Pixel size to print size

[John Di Leo]

Apologizing in advance for this, but my head is about to explode. Semi-retired and as photography has been a hobby for lots of years, I thought I'd try my hand at selling some of my work. So, I am dealing with an online gallery, and as my images are in digital form, I will sell some of them as open edition--The house prints them for sale from my digital upload.

My head is exploding over how pixel size and resolution relate to print size for fine art quality prints. I understand that pixel size for a digital image relates to it only digitally, and only serves as a "numerator" divided by resolution to give what that pixel number would translate to if printed ---"directly?" Is that the right term? in other words if a digital image is 3400x2600 at a resolution of 300dpi (out of LR6) then that image would print at its best detail at (3400/300)x(2600/300) or 11.33"x8.66", is that correct?

If the above is correct--and I am confused enough to think it may not be, or I may be missing something very important--then, if the image was later exported from LR6 at a res of 72, then it could be printed, again at its best achievable detail, to (3400/72)x(2600/72) or 47"x36". However, the "quality" of the image would not be as good, though it would be much bigger. Is that correct?

 

So, if the above is correct, there is a balance to achieve between size of finished print and detail in the print, correct? The attached file is 3422x2667. It was exported at 300 dpi--MAC Preview shows this res (though an online exif reader says 72).

Two questions arise: If the 300dpi image were printed to 47x36 size of the 72dpi print (is that called a "native print" or wrong term?), would it, the 300dpi export, look better than the 72dpi version, having 300/72 or 4x better resolution?

And

could this attached image tolerate printing to 11x14 or 16x20 ish without appreciable loss of quality?

 

Fully realizing large amounts of misunderstanding and ignorance may be on full display, but I've gotten to the point where the more I read the more my eyes glass over.

Thanks (Please move to another forum if appropriate)

Edited March 4, 2017 by John Di Leo

[Tim_Lookingbill]

I'ld suggest you stop thinking about resolution and focus on sharpening the image for possible enlargement. Currently any commercial printer driver can do a decent job of up sampling/increasing resolution to maintain quality just as good if not better than can be done in Lightroom or any other image editor. But it will depend on whether it's printed with halftone dots (commercial digital printer) or stochastic dithered dots of an inkjet. Consult your printer about this.

<br><br>

If you're this concerned just have a test print made at any size you want. Or have a crop of some detail at 2x3 inch area that the printer driver will enlarge to 8x12 (4 times the size) and view the print from a distance most would view if you enlarged 4x the full non-cropped 3400x2600 digital image at its default size as it came out of the camera (no futzing with Photoshop's "Image Size..." or changing it in Lightroom).

<br><br>

Distance the print is viewed has more effect on perceived quality than the resolution it's printed. For example I'm having 8x10 crops made from my 6MP 3000x2000 Raw files on a Fuji Frontier inkjet dry lab at Walmart with microfine picoliter dithered dots. Since I can view the entire print at arms length (very close) I can see all the dithering noise in the image from shooting at ISO 800 which stands out next to the smoothing in other areas I applied in the Raw converter viewed on my display at 100%.

<br><br>

The bigger I enlarge the print beyond 8x10 at a size requiring I view on a wall from a distance, the less I will see all these fine detail artifacts.

<br><br>

In short you're really over thinking this. Let the printer do his job you paid them to do which is provide a decent looking print good enough to hang on a wall.

[John Di Leo]

Thanks, Tim, appreciate it. I understand the overthinking part; my ignorance of the process makes everything seem equally important, so I can get lost in the details and not prioritize. Your response helped clarify.

 

John

[Bill C]

John, it sounds to me like you have a good handle on the details (but I'm not sure that I understand exactly everything that you said). And Tim puts things in a good practical perspective.<br><br>

 

I don't do fine art work, but if I did, I'd probably want to put a general print resolution limit on things - probably I'd be ok with enlargements of an equivalent to 150 ppi or finer on the print. But if someone wanted to double the print dimensions, such that the camera image was only handing over the equivalent of ~75 ppi, I'd be hesitant to go that far. (Perhaps I would, but probably would want to make a test print first.) <br><br>

 

Something I learned about 20 years ago is that the subject matter can have a lot to do with what I would call "acceptable print detail." For example a close-up portrait of one person has a "lower demand" than a group shot of a dozen people. The headshot might appear perfectly sharp with 72 ppi being delivered to the printer, but if you do the same thing with a large group, many viewers will notice that it doesn't seem perfectly sharp. Our eye just seems to get more demanding in that case. And, like Tim indicates, you want to "properly" sharpen, in software, for the conditions. To be clear, I'm speaking of the situation where the viewer can go as close to the print as they want - if you can keep them far enough away then this doesn't happen. <br><br>

 

One last comment, about ten years ago, one of the regular posters here supplied images for a survey done by a New York Times staffer - to see if the public could see a difference between printed images shot with different pixel counts. As I recall they made 16x20 prints from three different pixel-count images, ranging from roughly 150 to 250 ppi equivalent (check the math for yourself) being delivered to the printer. See the details in this thread (and read the NYT link). https://www.photo.net/discuss/threads/nyt-article-breaking-the-myth-of-megapixels.249141/

[Bill C]

The NYT article used three 16x24 inch prints, made from original images of roughly 6, 10, and 16.7 megapixel.<br><br>

 

The result was that, for the most part, people were not able to reliably determine which image was from each supplied image size. So if you work backwards to the ppi equivalent supplied to the printer, this is probably a sensible limiting guideline for you to use, at least as a starting point.<br><br>

 

Personally I'd like to see at least a few test prints on any system that I was unfamiliar with. Should I mention that I've made a full-time living in professional photography and large scale lab work for many years? There's just no guarantee of how your printing service has their machines set up, including the sharpening routines in their printer drivers.

[John Di Leo]

Thanks, Bill. The idea of a print test run is sound. Being new to this process, the way I understand it, and this must be confirmed, is that the prints are sold first, then printed and delivered. IOW I never see the final product being delivered---again, I need to check that. I fully recognize that the weakest link in all of this could be my skill, or, it could be the online house's printing, and I could never know. The alternative is to do a "closed edition" making the prints myself and shipping to them. Obvious advantages and disadvantages for both schemes. I would think that this being a "major" house, their printing would meet the criteria, but I don't know that. They have standards for submission of a minimal pixel count.

re sharpening in the computer...yes, I do that; this is not a representative sample of what I want in terms of sharpness, but represents dancing flambeaux in a night time Mardi Gras parade lit only by the flames and a street lamp half a block away. It was dark, but this was the "feel." Thanks for the advice

[Ed_Ingold]

Camera resolution and print resolution are related, but not necessarily joined at the hip.

 

<br><br>The ultimate factor in detail stems from the camera resolution (plus lens, technique, etc.). With a print, it is important that you not see the granularity of individual pixels. This is usually taken to mean 300 dpi, but even at 240 dpi you can't see pixels in a print without close examination. The difference is largely the expected viewing distance such as 10" for a print in your hand, vs 2 or 3 feet for a large print on a wall.

 

<br><br>Within limits, the print driver will interpolate from the image in order to produce some multiple of 300 ppi in the print. However if you were to take the 6 MP image cited by the OP and print it at 12x18, the native resolution would be 167 dpi on the print, which might cause noticeable pixelation, depending on the subject. To avoid this problem, you can resample the original image in Photoshop or Lightroom, for example, so that the output resolution would be between 240 and 300 ppi. The interpolation algorithms in these products does a good job of removing pixelation while preserving detail. A rule of thumb is that you can double the resolution of a digital image (4x the MP) without creating major artifacts.

 

<br><br>Besides obvious detail, higher camera resolution allows a more natural appearance of textures, such as fabric, leather or skin. Low resolution for portraits creates a "plastic" appearance of the skin, and fails to capture details of the eyes. With film, you achieve this effect with medium format or larger. With digital, 24 MP is a good starting point (or 18 MP without an anti-aliasing filter).

 

<br><br>These numbers are subjective and somewhat arbitrary, biased by my personal experience. I think they are reasonable, but open to debate.

[Tim_Lookingbill]

Not that I'm taking up Ed's offer to debate this but from what I've seen of framed/matted poster sized landscape prints hanging in my local art gallery shot with the best high rez Canon Mark XX series DSLR's, there seriously is no accounting for taste in the sharpness and saturation boost department. Never have I seen a more uglier Southwest Landscapes in my life hanging in a gallery.

<br><br>

So with that as a friendly reminder to the OP and in particular to his posted Mardi Gras image I'ld like to suggest that though that image viewed at its full size appears kind of soft, I wouldn't attempt to sharpen it more because it is of a night scene and naturally our eyes don't make out fine edge detail in such scenes. Overall consistency across the entire image of this slightly soft appearance is more important to maintain but will have a tendency to amplify individual sharp elements and make them appear to stand out when viewed enlarged on a wall.

<br><br>

For example there are pronounced halos on one of blown out spectral highlights of the flames that needs to be softened to match the others. I've applied ACR's Adjustment brush set to -100 Sharpening and touched up similar halos and other inconsistencies like noise to have it blend in with the rest of the image.

[John Di Leo]

thanks to all...I appreciate the info, opinion and advice.

Tim...

"there seriously is no accounting for taste in the sharpness and saturation boost department." Meaning over-sharpenming, out of focus or both?

and

"pronounced halos on one of blown out spectral highlights of the flames that needs to be softened to match the others." thanks...more than one.

 

Ed...

even though a photo is exported from LR at 300ppi the printer will print at lower than that for different sized prints? There must be a chart somewhere that one can look up native print sizes for xxxxxpixels x xxxxpixels @ 300 ppi, so that one could choose the print size most ideally suited to a given image file, or at least not vary far from it. That would assume that the photo's qualities are good to begin with. or garbage in garbage out. A bad photo (for whatever cause) will be a bad photo even if printed ideally.

 

Would it be better to export from LR at a certain expected end print size? As a test I exported that photo to 11x14" size and the file size as well as the dimensions in pixels were very close to the same size as the file posted above.

It sure is easier to produce for a computer monitor, more forgiving. The whole printing process adds another layer, it seems, over which you can only attempt some control.

Thanks again for the patient answers. This printing thing is a whole 'nother world.

[rconey]

Lots of good ideas here. I think the critical one is "what is the maximum size print you will authorize". Ed notes that you can upsize in software (yea, I use Adobe Camera Raw). How much? Certainly up to 200%, maybe more. Starting with more megapixels at capture lets you get to a larger print, holding to 300 dpi as best you can. I suppose the publishing house will print different sizes for the Customer. An option would be to send in the largest size image(upsized in software, then adjusted as needed) you are comfortable with, knowing that some would be printed at a smaller size. My impression is that printing smaller than the native resolution causes less image degradation, so tis would preserve quality at the largest print size.

[Tim_Lookingbill]

thanks to all...I appreciate the info, opinion and advice.

Tim...

"there seriously is no accounting for taste in the sharpness and saturation boost department." Meaning over-sharpenming, out of focus or both?

and

"pronounced halos on one of blown out spectral highlights of the flames that needs to be softened to match the others." thanks...more than one.

<br><br>

I was referring to the over cranked sharpening and saturation in the southwest style landscapes at my local art gallery. That photographer made sure the viewer saw every detail of splintered old wood on the stagecoach, every hair on the team of horses and every needle of the prickly pear that no person would be able to see from the distance he shot from. And of course the sunset oranges and yellows on the coachman's skin, stagecoach wood glowed intensely.

<br><br>

Yours is fine. I just wanted to point out when too much sharp detail and clarity on certain images does not serve the overall feel of the image. The example of the autumn leaves image is just the kind of busy and distracting texture to avoid. It's suppose to come off as a tranquil scene but the water looks like it's vibrating oil. It's downright irritating to look at for me. I found a lot of this kind of unnervingly cranked up texturing in quite a few nature photos. But on a print it could get amplified if the printer upsamples and applies too much sharpening to compensate for the softness.<br><br>

Edited March 6, 2017 by Tim_Lookingbill

[Ed_Ingold]

There are two recent threads regarding sharpening for size and various output media. In particular, different types and zmounts of sharpening are suggested for display, inkjet, continuous tone, and half tone.

 

<bf><br>I find it useful to convert to CMYK for printing on certain devices, including color laser printers. Black and dark areas tend to block up when printed in RGB. In CMYK, the same image tends to use black sparingly, preserving better detail.

Edited March 6, 2017 by Ed_Ingold

[Bill C]

There are two recent threads regarding sharpening for size and various output media. In particular, different types and zmounts of sharpening are suggested for display, inkjet, continuous tone, and half tone.

 

I'm not gonna try to talk about sharpening - I don't have much expertise in that area. But an observation: those threads sort of promote the commercial "Photokit Sharpener." If one looks at testimonials on their web site -

https://www.pixelgenius.com/sharpener2/testimonials.html - we find that photographer Jack Resnicki says, " I thought I knew sharp until I saw what PK Sharpener can do. When I show my clients 24x36 prints from a 6 megapixel capture and you can see chin hairs on the 20 year old female model like it was an 8x10 print from a 4x5 chrome, you got sharp. No crunches, no pixelation, just mind blowing sharpness. " <br><br>

 

Now, is this a useful data point? I dunno. Is he speaking of a head shot that is not so visually demanding like I just mentioned? I dunno. Is it from the earlier days of digital photography with different expectations? Maybe. At any rate it represents a situation where the photographer seemed satisfied with ~ 83 ppi being delivered to the printer. <br><br>

 

In the large outfit where I spent a lot of time, we would establish our sharpening routines by systematically testing combinations of printer driver sharpening and image sharpening via the lab software (such as Kodak's DP2). We might lay 8 or 10 large prints out in a color booth for evaluation, then pick out the best overall combination. But we didn't do fine art, rather it was high volume work printed to meet certain quality standards, not the "best" achievable.

[digitaldog]

There are two recent threads regarding sharpening for size and various output media. In particular, different types and zmounts of sharpening are suggested for display, inkjet, continuous tone, and half tone.

 

<bf><br>I find it useful to convert to CMYK for printing on certain devices, including color laser printers. Black and dark areas tend to block up when printed in RGB. In CMYK, the same image tends to use black sparingly, preserving better detail.

CMYK or RGB are sharpening agnostic as output sharpening is all about the output device, color models (output color spaces) are entirely separate parts up the workflow stream using color management. Depending on the print driver, you could send an Epson either output ready CMYK or RGB independent of output sharpening based on the paper/inkjet etc.

[Ed_Ingold]

I get better results with my color laser jet if I print CMYK vs RGB. The print driver seems to substitute process black for black toner when converting RGB to its native CMYK. It has nothing to do with sharpness or sharpening, although CMYK images print look cleaner and sharper. My Epson 2400 never seemed to care which mode I used. The Epson driver is probably more sophisticated than the one from Xerox, perhaps something about a photo printer vs a fairly good office machine.

 

<br><br>I'm disappointed in print drivers for my iMac. Most of the settings for my Xerox laser jet and Microboards CD printer are not accessible. If I create a document in InDesign, I have to export it as a PDF and print from my aging HP workstation. (Many of the fonts are mutually incompatible, so sending the native InDesign document doesn't work without a lot of extra effort.)

[John Di Leo]

excellent discussion.

It would seem that it is important for me to find out from the online house what printer they are using, if any sharpening tools or resampling are appropriate, and any other specific direction from them. If their printing is in house, they should know or could get the answers. If they out-source, I may be SOL if they cannot get the info, but that might say something about their quality control.

 

This is what they say about file size vs print size, no mention of sharpening, or printer used, just this:

"Your image must be:

• 
  
  
  • A JPEG file

• At least 1200 pixels x 1500 pixels

• Less than 50MB

If you're selling prints, here’s a general rule of thumb regarding image size: the dimensions of your JPEG file should be at least the size of the print, multiplied by 150 with a DPI of 150 or:

(Desired print length x 150) x (Desired print width x 150) at 150 DPI

For example, if you want to sell an 8x10 print, the image size should be 1200 pixels (8 x 150) by 1500 pixels (10 x 150) at 150 DPI."

That's it.

[Ed_Ingold]

150 dpi is pretty coarse for a print you can hold in your hand or approach closely (12" or so). I have machinist's scales graduated in 1/100" or 0.5 mm increments. The divisions can be clearly seen, without magnification. At 150 dpi, you could easily see pixelation, which I find distasteful. IMO, a better rule of 300 dpi for hand-held prints and 240 dpi for large prints, mainly to reduce the file size while remaining nearly invisible.

 

<br><br>I have been distributing JPEG images to my customers sized at 8"x12" and 300 PPI (a little more than 25 MB). That is more than adequate for printing in newsletters or on the web, even with cropping. If they need more, and have the facility to handle native 220 MB TIFF files or 125 MB JPEGs, all they need to do is ask. At that resolution, you can see individual faces in a group of 80 or more, even the color of their eyes.

[Bill C]

It would seem that it is important for me to find out from the online house what printer they are using, if any sharpening tools or resampling are appropriate, and any other specific direction from them. If their printing is in house, they should know or could get the answers.

 

John, I doubt you'll get any real useful answer from this. For example, they might say that many of their customers likely to sharpen a little bit more, but in practical terms that doesn't really tell YOU what to do - after all, how "normal" are you compared to their other customer? If you ask about the machine, you may get something like, "The x size prints are made on a Durst model xyz machine, and we run it on the default print driver settings with version x dot y of the software. The thing is, that the only way this means anything to you if you already work with a similar machine and you can notice, "oh, their sharpening is set a little higher than my machine." <br><br>

 

I think your best is to run at least one test print, in a size you want to sell. If you let them know your concerns, it's likely that they'll give you a free sample print. You can get more mileage out of this by making a composite image, using a half-dozen sharpening variations. (If you label them with text in the image the notes will never get lost.) Once you know what to expect, you can prep your images accordingly.<br><br>

 

Q. Since you haven't mentioned it, do you mind if I broach another subject? Are you up on the details of "color management," using a calibrated and ICC-profiled monitor? If not, you might want to let the lab do the color balancing for you, realizing that you are giving up your creative control to them. Or pay to have them run sample proofs out for your approval. Just fyi in case you didn't think this through.<br><br>

 

For what it's worth, my adult work life has been full time in professional photography, including a long stint with large scale labs. I spent years as the Quality Control Dept manager in an outfit putting out volumes of paper equivalent to well over a million 8x10 units per week. I've worked with dozens of different professional printing machines, probably near a dozen models of pro dye sub printers alone. So I've seen a bit of lab output before.

[John Di Leo]

I am going to see if I can get a test print out of them, saying I want to see what MY product looks like.

Thanks to all for the great advice, a real education.

[Uhooru]

Your image should do 11X14 at 300 dpi without too much problem, but 16 x 20 will fall apart without some up razing and even then you would be pushing it I think. You can do a test and see how it holds up. You can print with a bigger print all the way down to 180 dpi if not looking too closely. I've gotten good results with roughly 4500 xy 3400 at 18x24 at 180 dpi and it looked good. I may have up-rezed to 280, but can't remember.

[Tim_Lookingbill]

Check out my demo of the deer print low rez thresholds. You can test the minimum ppi to size of print by turning on your rulers in Photoshop and zooming to whatever makes an inch line up exactly or close to it holding up a real ruler to your display. If you see sawtoothed edges lowering the ppi, you'll have to up sample, but this is where a minimum up sample setting can be applied until you can't see the artifact.

 

https://photo.net/beginner-photography-questions-forum/00acnx

[John Di Leo]

again, thanks to all. I uploaded a couple of photos, during the process they showed prices--that I could adjust-- for sizes up to 16x20. So far, I don't see a way of telling them "don't enlarge to 16x20. I have a call into them to find out how to limit size availability.

I am fine with only going up to 11x14 as long as the ppi will support it.

[Tim_Lookingbill]

How it worked for me using Walmart's one hour photo kiosk for printing 8x10's (which had to be cropped reducing resolution from 6MP), the actual choosing of the size of print determines the resolution. All I did to the file on my end was set 8x10 dimensions in inches in Photoshop's "Image Size..." and uncheck "Resample Image". Then I just got tired of futzing with it and submitted as is preset by the raw converter where opening in PS has "Image Size..." at its default 240 ppi. The dimensions in inches were somewhat close but not exactly 8x10.

 

No difference in quality between the two prints. Walmart's Fuji Frontier printer driver did a great job on the print.

 

Since you don't know what your printer's printer driver or server RIP decides what is enough resolution before they have to resample/interpolate pixels, you'll just have to pull the trigger and have them print an enlargement or do a smaller test print.

 

Like I said before, let them screw it up. At least you'll also find out how well they maintain quality control. You can always do the ppi to actual size saw tooth edge test I showed in that link to the deer print. If their results show they over interpolated unnecessarily due to some default ppi minimum, make them do a reprint. In this fiercely competitive photo printing market, the buyer is king.

[DeanJones]

For a 4" x 6" print, the picture determination ought to be 640 x 480 pixels least. For a 5" x 7" print, the picture determination ought to be 1024 x 768 pixels least. For a 8" x 10" print, the picture determination ought to be 1536 x 1024 pixels least. For a 16" x 20" print, the picture determination ought to be 1600 x 1200 pixels least.