing field, i.e. converting analogue to digital (by scanning) and comparing the results.

Your tests confirm that if the comparison were made on the analogue field, i.e. converting the digital image to analogue, and optically printing, or projecting, the two images, digital would come out of the comparison in a different way.

In my case, though, as I send to stock agencies most of my pictures, I have to compare the technologies on the digital "field". I hope that my trustful Super Coolscan 5000 will last for many many years to come, but it's nice to know that, when the moment comes to replace it, I might replace it with something more performing and still hopefully find more information on my slides.

 

[Pioneer]

It was said that samples will be handed out at this photokina. You can order them at fotoimpex as well. Film and developer are marked as being on stock.

Thanks for the link. I have some on the way.

 

[ADOX Fotoimpex]

The method for highest performance with slides is already existent: Slide projection!

On our booth at Photokina we are projecting slides from our new Silvermax film.
Unfortunately the cheap digital camera in my mobile could not capture the full contrast of the scene ;-)

 

[Andre Noble]

I think Henning is correct. Seriously.

I have said Nikon 5000 scanner - and I feel I can extract much more information from optical printing from a 35mm frame than I can with the Nikon 5000 scan of same frame (and then printing via ligthjet/lamda).

 

[nworth]

Diffraction limited resolution for an f/2.8 lens is about 536 lp/mm; at a more reasonable f/8 it's about 188 lp/mm. Practical photographic lenses do not even approach this theoretical resolution over their focal spans, although some, like Zeiss, are quite excellent. Problems with focus, camera construction, film construction, and holding the film correctly in place reduce the resolution, and the resolution is never equal over the entire frame. When you combine the theoretical maximum resolution at f/2.8 with a film giving 160 lp/mm resolution (seldom achieved) you get something like 123 lp/mm. Even at 80 lp/mm (a frequently measured resolution with good film and equipment), the results we get with film look fantastic. (Note that traditional resolution measurement is difficult and somewhat subjective. More modern MTF measurements can be automated and give more and better information, although they still vary somewhat, depending on the instruments.) There are plenty of good reasons for using film, even in small formats, beside resolution - e.g. preservation, storage, and sometimes rendition. I've been in a position the last few years to see a lot of work, both film and digital, from a lot of photographers, and I will stand by my statement that digital photos, at their best, are consistently of better quality than 35mm film photos. At larger formats, I think film is consistently better.

 

[Henning Serger]

Hello Drew,

Lens resolution cannot evolve at the
same pace as sensors, ...

right, but it is not necessary at all that lens resolution evolve at the same space.
I've had quite a lot of interesting talks about that with lens designers: Lots of the current prime lenses are only diffraction limited, and have therefore a resolution limit given by the wavelenght of light.
As explained above, for example we've reached the diffraction limit for white light at f5,6 with both the old Nikkor 1,8/50 AI-S (a lens design from 1978) and the modern Zeiss ZF 2/50. We've reached this physical limit at that f-stop in combination with the CMS 20 (Agfa HDP) film. And that at a quite low / moderate object contrast of only 1:4 (two stops). You need an 200 MP 24x36mm Sensor to get the same resolution under the same test conditions.
With Agfa Copex Rapid still extremely high 165 clearly separated linepairs per millimeter ( you need a 94 MP Sensor to get the same resolution under the same test conditions).
And even with my old trusty workhorse Ilford Delta 100 very high 130 clearly separated linepairs (you need a 58 MP Sensor to get the same resolution under the same test conditions).

The physical resolution limit (Nyquist frequency) of the D800E is about 105 lp/mm. No matter how high your object contrast is, it is impossible to surpass this limit. And a better lens wouldn't help you either.

So we see the lenses are not the main problem. In most cases their resolution capabilities are more than good enough.

and all those DLSR's etc cannot compete with basic ancient movements provided by LF cameras. Tilt/shift lenses might be OK for some things, but overall, are a pretty poor
substitute.

Of course you have a much greater range in movements with LF cameras.
But there have been just introduced some very nice new T&S lenses for 35mm camera systems by German lens manufacturer Schneider. I've seen them at Photokina. Very impressive, huge image circle, excellent mechanics and optics (new lens design). Really worth a look.

If you want to venture into LF quality, then figure out all the above, esp a lens with fifty
times more resolution, not two times more!

Well, no one is denying the outstanding quality of large format. But sometimes there are exceptions from the rule.
When we test lenses, films and sensors, then we not only do that in standardised (lab) tests, but also always in real world shooting, 'out there in the fields'.
And there we've compared Agfa Copex Rapid and CMS 20 (in combination with their dedicated developers) in different formats to standard films in different formats.
Agfa Copex Rapid (CoRa) in 35mm surpassed APX 100, RPX 100, FP4+ in 6x6 medium format.
And CMS 20 35mm surpassed CHS 100 in 4x5".
We needed Acros 4x5" in the Arca Swiss to get better detail rendition compared to 35mm CMS 20.
And both CoRa and CMS 20 are available in 120, too.

Best regards,
Henning

 

[Henning Serger]

Hello Fabrizio,

That was one of the most interesting parts of your post. When analogue is compared to digital the comparison is always performed on the digital playing field, i.e. converting analogue to digital (by scanning) and comparing the results.

yes, that has been the major problem during the last 12 years in 99,9% of all these film vs. digital comparisons.
Almost all of these comparisons really have been digital vs. digital comparisons.
The digital capture medium scanner was compared to the digital capture medium camera sensor.
And the scans have been only max. 4000ppi scans in most cases. But a 4000ppi scanner can not resolve all the detail, which is in the film. Even the the best high-end drum scanners can not fully exploit the full potential of film.

One result of our numerous tests have been that you get the best detail rendition with optical enlarging with APO enlarging lenses and with slide projection (colour and BW slides).

A lot of these superficial and generalising "digital is better than film" statements are caused by comparisons of 4000ppi scans to sensors.
But film can do much better in the completely optical imaging chain.

If I look at 35mm slide projection with an excellent projection lens, then the quality is even surpassing digital medium format projected with a beamer. Because the imaging chain film+projection is much better than the digital imaging chain digital file+beamer. The beamer is the very weak part in this imaging chain and decreases the quality of the file to a big extent. From your 40,50, 60 or 80 MP only 1,2,4 or max 8 MP (most expensive gear) remain as beamer output.

Best regards,
Henning

 

[AgX]

Good point, Henning.

We should only compare things (pictures that is) as results of their own production line.
This does not say anything about the outcome though! And I don't start discussing issues like resolution etc.

 

[ooze]

This thread has been a delight to read. Thanks a lot for all the info Henning. It's always nice to be reassured that optical printing is the way to go.
I wish Spur HRX-3 was available in Turkey. Although part of me has stopped searching for a magic potion (good old ID-11 user here), the other part remains intrigued by potential improvements in neg quality.

 

[madgardener]

I have been enjoying reading this thread that I started. I'm learning a lot, even though the conversation has gone away from the topic. This discussion brings up another good question.

I know companies like Freestyle and Photo Warehouse are big buyers of film for their house brand. Is it conceivable that they might drive some innovation in the film market? While they don't manufacture film directly, they might have enough buying power to have an exclusive emulsion made someday?

 

[dwross]

I know the direction of the OP was for improvements or innovations to commercial film, but I'll throw my take on the subject into the hat. (Those of you who already know what I'm going to say -- no groans please)

I think a significant part of the future of film is R&D in the home darkroom, with the goal of making diy film, plates, and paper as easy and satisfactory and accepted as other alternative processes have become.

d

 

[Pioneer]

Roll Your Own

I know the direction of the OP was for improvements or innovations to commercial film, but I'll throw my take on the subject into the hat. (Those of you who already know what I'm going to say -- no groans please)

I think a significant part of the future of film is R&D in the home darkroom, with the goal of making diy film, plates, and paper as easy and satisfactory and accepted as other alternative processes have become.

d

I feel in the minority but I tend to agree with you. There are many on this forum who have convinced themselves that film will always be available, and they may be right. The problem will be the cost of film in my humble opinion. We are already seeing this and it is happening faster and faster. The major players have increased their costs. Smaller players are going out of business because their costs to produce are too high to recover in the current market. We are presently in the last stages of a vicious cost spiral that continues to climb.

Meanwhile digital is getting less and less expensive. Full frame cameras are getting cheaper. APS-C cameras are getting more and more features.

There will always be a few die-hard fans that will do whatever possible to continue shooting film. But to do that they will have to become more and more involved in the process. Just in the past 3 years it has become more and more difficult (and expensive) to get film developed. So serious practicioners are developing their own and printing their own, as much for cost containment as for control of the process. We are starting to move into the final stages where people will have to be prepared to either spend a lot of money on film and developing with the very few remaining manufacturers, or they will begin to make their own films, or they will have moved to another capture medium.

Many of those will move over to DPUG and continue to complain about how the old days when film was cheap and plentiful were so much better.

At some point it will become more cost effective, and give higher quality, to roll your own.

 

[Henning Serger]

I have been enjoying reading this thread that I started. I'm learning a lot, even though the conversation has gone away from the topic.

O.k., than back to the original topic:
At Photokina I've had a very long talk with Mr Boll from Fujifilm (Product-&Key Account Manager, Photo Imaging Products), and we've also discussed the situation in R&D. He said that there is still quite a lot of 'silver halide R&D' at Fujifilm, resulting in new products like the latest Eterna archival films, two new silver halide colour papers (for photobooks, an increasing market in Europe) just introduced at Photokina (I've seen them there), and the re-introduced Neopan 400 (they have solved the problem with the raw material which caused the former discontinuation).
Furthermore they absolutely need this R&D in anti-oxidants etc. for their cosmetic line Astalift.

Best regards,
Henning

 

[j.c.denton]

Henning, thanks a lot for the inside view. Is it valid to assume that this kind of R&D include benefits for either slide or negative film or can that be excluded? I am not very familiar with paper technology concerning how it corresponds to film technology.

Christian

 

[Henning Serger]

Henning, thanks a lot for the inside view. Is it valid to assume that this kind of R&D include benefits for either slide or negative film or can that be excluded? I am not very familiar with paper technology concerning how it corresponds to film technology.

Christian

Well Christian, progress in anti-oxidants for example is a field where film can benefit. They are used in films. And currently this film knowledge is indeed partly used in Fujifilms Astalift cosmetics. Perhaps we will see the opposite direction in R&D knowledge flow in the future. Maybe imaginable if the colour film market one time really has stabilised and then is growing again (well, the market for professional films is already in this stabilising process, but consumer CN film sales are still decreasing).

Best regards,
Henning