large prints, viewing distance, and screen preview

[Los]

I have a question about using my second monitor as a print preview screen for large prints. I work on a laptop, with a second larger monitor that i've calibrated and get matched prints from. i'm thinking of making prints 20x30 and larger, and i wanted to preview the quality (sharpness, aliasing, noise level) of the prints on the second monitor before sending them off to be made. here is my idea posed as a question:

it is my understanding that at normal viewing distances, larger prints can be made at ppi's lower than 300. if i stand at the intended viewing distance away from my second monitor, with the image magnified at 100%, would this be the same as viewing a portion of the final print at 100 ppi (or whatever my monitor resolution is)? and to take it further, would viewing the same image at 50% on the monitor be like viewing the print at 200 ppi at a normal or intended viewing distance?

does this make any sense?

[Kirk Gittings]

No, a screen is not a print. Print out a small section at home-say an 8x10 crop of the 20x30 and view that.

ALTHOUGH, I am not a believer in the "normal viewing distance" theory. When I go to shows people (me included) are pressing their noses against the glass to look at fine detail. IMO the NVD people are looking for minimum quality.

[Bruce Watson]

...if i stand at the intended viewing distance away from my second monitor, with the image magnified at 100%, would this be the same as viewing a portion of the final print at 100 ppi (or whatever my monitor resolution is)?

No.

Two main reasons. First, a monitor is a light source. A print is a reflective source. Second, a monitor has considerably less resolution than a print (unless you're using an iPhone, which has the resolution but not the size).

I'm not even going to get into things like paper texture, glazing in framing, how the ink colors and the phosphor colors differ, etc.

The bottom line is just no -- viewing on the monitor will not be "the same as viewing a portion of the final print at 100 ppi." It can be close. It can not be the same.

This is why we have proof prints. ;-)

[Los]

thanks, Kirk. The prints would not be for fine art purposes, so the NVD is a good starting point for me. the prints would be over a piece of furniture or high up on a wall out of a nose's reach :)

i'm exploring what degree of enlargement my 12mp dx sensor can tolerate and still make a "good" looking print. also, my research prior to this post suggests that i can maintain quality in the enlargement by reducing the ppi first, then only interpolating up as little as necessary. recently, i had interpolated up to 300ppi for three 20x30's from the same sensor and i felt they looked a little "rough" from all the over sampling.

i don't have a printer at home, so i thought there might be a way i could use the second monitor to help preview (or previsualize) what the prints would look like, since the color and contrast of that monitor are pretty consistent with the prints i get back.

carlos

[Los]

also Bruce, i found out that my monitor is about 90 ppi. i know that's low for most purposes other than billboards. i sharpen 16x20's at 50% magnification in PS, and adjust gamma and contrast at 25% magnification, fyi. i've been happy with the results from that work flow, but as i said earlier, the 20x30's looked roughed up by processing.

for what it's worth, i've begun shooting unit stills on movie sets. i know of at least one instance where another unit still photographer's work was cropped and used for the release poster. i suspect he was using a 12mp or 24mp fx sensor. either way, the cropping would have made the picture element very small for a 40" high onesheet.

http://www.keithbernstein.co.uk/gallery.html?folio=pictures&gallery=invictus

keith wrote me that all three elements of the poster were taken from his unit stills. i wouldn't suspect that they were shot on medium format digital systems. those don't fit in blimps :)

[Bruce Watson]

...I found out that my monitor is about 90 ppi. I know that's low for most purposes other than billboards.

Billboards can be in the range of 3 ppi. I seem to remember that those huge multi-story high billboards in Times Square in NYC are printed in that range. It works because it's impossible to get up close to them, and from the sidewalk on the other side of the street they look gorgeous and sharp.

[John Rodriguez]

i'm exploring what degree of enlargement my 12mp dx sensor can tolerate and still make a "good" looking print.

carlos

It really depends on your opinion of a good looking print. I shoot landscape with a lot of detail and the largest I ever felt comfortable blowing up a print from my D200 (10 mp) was 12x18. Same with my 5D (12 mp). Any larger then that is too soft for my preference (which is how I wound up shooting 4x5), but YMMV. An image without micro detail can be printed larger.

[jonreid]

my printer has always instructed me to judge and set sharpness with the image at 50% in Photoshop, not 100%.

[Lenny Eiger]

i don't have a printer at home, so i thought there might be a way i could use the second monitor to help preview (or previsualize) what the prints would look like, since the color and contrast of that monitor are pretty consistent with the prints i get back.
carlos

I'm with Kirk on this one. Further, if you don't have a printer at home, then go buy one. If you are actually working and getting paid, there is no reason not to. They are just not that expensive. There are a hundred things you can learn about your images, about shooting, about color, lots of different qualities, etc, from actually looking at what the print will look like.

Lenny

[sanking]

I agree with Kirk about normal viewing distance. When prints are displayed in galleries as fine art most people look at prints, at least those prints that interest them, at very close distance, from 12"-18" to as close as 3"4". I assume, like Kirk, that they are interested in fine detail in the print.

And while you need to calibrate your monitor so that what you get in a print closely approximates the density and tonal values you see on the monitor, no monitor can give you an exact idea of what the print will look like. To know what a print will look like, you must make a print.

Sandy King

[Ivan J. Eberle]

You will want to preflight them on a printer, but you won't easily learn all the ins and outs of printing huge prints simply by dicking around with consumer grade Epson printers and prosumer software. Find a commercial fine art printer who has the spendy uprezzing RIPs (subscription key and dongle type that are licensed per machine, and for which these folks often pay thousands annually, not just uprezzing from CS5), and you might just be amazed at how good a 6 foot print from a 10MP or 12MP DSLR can look.

That is, provided you used top flight glass and technique in the first place, and did everything right in the post processing stages.

[rdenney]

To answer your question, looking at a monitor from three times minimum viewing distance (which is maybe 8 inches--so three times that is a couple of feet) gives me an impression of the resolution in a print, assuming that the image is displayed at three times the resolution on the print (300 vs. 100, let's say). That word "impression" is the important part. For me, it's a threshold test, not a quality test.

Part of the problem is the mechanism by which things appear to be sharp. It isn't just resolution, but it is also contrast. Really, it's a matter of the accuracy of modulation transfer--how accurately the details of the scene are rendered. If they are rendered with much lower contrast, the print will not seem as sharp, even though the details may be present. Monitors present modulation fundamentally differently than prints, because they are illuminated sources rather than reflections. My experience is that monitors may seem "sharper" than prints with the same apparent viewed size.

So, if the image on the monitor viewed at the same apparent (from the viewer's distance) resolution seems unsharp, it will probably seem at least that unsharp, and probably worse, on a print. I can learn nothing better than that, however, by looking at the monitor.

Rick "who might reject an image at that size based on a monitor view, but who can never accept one until it's on real paper" Denney

[Los]

thank you for all the replies, guys. i appreciate the wealth of experience here and your kindness in sharing it.

Ivan, i'd like to learn more about the commercial RIPs. Are there any particular product names that I could research on the internet?

Rick, i think you hit the nail on the head with the word "impression". this is what i'm after. you suggested that viewing of the monitor be 3x the distance of viewing a print made at 3x the resolution (assuming 100ppi screen and 300 dpi print). did i understand that correctly?

carlos

[rdenney]

Rick, i think you hit the nail on the head with the word "impression". this is what i'm after. you suggested that viewing of the monitor be 3x the distance of viewing a print made at 3x the resolution (assuming 100ppi screen and 300 dpi print). did i understand that correctly?

Yes. If a "pixel" (and I don't want to get into the usual hair-splitting on the definition of that) is 1/300 of an inch on the print, and 1/100 of an inch on the monitor (both images displayed at native resolution, meaning that a pixel in the image is displayed as exactly one pixel), both will subtend the same angle at my eye if I'm three times as far away when viewing the monitor.

So, I set up the image size so that it will be three times larger than my print size, and view it from three times the distance. That reduces the apparent pixel size down to something like 1/300--similar to the print. But I only use it to weed out images that are not sharp enough. I can't tell enough doing this to know whether it will succeed on real paper. But if it breaks down in that view, it will more than likely break down on a physical print of the same apparent size (which is one-third the size viewed from one-third the distance), too.

Rick "just what I do" Denney

[Los]

cool beans, rick. thanks.

[SolsticePhoto]

There is some basic math here (you can find this in literature) that states the human eye at MFD (Minimum Focal Distance) -- about 10-12 inches can resolved 5 lp/mm over an area of an 8x10. This lines up with the definition of 20/20 vision. 5lp/mm comes to about ~250 dpi. With this as a base and some basic trig you can guestimate the needed dpi at a set viewing distance. As others have stated its not the whole story on how your print is going to look, but it will get you in the ball park.

See p52 of "The Manual of Photography" by Jacobson

[Brian Ellis]

It really depends on your opinion of a good looking print. I shoot landscape with a lot of detail and the largest I ever felt comfortable blowing up a print from my D200 (10 mp) was 12x18. Same with my 5D (12 mp). Any larger then that is too soft for my preference (which is how I wound up shooting 4x5), but YMMV. An image without micro detail can be printed larger.

True. Another variable is the technical quality of the original image. Was a tripod used? Was the optimum aperture used? How good was the lens? How good was the photographer? Etc. It's much easier to make a technically excellent print when the original image is technically excellent.

There also isn't any complete agreement I'm aware of on the lowest ppi that can used when printing and still make a "good print." 240ppi, 300 ppi, and 360 ppi are all sometimes stated to be the smallest or at least the optimum. OTOH, at a George deWolfe workshop I attended George gave an oddball number I've never seen anyone else give, something like 410pp, and said it resulted from extensive testing he, John Paul Caponigro, and other digital luminaries had done.

FWIW, I've made 20x30 prints from a 21 mpx camera at 240ppi that I thought were "good prints." I do always use a tripod, I have excellent equipment, I hopefully know what I'm doing some of the time, etc.

[Los]

solstice, what you wrote vibes with printing research i did when shooting 4x5. i learned that resolution and sharpness should be considered on the whole of the camera system (lens, film or sensor, and paper or monitor), with the lp/mm getting smaller with each step. i recall setting about 8 to 15 lp/mm as a high mark for fine 16x20 prints from 4x5 negs.

what i never considered then, was viewing distance. 10 to 12 inches is very close in some commercial uses like bus wraps, billboards, and "wall dressing" in large corporate environments. i remember seeing beautiful black and white photography when stepping off the elevator of a nice hotel, only to discover when i walked up 10-12 inches close to it, that it was of really low resolution. so now, i'm very aware of the intend use of an image when judging the quality of it. and i must say that since i've started noticing the use of movie print advertising, i've been very impressed with how clean the enlargements are, even when the resolution is a little on the low side (see MoneyBall posters around town). it seems the low res is masked with a good bit of post processing techniques that benefits skin tones and skin textures.

i tried the 3x3 softproofing technique that rick suggested and i hope to verify it with a print test next week. i'll let you know how it goes. i'm still hoping to read more about the hi-end commercial RIPs.

[Los]

if you're looking for the ppi of your proofing monitor, i found this resource online:

http://members.ping.de/~sven/dpi.html

© 2004-2010 Sven Neuhaus <sven@sven.de>. All rights reserved.