Pages: [1]   Go Down

Author Topic: How large can you print? Checking out Bart van der Wolf's tool  (Read 7744 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
How large can you print? Checking out Bart van der Wolf's tool
« on: November 16, 2013, 03:15:30 am »

Hi,

Theres was a recent question about how large we can print. Bart van der Wolf has a very nice tool on the net to calculate that:

http://bvdwolf.home.xs4all.nl/main/foto/dofplan/dofplan.html

Eyesight and viewing distance matters a lot. I have a few 70x100 cm size prints but I normally print just A2 (16x23"). I like large prints but suffer from acute wallspace limitation syndrome.

One of the experiments I have done was to make a large print from a 10 MP APS-C camera. The image was carefully processed and optimised for large size printing and than printed at 70x100 cm (27.5 x 39.4"). That print actually is quite good, if you view it at something like 80-90 distance. I have since shot the same subject, with 12, 24, 39 MP cameras in APS-C, 24x36 and 39x48mm format, but did not need to replace it.  I have 20/20 vision with corrective glasses, AFAIK, so normal quality in Bart's should suit my vision.

Filling in 0.9 m viewing distance and 10 MP gives 1003x674 print size, very close to my viewing experience.

Next experience was when I stated using an 24 MP full frame camera. I compared my 24 MP APS-C camera to my 12 MP APS-C camera, printing at A2 size. I made three comparisons. In two cases the 24 MP camera was slightly better and in case the two images were indistinguishable (by any tested observer). Let me make clear that the image quality on screen was significantly better with the 24 MP camera.

Bart's tool gives 41x62 cm size for the 12 MP camera when viewed at 50 cm distance. So, Bart's tool says that there would be little visible difference at 50cm viewing distance, and that was the case.

The third experiment was a shot made at the same time with 24MP APS-C and a P45+. In a sense this shot was slightly biased in favor of the 24MP camera, as it was used with a zoom lens giving the wanted framing, on the P45+ I used the longest lens I have 150/4 and I would have needed an 180 mm lens to match. That cropping essentially dropped the MP advantage of the P45+.

I made 1/3 crops of the images and printed at A4 size. Viewed on screen the P45+ image had an obvious advantage. The A4 prints corresponded to 57x86 cm (22x34"). At that size the prints were difficult to tell apart. Once you found a detail and compared details side by side you could see the difference. Just flipping around the copies and have a good look without referring to know details I guess my hit rate was about 30%.

In this case Bart's tool gives 87x58 cm print size for normal vision and 50 cm viewing distance. We also failed to see significant difference at shorter viewing distances, too.

To sum up:

Although what you see on screen are the pixels that go to the printer, the printing process is a great equaliser. The gain in image quality is much less in large prints is less than what I would have expected after viewing the image on screen.

Bart's tool gives print sizes that seems relevant to my experience.

As a side note, I don't think we view prints as close as we may believe, 25 cm is quite short, we need good illumination. For large prints I would guess that 50 cm is pretty close. That is about half an arms length.

Best regards
Erik
Logged
Erik Kaffehr
 

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #1 on: November 16, 2013, 04:55:06 am »

Hi,

Theres was a recent question about how large we can print. Bart van der Wolf has a very nice tool on the net to calculate that:

http://bvdwolf.home.xs4all.nl/main/foto/dofplan/dofplan.html

Eyesight and viewing distance matters a lot.

Hi Eric,

Indeed, visual acuity and viewing distance matter a lot, although sub par print quality and sharpening can waste some of the enlargement potential. It therefore also helps to incorporate that angular resolution into decisions (about DOF and magnification) we may need to make when planning a shot. For that we need to determine the angular resolution with which we view the final output, and my tool makes it a bit easier to calculate that.

The use of my planning tool is explained a bit further in another LuLa thread.

Quote
To sum up:

Although what you see on screen are the pixels that go to the printer, the printing process is a great equaliser. The gain in image quality is much less in large prints is less than what I would have expected after viewing the image on screen.

Bart's tool gives print sizes that seems relevant to my experience.


That's correct, once we exceed a certain standard output size (say when we start to need and turn our head to see the entire image), then the viewing distance will gradually allow to output larger than most people expect. This of course assumes that the shooting technique was good enough to offer adequate resolution.

Although excess resolution won't hurt (it just produces larger files than needed), it will also be unnecessary when we cannot resolve it by eye. That's were the reality check can certainly help. The 'normal' output quality setting in my tool is indeed what I also use for normal situations.

Quote
As a side note, I don't think we view prints as close as we may believe, 25 cm is quite short, we need good illumination. For large prints I would guess that 50 cm is pretty close. That is about half an arms length.

That's right, for normal people, not for photographers. ;) It may be nice to know that what I call 'low quality' in my tool is what Zeiss uses as their DOF criterion (2 minutes of arc), so they assume 2x normal viewing distance to make up for the lower resolution at the boundaries of the DOF zone.

That's why 'Normal quality' will usually be good enough, even for photographers, unless one wants to push quality to its uncompromised maximum, where 'Higher' and 'Very high' quality settings in my tool can be useful. Very good output processing will also be required to fully utilize the potential. Output Deconvolution sharpening after final resampling and high spatial frequency boost will help to get the most out of the viewing experience. Product photography is such an application where the almost tactile surface structures can really make a difference.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

PhotoEcosse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 712
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #2 on: November 16, 2013, 06:13:20 am »

Thanks for the link to Bart's tool. Very interesting.

 :)
Logged
************************************
"Reality is an illusion caused by lack of alcohol."
Alternatively, "Life begins at the far end of your comfort zone."

Hening Bettermann

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 945
    • landshape.net
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #3 on: November 16, 2013, 07:06:03 pm »

Thanks again to Bart for this tool, and to Erik for sharing your findings.

chez

  • Sr. Member
  • ****
  • Online Online
  • Posts: 2435
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #4 on: November 17, 2013, 10:24:23 am »

This artificial viewing distance is the equalizer. What if you did not have this "viewing distance". I have my work in galleries and the gallery does not have any roped off areas so the public is free to get as close as they want. Observing people in the gallery, many will stand back and take in the entire photo at one time, but then they come in close, very close to view some of the scene details. I've seen this happen on many occasions. If your print does not stand up to fine detail scrutiny, I think many would be disappointed. I just don't buy this viewing distance argument. One beauty of large landscape prints is that you can come in close and almost walk right into the scene. Viewing up close and seeing all the fine details of the scene is just another experience of viewing a print, one thst I myself do a lot when going through galleries.
Logged

Rob C

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 24074
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #5 on: November 17, 2013, 11:06:45 am »

Thanks for the link to Bart's tool. Very interesting.

 :)


The mind positively boggles.

Rob C

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #6 on: November 17, 2013, 11:19:40 am »

This artificial viewing distance is the equalizer. What if you did not have this "viewing distance". I have my work in galleries and the gallery does not have any roped off areas so the public is free to get as close as they want. Observing people in the gallery, many will stand back and take in the entire photo at one time, but then they come in close, very close to view some of the scene details. I've seen this happen on many occasions. If your print does not stand up to fine detail scrutiny, I think many would be disappointed. I just don't buy this viewing distance argument.

Hi,

That's no problem, just use the (default) reading distance as 'minimum viewing distance', and one can approach up to that distance without loss of visible resolution. When the distance is larger than that, then some detail will become impossible to resolve by eye, because it becomes too small, and other detail will be at the limit of what can be resolved by eye.

Cheers,
Bart


P.S. There are also situations where the public is not able to get close, in which case a larger output size will still be as good as can be, given that larger minimum viewing distance.
« Last Edit: November 17, 2013, 11:24:26 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Christoph C. Feldhaim

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2509
  • There is no rule! No - wait ...
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #7 on: November 17, 2013, 11:29:29 am »

I wonder whats a sane assumption for viewing distance in fractions/multiples of the image diagonal.
I usually move a bit back and forth when viewing images - peeping a bit into the image to explore a detail,
then standing back again to see the whole.

1-2 times the diagonal appears good to me - but what do you think?

There has bee this 6 MP discussion around, but what would be a MP limit for a large landscape shot,
since we usually don't watch 2x3 meter prints from 25 cm ...?
There should be a sort of upper Megapixel limit where more resolution does not help anymore since you'd stand back more.

Ideas?

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #8 on: November 17, 2013, 12:10:04 pm »

I wonder whats a sane assumption for viewing distance in fractions/multiples of the image diagonal.
I usually move a bit back and forth when viewing images - peeping a bit into the image to explore a detail,
then standing back again to see the whole.

1-2 times the diagonal appears good to me - but what do you think?

Hi Christoph,

I don't think there is a general consensus, also because aspect ratios of images can vary a lot. In my tool I adopted the 1x diagonal of output size as a starting point, which usually works okay for 3:2 and 4:3 kinds of aspect ratios. It allows to take in the entire composition without having to turn one's head, just scanning it with our eyes will suffice.

That is incorporated in the comment under section 1.4) of my tool, in the 'could be considered as having a 'standard' size' part (with a diagonal being approx. equal to the viewing distance entered earlier).

I may change that if something better can be found. Perhaps the vertical field of view of our eyes is a more important starting point? After all, a portrait orientation works differently compared to a landscape orientation (because with 2 eyes we cover a wider than taller field of view).
 
Quote
There has been this 6 MP discussion around, but what would be a MP limit for a large landscape shot,
since we usually don't watch 2x3 meter prints from 25 cm ...?
There should be a sort of upper Megapixel limit where more resolution does not help anymore since you'd stand back more.

Ideas?

In a way that is what the first section of my tool does. it establishes the required resolution, and scales it down with increasing distance, thus allowing a larger output size (dictated by available sensor dimensions) because it will be viewed form a larger distance.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

AreBee

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 638
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #9 on: November 17, 2013, 12:45:40 pm »

Bart,

Quote
...aspect ratios of images can vary a lot. In my tool I adopted the 1x diagonal of output size as a starting point, which usually works okay for 3:2 and 4:3 kinds of aspect ratios. It allows to take in the entire composition without having to turn one's head, just scanning it with our eyes will suffice.

Does this mean that the tool can be considered conservative when applied to 3:1 panoramas, because for the same height as a 3:2 or 4:3 print, the panorama is significantly wider? Or, is it the other way around, and for the same width as a 3:2 or 4:3 print the height of the panorama is significantly shorter?
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #10 on: November 17, 2013, 01:40:15 pm »

Bart,

Does this mean that the tool can be considered conservative when applied to 3:1 panoramas, because for the same height as a 3:2 or 4:3 print, the panorama is significantly wider? Or, is it the other way around, and for the same width as a 3:2 or 4:3 print the height of the panorama is significantly shorter?

Hi Rob,

The only part where aspect ratios matter at all, is with that comment which calculates based on diagonal output dimensions. All other calculations are based on angular resolution.

As far as the viewing distance of panorama's is concerned, it is precisely that where I think that vertical field of view is more telling for how we experience the image. It's no problem having to turn our head to view the full width, it even adds to the viewing experience of a wide scene, but having to tilt our head up and down is slightly less comfortable.

Given the fact that most sensors have a more modest aspect ratio, the diagonal as viewing distance will work okay most of the time, just imagine that a pano is wider (sue to stitching or cropping) than that suggested size, but the height is the same. So I'd just use the height as guidance for 'normal' landscape oriented output size, and let the width be whatever it is.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

AreBee

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 638
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #11 on: November 17, 2013, 02:02:53 pm »

Thanks Bart.
Logged

EinstStein

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 498
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #12 on: November 17, 2013, 07:18:34 pm »

The finish of the print paper and  the lighting condition matter a lot too. 
Logged

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 17835
  • When everyone thinks the same, nobody thinks
    • My website
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #13 on: November 17, 2013, 07:43:23 pm »

... Observing people in the gallery, many will stand back and take in the entire photo at one time, but then they come in close, very close to view some of the scene details. I've seen this happen on many occasions. If your print does not stand up to fine detail scrutiny, I think many would be disappointed....

Are you sure those were not photographers as well?  ;)

And if not, my assumption is they are just making sure it is what they think it is they are seeing ("It's A Bird... It's A Plane... It's Superman" type of thing). I have several 20"x30" made on canvas from an 8 mpx camera that, when viewed from a proper distance, tell one story (say a lighthouse), and then, when approached for close inspection, reveal a wealth of details (say seagulls) unobservable before. That is what viewers are looking for at close inspection, what the detail is, not how sharp it is. At that point, they are not interested in whether they can discern every feather on that seagull, just that it is a bird. After all, the main story they are looking at is the lighthouse in its environment, not an ornithological close up.

EinstStein

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 498
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #14 on: November 17, 2013, 09:48:26 pm »

I thing BVDW tool has a foundamental error. It assumes every pixel is equal fro camera to camera.
For example, if you try Leica C (P&S) and Leica M8.2, you'll find Leica C prints out bigger than M8.2.
It's interesting the tool can not compare Nex 6 vs. 7.
No, you can't do it just by counting on the pixel and than enlarge it.

 
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #15 on: November 18, 2013, 06:47:10 am »

I thing BVDW tool has a foundamental error. It assumes every pixel is equal fro camera to camera.
For example, if you try Leica C (P&S) and Leica M8.2, you'll find Leica C prints out bigger than M8.2.
It's interesting the tool can not compare Nex 6 vs. 7.
No, you can't do it just by counting on the pixel and than enlarge it.

Hi,

I've addressed your concern in this post in a bit more detail. In short, the tool is based on the limits of visual acuity, not on pixels. When visual acuity is unable to resolve detail, the 'pixel quality' of that detail is irrelevant. The Leicas mentioned indeed result in having marginally different output sizes beyond which blur becomes visible. Pixel quality is relevant for the detail that can be resolved by eye, but then more factors also play a role, starting with shooting technique and lens quality and ending with post-processing, output modality, and viewing conditions.

At your request, I've added the NEX-7 to the built-in list of selections. I will add other models as time goes by and new models are introduced, but if requested I may also add some older (exchangeable lens and fixed zoom) models or digital backs. One can always just type in the relevant numbers that are required for missing models.

That can also be useful for e.g. film-scans (there also pixel quality varies with the MTF of the total imaging chain) or panorama/stitched images. With a given aspect ratio in the dimensions, one can also experiment with a different total Megapixel amount, which will be rounded to 16-pixel multiples for the sensor dimensions (in compliance with e.g. JPEG compression in blocks of 8 or 16px).

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Christoph C. Feldhaim

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2509
  • There is no rule! No - wait ...
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #16 on: November 18, 2013, 06:53:31 am »

I wonder if your tool could be modified to work with the SQF as described here:
http://www.imatest.com/docs/sqf/

Would surely be a major undertaking, since it would only work for lens/sensor/printer/paper combos and would be something like an extension of the DXO website.
However - the tool as we have it now is a great thing to have and ... it works offline since you designed it so, and thats awesome.
Thanks for doing this! :)

Cheers
~Chris

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #17 on: November 18, 2013, 07:19:59 am »

I wonder if your tool could be modified to work with the SQF as described here:
http://www.imatest.com/docs/sqf/

Would surely be a major undertaking, since it would only work for lens/sensor/printer/paper combos and would be something like an extension of the DXO website.
However - the tool as we have it now is a great thing to have and ... it works offline since you designed it so, and thats awesome.
Thanks for doing this! :)

Hi Chris,

You're welcome. I may consider building a dedicated tool for print quality. The DOF quality planning tool allows to test some other (also image capture) scenarios, the print size resolution limit is just a by-product of (or input for) the need to determine the COC in a more meaningful/targeted fashion.

A dedicated print tool would indeed be a major undertaking (although I already have some modules available to make it possible), because one would need to somehow incorporate the system MTF and human Contrast Sensitivity at various spatial frequencies (and thus distances) into a total SQF. Without MTF input it becomes difficult to quantify subjective image quality, and most people have no idea about the MTF of their imaging chain, so it needs to be determined in another way (I'm working on that...).

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

chez

  • Sr. Member
  • ****
  • Online Online
  • Posts: 2435
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #18 on: November 18, 2013, 11:57:23 am »

Canvas has a way of obscuring the details due to its texture so any fine details are lost. I've gone through a lot of galleries in the South east where there are large prints on gloss material. Looking closely into these prints you can see the texture of the canyons, fine plants etc... which you cannot really see ( or at least I could not ) from standing back. It is these fine details of texture that would dissapoint if they were all mush when viewed up close.

Interpolating never adds detail...always takes it away.

Are you sure those were not photographers as well?  ;)

And if not, my assumption is they are just making sure it is what they think it is they are seeing ("It's A Bird... It's A Plane... It's Superman" type of thing). I have several 20"x30" made on canvas from an 8 mpx camera that, when viewed from a proper distance, tell one story (say a lighthouse), and then, when approached for close inspection, reveal a wealth of details (say seagulls) unobservable before. That is what viewers are looking for at close inspection, what the detail is, not how sharp it is. At that point, they are not interested in whether they can discern every feather on that seagull, just that it is a bird. After all, the main story they are looking at is the lighthouse in its environment, not an ornithological close up.
Logged

wolfbellw.

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 72
    • http://www.wolfgang-bellwinkel.de
Re: How large can you print? Checking out Bart van der Wolf's tool
« Reply #19 on: November 18, 2013, 12:11:24 pm »

its all about the right distance.
this image is about 4.50m wide and was printed from a
4,5/6,0 negative.

'Landscape of rest' - BACC Bangkok
04 October - 24 November 2013
Logged
Pages: [1]   Go Up