High accutance developers and resolution...

[Alan Johnson]

Sandy,
I guess it must be hard to see much difference as I thought the Xtol gave slightly higher resolution!
I wonder if any comparisons have been done for Pyro type developers.

 

[Photo Engineer]

Alan;

Try resizing the images so that they are identical in magnificaiton on-screen.

PE

 

[Chazzy]

Developers can be maximized for any two of the following: Grain, sharpness and speed. Normally, you cannot maximize all 3 but you can strike a happy medium.

Alan's experiment above is one of the best comparisons I have seen on APUG in spite of my comments. More like it should be done to "prove" a theory or observation.

PE

In the case of Xtol, I would say that it's a very happy medium indeed, with excellent grain, sharpness and speed.

 

[sanking]

Sandy,
I guess it must be hard to see much difference as I thought the Xtol gave slightly higher resolution!
I wonder if any comparisons have been done for Pyro type developers.

Alan,

Yes, two or three summers ago I started a very extensive project to compare pyro staining and non-staining developers. Part of that project involved contact printing the high resolution chrome on glass target that Ron mentioned. It took a lot of effort to set this up and get the right exposure and development that maximized resolution but I finally got it under control. I compared several films with two or three staining developers and two or three non-staining formulas. I used for sure PMK 1:2:100 and Pyrocat-HD 1:1:100 for the staining developers, and Xtol 1:2 and D75 1:1.

What I found was that PMK and Pyrocat both gave about 10% increase in resolution over Xtol and D76. However, I was never completely confident in the methodology since I could not get as much resolution out of the films as they are supposed to have. For example, Tmax-100 is supposed to resolve 200 lines per mm but in my tests I never got more than about 160-180 lines per mm. So I just threw in the towel on the project since in the end the greater resolution of the pyro staining developers, even if real, would not amount to much in practical terms since as your tests show, more than 100 lppm in the camera is a very high bar.


Sandy

 

[Alan Johnson]

Here are T-max 100 in Xtol 1+0 and in FX-1 resized.
Originally this was a test to see if the acutance developer produced what Pat Gainer called ,IIRC, the picket fence effect,where the lines from the acutance developer merged. TMX seems resistant to edge effects and I could not see any picket fence effect with this particular film.

 

[Kirk Keyes]

A very generous APUGger loaned me a set of chrome USAF targets and using Xtol and Acros, I think I was able to measure about 150 lpmm after a few trials. Like Sandy says, it's not easy, and I was still a ways from what Fuji says they get (200 lpmm).

 

[RalphLambrecht]

My results with T-max 100 film in Xtol 1+0 (solvent) and FX-1 (Acutance) don't show much difference in resolution.Photomicrographs of negatives of test chart, numbers on chart are resolutions in line-pairs per millimeter.
Lens was Canon ef 50mm f1.8 at f5.6.

Yes, this proves the point. Lens and film emulsion control resolution to the largest degree. Exposure and development can optimize it, but not radically change it. However, overexposure and overdevelopment can really mess it up!

 

[Photo Engineer]

Very nice.

Now, look at the apparent Dmin on the left vs the right. This has the effect of lowering the visual contrast thereby reducing the effective grain and effective sharpness of the image on the left. If you have a scanner utility that lets you set the white level of both of these images, then you might see another revealing change.

This, in analog terms is no more than changing the exposure level to make white white. You can then do the same to black by changing contrast, the same as if you were using an MG paper. This will then normalize the two images. Keep at it and we will have a near perfect (as far as can be done with these images which are quite good) test.

PE

 

[RalphLambrecht]

I do own a copy.

While it explains the difference (which is rather obvious) between accutance and resolution; it does not present any useful information on how developer type affects resolution.

While Bill Troop is a photochemist, the primary author (Steve Anchell) is not.

Because developers don't significantly affect resolution.

 

[RalphLambrecht]

Not entirely. Digital has a linear grid for the 'dots'. Film grain moves with the light and processing, thus is uneven in it's distribution. This gives the blacks in film a much richer look and the greys in digital a much more linear graduation.

Depending on the shot, film and digital will show different aspects of the lens. The same, yet different.

Well, I'm not so sure about linear grids or about linear gradation, but I love film, and this is analog forum anyway. The point only was that digital concentrates on contrast rather than resolution for one reason... it is subjectively more important to sharpness than resolution.

 

[RalphLambrecht]

Developers can be maximized for any two of the following: Grain, sharpness and speed. Normally, you cannot maximize all 3 but you can strike a happy medium.

Alan's experiment above is one of the best comparisons I have seen on APUG in spite of my comments. More like it should be done to "prove" a theory or observation.

PE

I agree with that, but I think in this statement 'sharpness' mostly means acutance and less so resolution.

 

[RalphLambrecht]

But is not the lens the limit of resolution in Alan's test, not the developer? I have done tests by contact printing high resolution chrome on glass targets and with Tmax-100 I was able to get somewhere around 160-180 lines per mm with pyro staining developers.

Sandy

Sandy

What lens did you use to do this test?

Resolution measurements are tricky without an optical bench, because the system resolution always is a combination of the individual resolutions of film and lens, and of course, also of the resolution of your measuring optics (loupe, microscope), not to mention your MTF cut-off criteria. In other words, what contrast between the lines do you accept to call it resolved?

The following equation is usually applied:

1/R^2 =1/r1^2 + 1/r2^2 + 1/r3^3 + 1/rn^2

It is sobering to see, for example, that the combination of a film and lens, each with a resolution of 100 lp/mm, limits the overall performance to just 71 lp/mm.

 

[sanking]

Sandy

What lens did you use to do this test?

Resolution measurements are tricky without an optical bench, because the system resolution always is a combination of the individual resolutions of film and lens, and of course, also of the resolution of your measuring optics (loupe, microscope), not to mention your MTF cut-off criteria. In other words, what contrast between the lines do you accept to call it resolved?

The following equation is usually applied:

1/R^2 =1/r1^2 + 1/r2^2 + 1/r3^3 + 1/rn^2

It is sobering to see, for example, that the combination of a film and lens, each with a resolution of 100 lp/mm, limits the overall performance to just 71 lp/mm.

Ralph,

No lens, I was contact printing to take optics out of the equation since I just wanted to test resolution. Even in contact printing the resolution figures varied a lot depending on whether the light was diffuse or point source. If I ever do this again I will try to find out what type of light is used, and at what distance, because ultimately I felt this was the biggest wild card in the equation.

I used a microscope at about 40X to examine the bar lines. There is as you know some subjectivity in this.


Sandy

 

[RalphLambrecht]

Ralph,

No lens, I was contact printing to take optics out of the equation since I just wanted to test resolution. Even in contact printing the resolution figures varied a lot depending on whether the light was diffuse or point source. If I ever do this again I will try to find out what type of light is used, and at what distance, because ultimately I felt this was the biggest wild card in the equation.

I used a microscope at about 40X to examine the bar lines. There is as you know some subjectivity in this.


Sandy

Thanks for the details of your test. I wonder if you get into the wavelength of test light making a difference with these high resolutions.

In any case, in my tests, I try to take some of the subjectivity out by taking a digital photograph through the microscope and then normalize the exposure in Photoshop. Then, I take the eye-dropper tool to find the line pair with a 10% difference in 'density' between a black and white bar, and use that for the resolution calculations.

I also think that PE is correct in saying that all test taken with the same lens eliminate the effect of optics and are valid. The benefit of that approach is also that it makes the test and the results more realistic. Who cares about a 200 lp/mm performance if you can only get it from contact printing a $500 resolution target? OK to compare emulsions or developer effects, but how much is left of the material difference after a camera lens took care of it?

 

[wogster]

I agree with that, but I think in this statement 'sharpness' mostly means acutance and less so resolution.

I don't know if it's so much sharpness, as perceived sharpness, if you look at two images, the one that is measurably sharper may not actually look sharper. A brighter and higher contrast image can look sharper then an image that is darker and lower contrast, even though the second image may in fact be measurably sharper.

 

[RalphLambrecht]

I don't know if it's so much sharpness, as perceived sharpness, if you look at two images, the one that is measurably sharper may not actually look sharper. A brighter and higher contrast image can look sharper then an image that is darker and lower contrast, even though the second image may in fact be measurably sharper.

Exactly!

 

[sanking]

Thanks for the details of your test. I wonder if you get into the wavelength of test light making a difference with these high resolutions.

In any case, in my tests, I try to take some of the subjectivity out by taking a digital photograph through the microscope and then normalize the exposure in Photoshop. Then, I take the eye-dropper tool to find the line pair with a 10% difference in 'density' between a black and white bar, and use that for the resolution calculations.

I also think that PE is correct in saying that all test taken with the same lens eliminate the effect of optics and are valid. The benefit of that approach is also that it makes the test and the results more realistic. Who cares about a 200 lp/mm performance if you can only get it from contact printing a $500 resolution target? OK to compare emulsions or developer effects, but how much is left of the material difference after a camera lens took care of it?

The only step I took with regard to the wavelength was to balance the exposing light to daylight.

I agree that practical tests for resolution made in the camera give a more realistic indicator of the capability of the film/camera/lens system. However, if the weak link in the chain is a lens that can not resolve more than 100 lpm a test for resolution with developers will not reveal any useful information about resolution if all of the film/developer combinations tested are capable of more than 150 lpm, which was the case in my testing.

Ultimately perceived sharpness is more important than any other single factor, as has been pointed out. I have made prints 12X18" in size from a Canon G9 (7mm wide sensor) that look sharper, at viewing distance of 10-12", than some prints I have made from 5X7 negatives that have a huge amount of detail.

BTW, I print with a hybrid process. Either film or digital capture, mostly the former, correction in Photoshop, printing a digital negative, and finally a carbon transfer print. The file is adjusted with a curve to give a digital negative with a totally linear output so that what I see on screen is pretty much what I get in the final print, regardless of whether the original capture is film or digital.


Sandy King

 

[aldevo]

Aldevo,

I think you have it exactly right, except that D-76 is considered a moderately fine grain developer rather than an accutance developer. For the best resolution, you probably don't want the highest accutance developer. Best resolution is usually acheived with a lower contrast light-to-dark transition, whereas highest accutance is acheived with a higher contrast light-to-dark transition.

Charlie Strack

Understood. It's my understanding that D-76 is not heavily-reliant on the so-called solvent effect to achieve fine grain. Earlier fine grain developers, such as those using PPD, were heavily reliant on the solvent effect.

Your explanation of resolution vs contrast is exactly what I would expect. Unsharp masking in Photoshop is probably nothing more than accentuating local contrast gradients around areas that meet some candidate contrast threshold.

 

[aldevo]

I agree with Ron. These results are technical and can usually be measured. Unless it's measured it's subjective and fairly useless for anybody else.
I do feel that people put too much energy and thought into film developers. Pick one and work with it. By varying your technique, evaluating your results, and above all - printing your negatives often, you learn a lot.
The other week I shot two rolls of Kodak TMY (the old kind) using a 35mm camera. I processed one roll in Rodinal and the other in Xtol. According to what everybody claims they observe with these two developers, the difference in grain should be huge. In reality, in the resulting 11x14 prints, the difference (to my eyes) is really negligible. Sharpness was very similar too. I used a Rodenstock Rodagon enlarging lens, which is an OK lens, so no sub-par equipment either.

What I'm getting at is that if you're truly interested in resolution, sharpness, grain, acutance and all that, at least make an evaluation based on how it makes a difference for you in practical terms before you take it too far. I found, by comparing prints of the normal size I make them, that I don't really care whether I get the look of the Rodinal or the look of Xtol. As an artist the difference between them is not even close to feeling important, that's how subtle the difference is. It may be different for you, but it's worthwhile comparing prints before you get overly concerned with which film developer you use.

- Thomas

Great points, all.

I've finally learned not to obsess about such things because I never enlarge my negs beyond about 9x and, most frequently, never beyond 8x. The differences in sharpness & grain between developers at relatively modest enlargements will not be significant.

Shifting gears a bit it certainly appears that the limits of resolution are only attainable under highly-optimized conditions that can only be realized in a heavily-controlled environment. We never realize such conditions in the field.

 

[wogster]

Your explanation of resolution vs contrast is exactly what I would expect. Unsharp masking in Photoshop is probably nothing more than accentuating local contrast gradients around areas that meet some candidate contrast threshold.

That's exactly how it works, it's tough to get it right though, there are lots of digital images that have been over sharpened.

 

[RalphLambrecht]

The only step I took with regard to the wavelength was to balance the exposing light to daylight.

I agree that practical tests for resolution made in the camera give a more realistic indicator of the capability of the film/camera/lens system. However, if the weak link in the chain is a lens that can not resolve more than 100 lpm a test for resolution with developers will not reveal any useful information about resolution if all of the film/developer combinations tested are capable of more than 150 lpm, which was the case in my testing.

Ultimately perceived sharpness is more important than any other single factor, as has been pointed out. I have made prints 12X18" in size from a Canon G9 (7mm wide sensor) that look sharper, at viewing distance of 10-12", than some prints I have made from 5X7 negatives that have a huge amount of detail.

BTW, I print with a hybrid process. Either film or digital capture, mostly the former, correction in Photoshop, printing a digital negative, and finally a carbon transfer print. The file is adjusted with a curve to give a digital negative with a totally linear output so that what I see on screen is pretty much what I get in the final print, regardless of whether the original capture is film or digital.


Sandy King

That makes sense to me because it reflects my experience as well.