Top-end digital concerns

[bglick]

> This guy prints only from jpg's files and only @150 dpi from 8x10'' to up to 40''x120'' claiming that where is no visible difference to printing larger than 8x10 from tiff's over jpg's files and 300 dpi over 150 dpi.


Sheeesh, lots of issues in your post...but I will address this particular one... 150 dpi = 3 lp/mm, assuming no losses in the printing process. If you feed the printer 300 dpi, there "should" be a gain in printing resolution.... but, this is where details come into play....

1) this assumes, the operator is printing at high enough dpi to lay down the added data

2) this assumes the paper and ink combination can exceed recording 3 lp/mm per mm... many non coated fine papers can NOT!

3) this assumes, you are viewing at normal distances, say 30", 3 lp/mm on paper is already a high standard, as its beyond what the eye can resolve at that distance.....more resolution would not easily be discernable, unless there was extreme lighting on the print.



> My dream is to have 8x10 digital back in size of 8x10 film holder: one side is 8x10 size sensor another side is 8x10 1600x1200 touch screen.

If you followed this thread, you will learn, a sensor this size will be greatly limited by apt. diffraction due to the long fl and higher f stops used on 8x10... its in conflict with the laws of imaging, hence why we will never see this product.... However, a 8x10 scanning back would be interesting....but this requires static subjects, in which case, its easier to stitch with off-the-shelf DSLR, better results and less back pain....

[bglick]

Moving this thread to the DOF / print size issue Sheldon asked about.... which answers the "what format" question we all struggle with... regardless whether its a purchase decision, carry decision, cost decision, its a major issue for all landscape / architectural photogs...


First the disclaimers.... many people like larger formats for the big ground glass, the slow technique of view cameras, the beauty of a Big Honkin Ebony 8x10 sitting on a tripod, etc. This discussion does not address any of those likes / dislike issues. It only pertains to how DOF and its associated diffraction effects OFR (On Film Resolution) and final print sizes.


The recorded MP's of resolution described previously were ideal in comparing formats of all different sizes / types. However, the MP's value is a bit deceiving for some, as ultimately, it's the final print size that matters. Since it takes 4x the recorded MP to make a print 2x as large, while holding the SAME OPR (On Print Resolution), the recorded MP's doesn't deliver the ultimate information needed for those who make prints as the final product. So, I will express how the effects of 1/R at different format film sizes, will effects the max. size final print, at a given resolution. This assumes regardless of the size of the print, it will be viewed at the same distance, which today, seems to be commonplace as everyone has become resolution obsessed. Till the prints become 60" +, whereas often you are forced to view at a greater distance to fill your field of view to "take in" the entire print. Let's use an aggressive 5 lp/mm to the printer... with printer losses, you may yield 3 - 4 lp/mm to the final print... still a VERY high standard!


I will use 3 formats, 4x5, 8x10 and the little gem previously mentioned, Mamiya 7. Lets use color film at an avg. MTF of 60 lp/mm (expect better results with B&W). Format sizes..., I will use 4x5 at 90x120 and 8x10 at 180 x 240 to keep these two at exactly 2x, or 4x area. Therefore, for same composure and DOF, you always double the fl for 810 vs. 45, and also double the f stop, pretty simple.

The M7 will have to stand on its own unique size, not conforming to the 2x vs. 4x5 rule. But since it has a 1.23 aspect ratio, nearly identical to 4x5's 1.25, it makes for an ideal comparison. Since all the aspect ratios are the same, I will express the print size on the horizontal dimension only for all films. With a 89mm film diag. on the M7, the increase in fl to 4x5 is reduced to a 1.7x factor, as well as the increase in f stop.


And of course, we are making the assumption your scanner can extract all the data from all the formats. This post also assumes you understand basic DOF premise, as it relates to format size...i.e. same composure with different format sizes requires proportional increase in f number, vs. lens fl increase.....which will always yield the SAME resolution at EQUAL size final prints. (all else being equal)


First, we display LF in all its GLORY -

SCENARIO #1 - Infinity focus or flat plane subjects....

M7 = f4 1/R = 52 lp/mm

52 * 69mm film / 5 lp/mm print / 25.4
= 28" print on Horiz.


I won't repeat the math from here on...

45 = f11 (a lens optimized at f11) = 39" Horiz.

810 = f11 (for illustration purposes) = 79" Horiz.


Although there is only a few lenses on 810 that have great performance, I used it to demonstrate the expected 2x increase in print size when doubling the format. You can adjust numbers accordingly, maybe f16 or f22 is more realistic with the lenses you use. Anyway, the obvious message here is..... with NO DEPTH, LF film still rocks the imaging world, nothing compares - SLAM DUNK!


Next, the sad reality of the DOF, and how the imaging laws rear their ugly head for scenes with depth. I will offer a Mild DOF scene scenario and an extreme DOF scenario, as the numbers vary greatly. Each time, always increasing the f stop by the same factor the lens fl increased.... Since the M7 lenses are diffraction limited at low f stops, we can use their true values in 1/R.... as well as the 45 / 810, as at such small apertures, most of the modern lenses are diffraction limited also...


SCENARIO #2 - Mild DOF (rez holds up relatively well)

M7 f9.4 = 24" Horiz. print

45 f16 = 35" Horiz. print

810 f32 = 50" Horiz. print



Now, the sad news, where most of us end up, more extreme DOF scenes..... and addresses Sheldons question about f32 on 4x5.


SCENARIO #3 - Extreme DOF

M7 f19 19" Horiz. print

45 f32 25" Horiz. print

810 f64 49" Horiz. print



This is why I will never sell my 8x10 gear.... as apt. diffraction attacks 4x5 and 810 equally with extreme DOF, hence the incredible value of added film Real Estate area... But at the smaller end, M7 due to the its smaller f number is more 1/R friendly, and nearly matches 4x5....and when you consider the extreme precision of the M7 system vs. 45, and it's ultra high contrast lenses (not accounted for in this calculation), I will agree with Sandy, M7 actually does match 45 in the extreme DOF scenario, (color film) hence why I call this camera system a gem..... I carry mine in a waste fanny pack...crazy, huh.


Now, before the machine-gun fire ensues.... keep in mind, these are for the same viewing distances, and a very high standard of OPR (On Print Resolution) If you increase the view distance by 2x, you can increase the horiz. print dimensions by 2x, and still hold the same perceived resolution. If you decrease the OPR requirements in half, you can also increase the print sizes 2x.


Also, the lighting you view prints under, is another huge variable not mentioned, which can also deceive many when comparing prints in galleries vs. their homes, etc. The MTF of the eye also plays a critical role in perceived resolution, as it too is effected by it's f number... as it has a fl and a varying apt.... but that's another thread. This is one of the reasons you hear a lot, "that's not what I see" comments. Just making it clear, the print resolutions above is just ONE factor in the imaging chain... although often the most critical one.


Sheldon, does this answer your question?

[andress007]

bglick, I don't know how this guy print, but I sent him jpg @150 dpi and got back quality about the same as my 300 dpi tiff prints.

where is no substitute for square inches of recording media: for digital or for film - it doesn't matter - the larger the better. That's why they made 35 mm film cameras and up to 20x24 film or bigger cameras, and that's whey they make 1/3, full 35 mm and 51x38 digibacks and probably soon 6x6 or even 4x5 digi.

For sure they are different cameras for different jobs and different situations. But always 4x5 film will blow 6x6 and 35 mm film, and 51x38 digital will blow 35 mm digital.

[Sheldon N]

Sheldon, does this answer your question?

Yes thanks everyone, it certainly helps. I guess it's not quite as simple as a "rule of thumb".

Thanks for the kind words on my pics, too. It's funny that I read this thread on high resolution film and digital images with great interest, when the majority of my photos end up getting displayed as 1000 pixel wide jpegs on the internet.

I don't have to do a "Michael Reichman double blind print test" to know that there's no difference between pictures taken with my Canon G10, 5D, or 4x5 Arca Swiss when it comes to Flickr!

[Tim Povlick]

Interesting thread...

A question if I may. The attached is a spot diagram that shows on/off-axis performance of an 80mm-f/7,5 lens. The black circles indicate the size of a diffraction limit spot. On some wavelengths (green in particular) the spot diagram shows it is well inside diffraction limit. It looks as if one could use a green filter some b&w film and 'beat' diffraction limit. What did I miss.

_ .. --
Tim

[rvhalejr]

DQE == quantum detection efficiency
(apparent english order: Detection Quantum Efficiency)

Whereas there is probably no future for better films with a better DQE

1.) 20nm polymer film crystallization (attached image)

"Organic and Polymer Photovoltaics"
ref: ...mrc.utexas.edu/NSFworkshop/Presentations/mcgehee1.pdf

2.) Ag depth of ~13 nm

"Polymer photovoltaic devices with highly transparent cathodes"
ref: ...ntur.lib.ntu.edu.tw/bitstream/246246/87082/1/41.pdf

3.) Current T-Grain Films

....en.wikipedia.org/wiki/Tabular-grain_film

[Emmanuel BIGLER]

the size of a diffraction limit spot

Hello !
Well we are drifting a bit off-topic but let me make a comment !

Well the only problem I see is that I do not know what is the diameter of this diffraction spot ;-) as you probably know, a diffraction-limited spot has no sharp edges, so it is somewhat arbitrary to define its actual diameter.
Lord Rayleigh has a definition, and since He Was Lord Rayleigh, for more than a century people continued to use His Definition

Well assume that we have a reasonable definition for this diameter, for exampe the full-width at half-maximum of the intensity per square area (this is different from the optical density on a film !).
Usually, optical simulations considering white light go up to 0.7 microns, the actual limit of what the human eye can see. If you stop the simulation to 0.5 microns, in the green region of the spectrum, the diffraction spot is smaller and the lens performance, better. And if you reduce the range of wavelengths actually in use, you also improve performace in terms of residual chromatic aberrations.
I have used a long time ago a high performance process lens made by Cerco ( France) for photomask fabrication with high-resolution silver-halide plates ; the lens had a permanently attached green filter. The lens Was actually diffraction-limited @f/4 but for a narrow range of magnifications, I do not thinks there would be any interest for photography, but who knows ?

Regarding the spot diagram method and the definition of the size of a combined aberration/diffraction spot, one of the reasons why engineers like the MTF curve so much in their simulations is probably as follows.
If you make a simulation neglecting diffraction effects, e.g. to compute geometrical aberrations, you eventually end up with infinite values for light intensity per square area detected in the image plane. For example a pure spot of spherical aberration has an infinite intensity per square area at the centre.
So it is meaningless to define the full-width at half maximum for a quantity wich is infinite at its maximum !
Hence, you have to find someting else ; another criterion mentioned in old textbooks was the encircled energy which is not infinite, it is the integral of the light intensity per square are, inside a given circle. This gives a finite result. So you can define the size of the spot by the diameter that will encircle, say, 90% of the total energy in the beam ; doing so, you avoid the problems of manipulating infinite quantities, but this is still an arbitrary definition for the diameter of an image spot.

Eventually the MTF approch is even better since you can define the MTF for a simulation where diffraction is neglected ; like the encircled energy, the MTF is finite and can be computed even if the intensities per square area are infinite.
And you can combine both aberrations + diffraction of course.

The advantage of the MTF simualtion is that there no longer any need to define an arbitrary diameter for a spot, hence this reduces some sterile controversies between lens manufacturers too bad that the MTF approach is so difficult to figure out.

There is a nice tutorial by Carl Zeiss recently published on this subject, the author is H. H. Nasse and there is little to add to this tutorial written by an expert.
http://www.zeiss.com/C12567A8003B8B6..._Kurven_EN.pdf
This is a companion article to the last issue of Zeiss Camera Lens News issue # 30, Dec. 2008
http://www.zeiss.com/C12567A8003B8B6...30_English.pdf

[Emmanuel BIGLER]

... 20nm polymer film crystallization ... Ag depth of ~13 nm ... Current T-Grain Films

Sensitive Polymer & al. research news : many thanks for the references : I appreciate !

Yes ! Analogue and Photo-Chemistry Rule

If our readers are not convinced that Analogue Imaging & Photo-chemistry Actually Rule : figure out how your silicon image sensor is manufactured

Including the last sensor, 12 Mpix by Sony for mobile phones and a pitch of 1.7 microns for the pixels !

[rvhalejr]

...
N=5,6 for the 50 mm in 24x26 format
...

S/B N=5,6 for the 50 mm in 24x36 format

[mcfactor]

One aspect that is being ignored in this post is the greater tonality that you get as you increase film (or sensor) size. While I understand that resolution is negatively affected by smaller aperatures, the tonality of the image is not. So, while an 8x10 negative taken at f/90 may not have much more (if any) resolution that a medium format digital back taken at the optimal aperature, it will certainly look richer and have deeper tones. Which is really why many people continue to use large format.

I realize that this post is about resolution (and in that regard it is very interesting), however the underlying implication is that with higher resolution backs, large format film cameras are becoming obsolete. And I think it should be noted that resolution is only part of the decision to use large format.