Resolution of color media

[warrennn]

I admit that I am fairly obsessive about resolution and purchased a medium format system with this in mind. I recently tested my 80 mm lens using a USAF resolution test chart and obtained the expected ~70 line pair per mm. I used Tmax 100 film to make the exposure and bracketed the focus and, of course, used a good tripod.

I recently did the same test using Velvia 50 and found that the resolution was one half that of the Tmax film. I also noticed that my 8x8 RA-4 prints (Fuji CA) seemed slightly fuzzy and used the B&W negative of the USAF chart to test the resolution of the paper. I bracketed the enlarger focus to convince myself that the problem was not a mis-calibrated grain focuser (also I had two of the latter and they both agreed with each other).

Looking through the grain focuser, I was able to see all of the patterns that I could see by examining the negative with a microscope. Thus, the enlarger lens (an el-Nikkor 80 mm f/4) and enlarger light source (dichroic, i.e. diffusion) were not at fault.

The results were that the paper seemed to have a resolution which was about 5 times worse than the tmax film for 8x8 prints (this would be less of a problem with larger prints). Are these resolution degradations the usual case with color media? If so, I wonder why I go to the trouble of using MF for color, since I would imagine that a 35 mm system would do as well.

Any suggestions would be welcome. Thanks.

Warren Nagourney

 

[AgX]

Negatives and slides are typically enlarged today. So they should yield a resolution that makes that possible.

Papers are typically not enlarged. The most critical case would be looking at them with blank eyes at minimm viewing distance.
If they yield enough resolution to match those of the eays (or twice that...) it's okay.

Though there are theories that even higher resolutions are beneficial to the perception.

 

[Film-Niko]

Hi Warren,

from my own tests and experiences:

Resolution of T-Max with my best 35mm prime lenses at lower to medium contrast (2 - 3 stops): 140 - 155 Lp/mm.
Resolution of Velvia 100 with the same lenses: 130 - 150 Lp/mm.
Resolution of Velvia 50 with the same lenses: 120 - 130 Lp/mm.
Velvia 50 has a bit lower resolution than 100 because of significantly coarser grain.
Resolution of Ektar 100: 90 - 100 Lp/mm. Fuji Reala gets 110 Lp/mm.

With my best MF primes I got 90 - 110 Lp/mm with TMX and 85 - 100 Lp/mm with Velvia 100.
A friend of mine got with TMX 130 Lp/mm with his 80/f4 mm Mamiya Sekor at his Mamiya 7. Marvellous glas.

I have never seen a difference in resolution between Ilfochrome with my slides and Ilford Multigrade with my BW prints. I am using APO enlarging lenses.
But I have not tested RA-4 paper, because I don't use RA-4 in my own darkroom.

 

[warrennn]

Thanks for your replies.

From my tests of a 2.25x2.25 negative blown up to 8x8 (magnification of ~3.5), it would seem that there will be loss of resolution for small enlargements (less than or equal to 8x8) and one should therefore use at least an 11x11 print to avoid this. I received a PM which concurred about the much lower resolution of papers (and incidentally suggested that an APO enlarger lens is a waste of money).

I am wondering why my Velvia slides didn't do better -- their measured resolution was about 35 lpm, which is much lower that Film-Niko's 125 lpm. I confess that I was using film that was about 2 years out of date and that I developed it myself (my first attempt at E6), but the colors seemed perfect, so I wouldn't necessarily expect processing errors. I also wouldn't expect that out of date film would have lower resolution, but I might be wrong.

Cheers,

Warren N

 

[Photo Engineer]

Since Velvia is a reversal slide film, and RA4 is intended to make prints from negative C41 films, how did you do the print process? Did you make an internegative? Did you cross process the paper?

This would help and perhaps explain the problem.

PE

 

[warrennn]

The prints were made from a B&W negative (tmax 100) which was taken of a USAF resolution chart -- the colors would be wrong of course, but the resolution test should still be meaningful (the chart had 100% contrast). I was prompted to test the paper resolution since some of my prints seemed a little fuzzy.

The slides were of course a reversal film. My question was about possible resolution degradations in color media, hence I talked about both RA-4 and E6 with no intention to connect them in any other way.

Warren N

 

[Photo Engineer]

Ok, then here is a possible explanation for you.

Color paper has the cyan layer on top and the yellow layer on the bottom. Due to multiple internal reflections, the cyan layer is like an object on top of a mirror with some plastic between. These reflections cause degradation of the cyan layer and the human eye is most sensitive to cyan and magenta images as far as sharpness is concerned.

To alleviate that, Kodak puts a cyan and magenta dye into the paper to absorb back reflection and to improve sharpness. These dyes are called acutance dyes. That gives the Endura papers their blue purple color. IDK what Fuji does.

In any event, expect color prints to be less sharp than B&W prints. The acutance dyes are not perfect and the thicker layers and spread of dye image clouds contribute to degradation of sharpness.

OTOH, the papers are adjusted to give roughly equal apparent image sharpness at normal viewing distance, such that as you enlarge a print, you view it from a greater viewing distance.

PE

 

[warrennn]

Thanks, I should make 11x11 prints to prevent the paper resolution from being the limiting step.

Do you have any idea why an E6 transparency would have much lower than expected resolution?

Warren N

 

[Photo Engineer]

Look on the Fuji and Kodak web sites. They show resolution I believe as part of the product specifications. Maybe that would help. Just remember thought that the same limitations of 3 layers and turbidity apply to all color products.

PE

 

[AgX]

Look on the Fuji and Kodak web sites. They show resolution I believe as part of the product specifications.

No, they don't.The only datashests that typically gave that information were those for the Polacolor materials. But of course due the technique behind them they were out of the game anyway.

 

[Photo Engineer]

Kodak gives MTF which has a direct correlation with sharpness or resolution.

See here, Page 18: http://www.kodak.com/global/en/professional/support/techPubs/f4016/f4016.pdf

and here: http://www.normankoren.com/Tutorials/MTF.html for some interesting information.

PE

 

[AgX]

Thanks, I know that Koren tutorial It is very informative, but I haven't looked at it for a while as I have to admit with a large pile of books and a the computer filled with files it is sometimes hard to decide where to look for something.

However it does not give listing of colour paper resolutions as far as I remember.
And that first link leads me to the T-max data-sheet.

 

[Photo Engineer]

All of Kodak data sheets for all film have MTF data. No data is reported nor has ever been for ANY paper AFAIK, as the reading of reflection MTF is very very difficult and fraught with possible error due to viewing angle and other factors including reflectance substrate (RC vs Baryta for example). It is also tied up with magnification of the image and viewing distance whereas film values are an absolute (relatively speaking ).

PE

 

[warrennn]

Thanks, it turns out that an MTF curve is available for Velvia 50 from Fuji's data sheet. The curve doesn't go above spatial frequencies of about 60 lpm (line pairs or cycles per mm). I used Koren's procedure of parametrizing the curve and duplicated the curve provided by Fuji with the frequency axis extended to 200 lpm.

Converting from the USAF chart to an MTF is very risky, since the selection of the pattern which is considered barely resolvable is a very subjective matter and usually corresponds to an MTF (contrast) of from 2% to 5%. Taking 5%, the USAF chart-derived resolution of Velvia 50 is about 140 lpm. This is 4 times what I observed in my test.

Warren N

 

[Photo Engineer]

Did you use a vacuum frame or pressure frame to hold the film against the chart?

Was it the evaporated metal USAF chart?

Those questions are also important in this evaluation. It is not clear how your test was actually performed.

PE

 

[warrennn]

My procedures were not very sophisticated. I used a cardboard chart which I purchased from Edmund Scientific and photographed it outside on our deck from a distance of about 10 feet. I have used the same technique on a number of occasions in the past to evaluate my lenses and obtained the expected resolutions of about 70 lpm. I should do it again with fresh film and using more careful procedures.

Thanks for the suggestions.

Warren N

 

[Photo Engineer]

Well, I can probably solve part of your problem then.

The data reported by Kodak and Fuji are for the film, in vacuum contact with a metal deposited chart emulsion to metal. Your data includes the MTF of the lens since you used a camera. So, you need OTF (O = Optical) in the mix.

PE

 

[warrennn]

I believe that the resolutions add reciprocally (like resistors in parallel), so my observed resolution should be 1/(1/70+1/140)=47 lpm, not too far from what I observed. Thanks!

(MTF vs frequency functions are multiplied, like any transfer function.)

Warren N

 

[Athiril]

Thanks, I should make 11x11 prints to prevent the paper resolution from being the limiting step.

Do you have any idea why an E6 transparency would have much lower than expected resolution?

Warren N

The low-contrast figures are the most useful resolution, in the case of Velvia which has the highest it is 80 lp/mm iirc, about the same as the 30D and 5D II are capable of at maximum given good lenses to put in perspective if that helps.

High-contrast resolution is useful for some things, but you should base all yuor expectations on the low-contrast resolution, which is what you'll see across the texture of an object etc.

In which case, C-41 is up there with E-6 in low-contrast resolution iirc from looking at the Fuji specs.

 

[Film-Niko]

Thanks, I should make 11x11 prints to prevent the paper resolution from being the limiting step.

Do you have any idea why an E6 transparency would have much lower than expected resolution?

Warren N

Hi Warren,

as I have explained in my first post, resolution of Velvia at medium contrast is a little bit lower compared to TMX.
At low contrast of about one stop (1:2) Velvia 100 and Velvia 100F even have higher resolution than TMX. TMX is rather grainy compared to the extremely fine grained Velvia 100, Provia 100, E100G, Astia 100F and so on.
At these low contrast values the resolution of TMX is limited by the much coarser grain, the ISO 100 slide films all give much better detail as TMX at this low contrast.

So, there must be a test mistake in your set-up.

 

[Film-Niko]

Kodak gives MTF which has a direct correlation with sharpness or resolution.

See here, Page 18: http://www.kodak.com/global/en/professional/support/techPubs/f4016/f4016.pdf

and here: http://www.normankoren.com/Tutorials/MTF.html for some interesting information.

PE

Hi,

the Norman Koren website is not helpful for film photographers searching for realistic resolution data.
He has some big faults in his tests. For example are his resolution conclusions based on 4000 dpi scanned films.
But a 4000 dpi scan is not at all capable of getting all the film detail.
We have tested some scanners, with the 4000 dpi scanners we've got only half
of the real film resolution.
A 140 Lp/mm resolution of Velvia 100F was reduced to only 65-70 Lp/mm with a Nikon Coolscan 9000 4000 dpi scan.

The 140 Lp/mm visible under the microscope was only minimal reduced to 130 Lp/mm with slide projection, evaluating on the screen, with Leica Super-Colorplan P2 projection lens.

 

[Film-Niko]

I believe that the resolutions add reciprocally (like resistors in parallel), so my observed resolution should be 1/(1/70+1/140)=47 lpm, not too far from what I observed. Thanks!

(MTF vs frequency functions are multiplied, like any transfer function.)

Warren N

Warren,

first, this formular is not suitable for an exact calculation, it is more a "rule of thumb".
Second, your value of 70 for the lens is much too small. Here you have to use the value of the aerial resolution of the lens (good primes are in the 150 - 400 Lp/mm range at f5,6 with white light). It looks like you have taken the value of lens + TMX, wich is already a system resolution, therefore you got a wrong result.

 

[Film-Niko]

The low-contrast figures are the most useful resolution, in the case of Velvia which has the highest it is 80 lp/mm iirc, about the same as the 30D and 5D II are capable of at maximum given good lenses to put in perspective if that helps.

Fujis data is 80 Lp/mm at low contrast of only 1,6:1.
But warrens test target has probably an higher contrast. Most printet test charts are in the 1:4 - 1:10 range, two to about three stops. By the way an object contrast, which you can find in nearly all scenes. Therefore valid for normal photography.

Below you can find some of my test results (see my first post above) with film at this medium object contrast. Under the same test conditions our result with 24 MP FF digital was 75 Lp/mm.

"From my own tests and experiences:

Resolution of T-Max with my best 35mm prime lenses at lower to medium contrast (2 - 3 stops): 140 - 155 Lp/mm.
Resolution of Velvia 100 with the same lenses: 130 - 150 Lp/mm.
Resolution of Velvia 50 with the same lenses: 120 - 130 Lp/mm.
Velvia 50 has a bit lower resolution than 100 because of significantly coarser grain.
Resolution of Ektar 100: 90 - 100 Lp/mm. Fuji Reala gets 110 Lp/mm.

With my best MF primes I got 90 - 110 Lp/mm with TMX and 85 - 100 Lp/mm with Velvia 100.
A friend of mine got with TMX 130 Lp/mm with his 80/f4 mm Mamiya Sekor at his Mamiya 7. Marvellous glas."