The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Print Resolution & COC

the recent thread "Distance measurement" from Tashley, in which he recommended a DOF calculator lead me to think about DOF and COC.
i tried this app and another from the same author now am wondering which settings would give me optimum and usable results.

one of these apps let you set the near limit and the far limit and the focus point will be calculated as well as the needed aperture and the resulting COC. the algorythm also regards diffraction.


but which COC is acceptable to look sharp at a desired print size?

my sensor (6,7micron 39MPx) should bring good sharpness with a COC of 0,041mm when printed and viewed at a distance equal to the diagonal of the print size

i'm able to print @300ppi for 600x400mm without resizing.

but as a pixel peeper i want to view those prints at a closer viewing distance. let's say 250mm instead of 740 (normal viewing distance) - which would be 3 times closer - how big is the maximal COC (which affects my DOF) in this case?
0,014mm? (which would be 3 times more precise)

is there a rule of thumb to know: at a COC of "x" i'll get a decent print at size "y"?
 

tashley

Subscriber Member
Thomas, I am going through the same math at the moment and have decided to work to the print resolution parameter in these apps rather than the blur spot diameter, which is not exactly the same as COC because it includes diffraction. I work to 5lppmm for a 1.2 metre print and in the Pro version of the app that is about 23.5lppmm in 'ten inch print' language. I am currently testing this out by viewing on a MacBook Pro retina screen, which gives about the same impression of DOF as a a 120cm wide print. It seems to work so far. Still testing though! But it is roughly equivalent to a BSD of somewhere between 19 and 21 microns.
 
Last edited:

Shashin

Well-known member
CoC is based on print viewing--if it works for one print size, it works for them all (viewinf distance is a ratio of print size). Basically, 1/3,000th the diagonal of the print area would be photo quality. Since 1/1,500th the diagonal would be acceptably sharp based on DoF criteria, half viewing distance should be sharp as well. A 24 MP sensor if sharp at 100% would give you 1/4 viewing distance as acceptably sharp.
 

Bob

Administrator
Staff member
DOF and acceptable print sharpness are rather disconnected but related issues.
What can be done though, is to consider the angular resolution of the eye or about 4 arc-minutes.
From that and some basic trig it is all derivable
-bob
 

tashley

Subscriber Member
Respectfully chaps, I no longer subscribe or work to some of the accepted formulae. Every show I go to, and I go to a lot of shows and I buy as well as make some very large prints, is full of very large prints. Both at shows and in galleries and in people's homes, I see people very routinely getting much, much close then the diagonal when viewing a print. Sure they also step back for 'the full effect' but they also get their faces right up to the print. So do I. In the real world of people buying fine art prints, this diagonal stuff is, in my experience just not true. So untrue in fact that I think it should be forgotten by all photographers selling certain styles of fine art to a discerning audience.

My personal aim is to be able to make images which, if shot at an ISO where the camera performs well, and for pixel pitches in the current area of d800 or IQ180 territory or larger, are printed to an absolute minimum of 180dpi with 220 dpi or more preferred.

That means:
An rx1 image goes to about 33"/85cm max, 27"/70cm preferred.
d800 a 41"/104cm max, 33"/85 cm preferred
IQ180 57"/145cm max, 47"/120cm preferred

That establishes my personal parameters for prints of the type of subject that need to 'look quite sharp or pretty sharp' assuming good technique, lenses etc. though of course some images lend themselves to much larger reproduction if a dreamy, hazy or grainy type look suits or, possibly, if I am using a very textured paper and want a more painterly look.

Having established those personal preferences, I now have to wonder about DOF and that means establishing a 'safe' minimum viewing distance and for that I am happy with around 18". A really good proxy for all this is: when viewed on a MacBook Retina display, on my lap and at 100% zoom in C1 or LR current versions, are my desired picture elements subjectively 'sharp enough'?

Next, I fully agree with the OP in that COC based DOF is less attractive than the method used in the apps he references, which both use 'blur diameter' as a sort of 'COC that takes diffraction into account'. What I really prefer is to use the print dimension and resolution inputs instead. The problem comes when it comes to translating my preferred parameters into those you need to input into the calculators!

My working assumption for now is that I want 'viewing distance independent' print resolution of, I THINK around 5 lppmm minimum, which seems to equate to a blur diameter of about 19-21 microns for those who prefer it that way.

Having established all that, the calculators indicate (depending on which one you use) either what your DOF IS given your aperture and focus distance, or what you focus distance and aperture NEED to be, given your desired nearest and furthest points of focus are.

The devil is in the detail, however, so I am still testing if the outcomes match my math and my understanding of what is going on... So far so good but there will be some surprises I am sure!
 
Last edited:

Ed Hurst

Well-known member
I can't comment on the specific maths here, not having gone into it, but I totally agree that the issue of diagonals and viewing distance have never worked for me personally. I tend to scrutinise large prints up close as well as from a greater distance and only feel satisfied (visually convinced) if the close-up view retains loads of detail rather than breaking down. It's a tough model but anything else just doesn't work for me.
 

tsjanik

Well-known member
I certainly agree with Tim and Ed. For me, the whole point of many large prints is to invite inspection of details contained within, not just viewing the whole at some prescribed distance. Although the diagonal and CoC can be a guideline, the attempt to quantify a quality which is ultimately very subjective and dependent on print content, viewer’s preference and photographer’s intent is not often useful.

Two examples of content effects from my posts at GetDPI: The first is an image that prints very well at 180 ppi; there is no detail to examine. The second needs the detail in the small stones and does not print well below 300 ppi.

Tom




 
Last edited:

kdphotography

Well-known member
All those numbers and formula crunching apps make me dizzy. It takes the fun/enjoyment out of photography, imo.

It's a lot more subjective, but the only question I ask at the end of the photography and printing process is: Does it look good?

:) ken
 

tashley

Subscriber Member
I absolutely agree: this is a one-off process I go through so as to ensure that it is then fun again, and that getting the resolution and DOF where I want them is second nature so that I can then ensure that what I get in the print size I want is what I need!
 

Shashin

Well-known member
I print large for myself and other artist, museums, and archives. The theory is not wrong. The last print I did for a museum was a crop of digital file printed to 12' x 10'. That display has recently won an award. So, I have the experience.

A 24MP image should stand up to 1/4 viewing distance. For a 40" print, that is a viewing distance of 12". I don't know about you, by 12" is pretty uncomfortable to view at. A "close " distance is 18" for many people. Most art on a wall is viewed at distances far greater than standard.

And if we consider the abbreviation for medium-format digital is OCD, then you folks are not actually the best people to evaluate your prints (this is just one of those things like writers can't edit and proofread their own work). Most people don't give a hoot. What they see is a great image. Nor do they expect that image to "perfect" at 10" away. I have made 40x55' prints from the 645D with tons of fine detail (in fact, there is detail in the file that cannot be seen in the print) and the print is beautiful. You don't change an image by making it big.

And what is more important to a print is sharpness, which has nothing to do with resolution. This is why images from smaller sensor with less resolution can also be printed large and still be fine. I don't know how many times has someone given me something to print and told me they want it big but not too big because it has x number of pixel. I tell them for kicks and giggles lets make a test really really big. I have never had one of these folks leave thinking there is a limit to print size.

Now, if you want to use viewing distance to create the illusion of more detail (or more contrast), then print small. In many cases, you will proportionally get a longer viewing distance--it has nothing to do with dpi.

This is like one of those conversations trying to decide what is the best color. There is no objective standard you can apply here. And when you consider that this comes down to the viewers perception of an image, how do you even know what another eye perceives as beautiful?
 

Shashin

Well-known member
All those numbers and formula crunching apps make me dizzy. It takes the fun/enjoyment out of photography, imo.

It's a lot more subjective, but the only question I ask at the end of the photography and printing process is: Does it look good?

:) ken
+1

Folks like to try the math for some kind of justification, but actually the math says print as large as you want. You have no one but yourself to please because it does not matter.
 

tashley

Subscriber Member
I print large for myself and other artist, museums, and archives. The theory is not wrong. The last print I did for a museum was a crop of digital file printed to 12' x 10'. That display has recently won an award. So, I have the experience.

A 24MP image should stand up to 1/4 viewing distance. For a 40" print, that is a viewing distance of 12". I don't know about you, by 12" is pretty uncomfortable to view at. A "close " distance is 18" for many people. Most art on a wall is viewed at distances far greater than standard.

And if we consider the abbreviation for medium-format digital is OCD, then you folks are not actually the best people to evaluate your prints (this is just one of those things like writers can't edit and proofread their own work). Most people don't give a hoot. What they see is a great image. Nor do they expect that image to "perfect" at 10" away. I have made 40x55' prints from the 645D with tons of fine detail (in fact, there is detail in the file that cannot be seen in the print) and the print is beautiful. You don't change an image by making it big.

And what is more important to a print is sharpness, which has nothing to do with resolution. This is why images from smaller sensor with less resolution can also be printed large and still be fine. I don't know how many times has someone given me something to print and told me they want it big but not too big because it has x number of pixel. I tell them for kicks and giggles lets make a test really really big. I have never had one of these folks leave thinking there is a limit to print size.

Now, if you want to use viewing distance to create the illusion of more detail (or more contrast), then print small. In many cases, you will proportionally get a longer viewing distance--it has nothing to do with dpi.

This is like one of those conversations trying to decide what is the best color. There is no objective standard you can apply here. And when you consider that this comes down to the viewers perception of an image, how do you even know what another eye perceives as beautiful?
I can't discount your experience but it differs from mine. I have been at a number of sales recently in pretty significant places where the standard of buyers was international art collectors who buy some photography as part of the mix. One of the guys there, a good friend of mine, is on the photographic acquisitions committee of the Tate in London and a significant collector in his own right. He and other people were getting their noses up against the images and then stepping back (or vice versa) - I would define the average length of the nose as 18" as I said above :D

Clearly some types of subject, treatment, printing substrate can work gorgeously with a much lower resolution but the sorts of people I treat as standard bearers for very fine quality large prints are the sorts of people to whose quality standards I aspire - Nadav Kander, Burtynsky, Becker at al. And it is that level of quality in a print that my math above is aimed at approaching. So for me, and for the people I saw paying tens of thousands of dollars for images, the diagonal is dead. Dead.
 

Shashin

Well-known member
Tim, I simply may not be articulating myself well. I think Ken sums it up. Print to what pleases you. But I would say, don't work from some kind of arbitrary measure of DPI (I can't see how you see), but actually make prints and see for yourself. This is absolutely subjective. And experimentation is important. Just as we experiment with our processing, exposures, and a whole bunch of other stuff.
 

alajuela

Active member
I certainly agree with Tim and Ed. For me, the whole point of many large prints is to invite inspection of details contained within, not just viewing the whole at some prescribed distance. Although the diagonal and CoC can be a guideline, the attempt to quantify a quality which is ultimately very subjective and dependent on print content, viewer’s preference and photographer’s intent is not often useful.

Two examples of content effects from my posts at GetDPI: The first is an image that prints very well at 180 ppi; there is no detail to examine. The second needs the detail in the small stones and does not print well below 300ppi.

Tom


Hi
I don't know if this off subject or maybe wrong BUT I was taught that ppi - print resolution is printer specific.. For example I use Epson large format printers and the numbers should be either 180 or 360. Canon or HP I am not sure of but they are different but the high and low resolution related 1:2
Am I lost here?
Thanks
Phil
 

kdphotography

Well-known member
+1

Folks like to try the math for some kind of justification, but actually the math says print as large as you want. You have no one but yourself to please because it does not matter.
And for some folks (like me), I just hate math. It's the reason why I ran from the real sciences. Otherwise I would have studied medicine and not law.

:D ken
 

tashley

Subscriber Member
Tim, I simply may not be articulating myself well. I think Ken sums it up. Print to what pleases you. But I would say, don't work from some kind of arbitrary measure of DPI (I can't see how you see), but actually make prints and see for yourself. This is absolutely subjective. And experimentation is important. Just as we experiment with our processing, exposures, and a whole bunch of other stuff.
I think the confusion arises because of the OP question about DOF rather than resolution: I already know, for example, that the maximum print size I personally am happy with from the IQ180 is around 50" or 125cm assuming a good capture and no cropping. That's a shade over 200dpi and that's clearly a subjective opinion that depends then on artistic intent, subject type, paper and blah blah - but it is my default biggest size for a 'noseable' exhibition print of my own work. It translates to about a 36" print from a D800.

But going back to the DOF issue, what the OP and I are trying to work out is how to set the parameters in these two particular apps so as to get useful DOF information from them. Useful meaning, if we set our cameras as suggested in the apps, will we get the sorts of degree of focus we find acceptable in the indicated DOF zone at the print sizes we are aiming at.

This is of great current interest to me because I am using a lens that is new to me, the Rodie 40, and have shot little with it so far - so would like, rather than months of trial and error and possibly missing some unrepeatable shots, to get this sorted now. So I can then have some fun!

EDIT: I forgot to say that I am giving this so much attention because Rodenstock recommend using no smaller an aperture than F8 if full resolution is to be achieved, so I am working with a bit less DOF than I am used to. Upping my game in the process, I hope!
 
Last edited:

tsjanik

Well-known member
I certainly agree with Tim and Ed. For me, the whole point of many large prints is to invite inspection of details contained within, not just viewing the whole at some prescribed distance. Although the diagonal and CoC can be a guideline, the attempt to quantify a quality which is ultimately very subjective and dependent on print content, viewer’s preference and photographer’s intent is not often useful.

Two examples of content effects from my posts at GetDPI: The first is an image that prints very well at 180 ppi; there is no detail to examine. The second needs the detail in the small stones and does not print well below 300ppi.

Tom


Hi
I don't know if this off subject or maybe wrong BUT I was taught that ppi - print resolution is printer specific.. For example I use Epson large format printers and the numbers should be either 180 or 360. Canon or HP I am not sure of but they are different but the high and low resolution related 1:2
Am I lost here?
Thanks
Phil
Hi Phil:

I use an Epson as well. I was refering to the native ppi of the image. I would upscale to 360 ppi before printing, whether from 180 or 300. My intent was to say that I find upscaling from 180 to 360 works when there is not a lot of detail; when detail is important, I prefer to not have less than 300 ppi in the native image.

Tom
 

tashley

Subscriber Member
Hi
I don't know if this off subject or maybe wrong BUT I was taught that ppi - print resolution is printer specific.. For example I use Epson large format printers and the numbers should be either 180 or 360. Canon or HP I am not sure of but they are different but the high and low resolution related 1:2
Am I lost here?
Thanks
Phil
I think I was taught that once too, but I have also been taught to just feed the image to the printer and let it do any required upresing, to set the exact print resolution, to always set the highest printer resolution in the driver regardless of image resolution, and only to print at full moon. Ok, that last one was a lie.

What I do on my Canon iPF6300 is always set the driver dpi to 600 (which is the highest) regardless of print resolution, but make sure that my pixel width divided by my print width is no lower than 180. I once read a very convincing argument that this was the best way to go and it certainly makes nice prints though I have not tried every permutation to see if it makes the best prints possible. This printer's driver offers choice of 300 and 600 by the way, depending on which paper you use.
 

GrahamWelland

Subscriber & Workshop Member
Consensus after multiple, multiple, multiple rounds discussing this seemed to be pixel pitch * sqrt(2) (or * 2) for pixel peepers.

Just be warned that this will give you a very challenging DoF for non-planar subjects even when using tilt. i.e. not much DoF at all using optimum apertures on a MFDB with 5.2um or 6um pixel pitch.

Ultimately it's a decision as to what is 'acceptable' blur for the DoF determination.
 

alajuela

Active member
Hi Phil:

I use an Epson as well. I was refering to the native ppi of the image. I would upscale to 360 ppi before printing, whether from 180 or 300. My intent was to say that I find upscaling from 180 to 360 works when there is not a lot of detail; when detail is important, I prefer to not have less than 300 ppi in the native image.

Tom
Thanks Tom,
I generally bring in the image out of raw at 360. But also I default a lot of the up scaling to image print RIP, If I have to go really much bigger than native size. Then I try genuine fractals and there a bunch of opinions on that process, all at one time or in steps.......
Thanks
Phil
 
Top