Measuring film resolution

[Photo Engineer]

Imaging direct by lens is best to give overall performance. Subsequent measures of film and of the printing process shows the weak points. Is it camera, film, process or printing?

PE

 

[Bill Burk]

I have such a target, and will try to implement Bill Burk's idea of repeated resolution patterns which follow the step pattern on my Stouffer wedge. Having such a special resolution target would allow me to see (and measure) density, grain and resolution versus exposure in one single test clip, which gives me much more reliable results than determining the three properties from three separate test strips.

Since we have reason to believe that higher contrast yields better resolution unless grain grows out of control, I think I will try to create this resolution strip with repeated patterns on Fuji Velvia 50. Not only is this film plenty sharp enough to make a resolution target for my purpose (Tri-X, HP-5+, Delta 3200), it also saves me the second copy step or the (very expensive) inverted target.

I'd hoped PE would agree with the plan but his estimate is the duplication of the target would degrade it beyond usefulness even for this purpose.

My initial thought of using Velvia: It's a dye based result, and totally transparent to infrared. Thus E6 might be less suited to the purpose of making a sandwich with sensitometric strip... I could be wrong about this, because you really are looking at high resolution - high contrast signal and only looking to see if the signal exists.

Since you get good results already with a camera, maybe the test device you make could be a camera-based concept: Like maybe a half-frame SLR (e.g., Olympus Pen) with macro bellows and a suitable original target, where you expose at exact same f/stop for all... but vary the exposure time (or if using flash, flash intensity). Using a half-frame camera would give you more exposures per roll.

Another idea would be to make a special purpose-built optical printer device that only advances the film two or three sprockets per exposure, again a camera-type device so that you can reduce from a master target.

 

[Rudeofus]

Still not sure why you want to contact print sources to the films, rather than image direct by lens?

My biggest concern about having a lens between resolution target and film test clip is flare. Look at the two attached images of a Stouffer wedge printed onto B&W paper, one was contact printed, the other one through a Schneider Kreuznach lens. The enlarger was set up in such a way that both prints received the identical amount of light in the transparent section of the wedge.

 

[Rudeofus]

My initial thought of using Velvia: It's a dye based result, and totally transparent to infrared. Thus E6 might be less suited to the purpose of making a sandwich with sensitometric strip... I could be wrong about this, because you really are looking at high resolution - high contrast signal and only looking to see if the signal exists.

Since none of the films I want to evaluate are sensitive to IR light, transparency or opaqueness to infrared light is irrelevant.

Since you get good results already with a camera, maybe the test device you make could be a camera-based concept: Like maybe a half-frame SLR (e.g., Olympus Pen) with macro bellows and a suitable original target, where you expose at exact same f/stop for all... but vary the exposure time (or if using flash, flash intensity). Using a half-frame camera would give you more exposures per roll.

I have no plans of using a camera for this purpose, and ideally not even a lens between target and film test clip. As Michael writes, contact printing is the way to go. It's time to put that engineering degree of mine to use ...

 

[Photo Engineer]

Here are 4 examples of contact prints.

1. Handcoated on baryta
2. Ilford MC check
3. Handcoated on Strathmore smooth
4. Handcoated on Strathmore watercolor

This is a negative chart only test and shows how surface affects resolution of a high contrast chart being contact printed using a heavy glass pressure plate.

PE

 

[Rudeofus]

Ron, is there a chance you could post higher resolution images? At the moment resolution seems to be limited by pixelation, not by the actual medium.

 

[Photo Engineer]

Look in the book for higher resolution neg and pos images. I can re-scan, but the first scans (that I liked) would not upload due to size.

PE

 

[RobC]

Dead Link Removed

 

[Alan Johnson]

Rudi,
I do not think that flare would be a problem using a lens.Here is a description by Kornelius Fleischer of how he did tests for Zeiss:
http://photo.net/leica-rangefinders-forum/009t9L?start=10

 

[Rudeofus]

Rob, anyone with computer skills can create such a resolution test chart from scratch, getting it onto a real medium with the desired resolution is the trick that few companies can perform - for good money.

Ron: assuming you refer to figures 61-64, the printed pages suffer from the same issue as the scans you posted here: resolution of tested medium by far exceeds resolution of presentation medium. It would be very helpful if you could post a cropped version of these images in full resolution, e.g. the groups starting with 7.1.

 

[RobC]

Rob, anyone with computer skills can create such a resolution test chart from scratch, getting it onto a real medium with the desired resolution is the trick that few companies can perform - for good money.

Ron: assuming you refer to figures 61-64, the printed pages suffer from the same issue as the scans you posted here: resolution of tested medium by far exceeds resolution of presentation medium. It would be very helpful if you could post a cropped version of these images in full resolution, e.g. the groups starting with 7.1.

Well that chart printed and then photographed at a correctly calculated distance would do it except it will be limited by lens and film development. BUT according to zeiss, 400 lp/mm is possible on film using some of their lenses. So I figure that using a film such as CMS 20 and a high quality lens you should be able to do it.

If its a question of making a glass target for contact printing then yes you would need specialist equipment to make it, hence the cost of them.

 

[Rudeofus]

I do not think that flare would be a problem using a lens.Here is a description by Kornelius Fleischer of how he did tests for Zeiss:
http://photo.net/leica-rangefinders-forum/009t9L?start=10

Flare is obviously not an issue if you start with a 50:1 contrast ratio, like Mr. Fleischer claims he does. It turns into a huge issue if you start with 1000:1 contrast ratio, as exemplified by my test prints, and as I would likely encounter if I superimpose a Stouffer wedge with a repeated resolution test pattern.

 

[Rudeofus]

Rob, that 400 lp/mm lens that Zeiss is so proud of won't fit on my cameras, and the lenses that do give me substantially less resolution. We should also take another thing into account: a lens, that does 400 lp/mm, does not convert a 400 lp/mm 50:1 contrast pattern into a 400 lp/mm 50:1 contrast on the test medium. In fact I suspect that Mr. Fleischer checked up to which resolution he could discern line pairs under the microscope and recorded that number as limit resolution. That's an important number if you want to show what kind of detail one can resolve in a photograph, but it's not all that relevant if you want to copy a resolution test target.

 

[Photo Engineer]

Ron: assuming you refer to figures 61-64, the printed pages suffer from the same issue as the scans you posted here: resolution of tested medium by far exceeds resolution of presentation medium. It would be very helpful if you could post a cropped version of these images in full resolution, e.g. the groups starting with 7.1.

Rudi, it is a matter of rescanning the data. I have so many backed up projects it will take me a good long while to get to it. I have 2 projects at GEH going, the new book(s), and a host of other stuff that never even shows up here on APUG. So, be patient.

PE

 

[Bill Burk]

PE you could just tell us the numbers you can read. I'll take your word for it.

 

[Photo Engineer]

Bill, you see the examples, well the positive results are the same but different. And on my screen, in hi-res, they look fine. Judge for yourself.

PE

 

[Prof_Pixel]

Ron,

Looking at a JPEG file is meaningless. Because of the way JPEG compression divides the image into 'super pixels', I can start to see alaising in the 4.0 bars of the Ilford print.

 

[Athiril]

Diffraction pattern may work for you.

 

[Photo Engineer]

Ron,

Looking at a JPEG file is meaningless. Because of the way JPEG compression divides the image into 'super pixels', I can start to see alaising in the 4.0 bars of the Ilford print.

I recognize that Fred, but I wanted to post something that was at least viewable. I guess I did more harm than good, but I thought that the overall losses in quality might be useful.

In any event, I was unsure as to what format to use for the image when scanned and I am limited in file size for uploading. I was not sure of what resolution to scan either. Maybe we can talk about this sometime soon. I plan on being at GEH on Wednesday.

In any event, it may be a while before I can get to it with all of the other things on my plate.

Thanks.

PE

 

[RobC]

the PDF I posted a link to can be zoomed into (max 3600% in acrobat) and the detail is there. What you can print that to is another matter.

For those that don't know, all a PDf file is, is a file containg "PostScript" language and data which for the purposes of this document contains vector data for position and dimensions of shapes. Acrobat is just a rendering engine for PostScript landguage/data.

With the correct software you can write your own postscript to produce any graphic you like at whatever size you like but the above attached pdf already has it down to a certain size.

As has already been determined, unless you have a device capable of outputting that at the size you want, you are stuffed.

 

[RobC]

See http://www.danes-picta.com/

look in Pre-Press at bottom left. There's a microfilm good for 250 lp/mm but how much you lose in contact printing itself I don't know.

Only Euro 67 I think.

 

[Prof_Pixel]

For those that don't know, all a PDf file is, is a file containg "PostScript" language and data which for the purposes of this document contains vector data for position and dimensions of shapes. Acrobat is just a rendering engine for PostScript landguage/data.

Yes, but....

An Acrobat file can also include non-vector data bit mapped image files in a variety of formats (such as TIF and JPG). An Acrobat file is simply a 'wrapper' that can contain data in a variety of formats.

 

[RobC]

Yes, but....

An Acrobat file can also include non-vector data bit mapped image files in a variety of formats (such as TIF and JPG). An Acrobat file is simply a 'wrapper' that can contain data in a variety of formats.

I'm not interested...

 

[Rudeofus]

the PDF I posted a link to can be zoomed into (max 3600% in acrobat) and the detail is there. What you can print that to is another matter.

As has already been determined, unless you have a device capable of outputting that at the size you want, you are stuffed.

This reminds me of my childhood when I copied a circuit diagram from a book and then declared "now I have a 120W audio amplifier". Seriously, there are many companies out there which can create unbelievable fine patterns on glass, as anyone in the semiconductor industry can confirm, but they do it for a lot of money. BTDT.

The Danes test patterns are nice, but even further removed from MTF measurements than the Lasersoft targets.

 

[Photo Engineer]

I have a high end scanner that can scan at very high resolution, but the files created are huge and cannot be uploaded to APUG as I noted before. I can also output them in several different formats.

My regular charts are 4x5 on very thin film.

I also have the 35mm size USAF resolution charts made of metal on glass. They cost a pretty penny.

PE