Pages: [1]   Go Down

Author Topic: Resolution - again (and again)  (Read 1580 times)

MatkoD

  • Newbie
  • *
  • Offline Offline
  • Posts: 10
Resolution - again (and again)
« on: December 10, 2020, 05:43:24 am »

After reading tons of material, I still could not find a simple Thumb of Rules regarding print resolutions
1. Using Epson LF printers, the optimal picture resolution should be 360ppi / 720ppi when using Finest Detail
2. There makes no sense printing higher than 1440dpi on Matte / Art papers due to higher ink penetration and uneven structure of the paper
3. Canvases - see above
4. Therefore 2880dpi should be used only on coated papers suited for PK?
6. Does the 360ppi makes sense with 1440dpi, and 720ppi only with 2880 resolution? What about very big printouts, say 2,5 x 1 meter? Should all digital be set at 360ppi?
6. When I follow Jeff Schewe's table of Eye resolution and viewing distance and extend it, I come to only 25ppi resolution, viewing a picture of 165 x 110 cm from aprox 3,5 m distance...
7. What is the optimal corelation between digital image (ppi) resolution TO output resolution (dpi) TO viewing distance (printed image size)
8. I guess no one prints a 720ppi, 16bit/Ch, 165 x 110 cm image at 2880 dpi. Or am I wrong?
9. Who Prints with 720dpi and When?
10. 720 and 1440dpi: The VISIBLE Difference at WHICH size / viewing Distance using High Speed Checked (Bi-Directional)
Thanks
Logged

Binartem

  • Newbie
  • *
  • Offline Offline
  • Posts: 47
Re: Resolution - again (and again)
« Reply #1 on: December 10, 2020, 09:19:45 am »

I've been testing printers, papers, resolutions, and pixel peeping for more than two decades and I used to give a lot of technical detail and an "engineering" view on this topic.  These days, I try to just simplify it for people.  The "go to" resolutions as far as PPI (what you send to the driver) are 360 and 720 PPI for Epson printers and 300 and 600 PPI for everything else.  It often depends on how you intend to display your prints and how much scrutiny you expect from viewers.  You may only need 25 PPI for a print viewed at 3.5m but is it possible that some people will walk up closer to inspect smaller areas of the print?  I see no advantage to printing under 300/360 as the only down side is processing time.  Most papers can't hold more than about 720 PPI and if you print higher than that, it likely won't be visible without magnification.  So, what I recommend is:

- Print at 720 PPI (Epson) or 600 PPI (others) for anything about 13x19 and smaller
- Print at 360 PPI (Epson) or 300 PPI (others) for anything bigger unless you expect close scrutiny (then you can still print at 720 or 600)
- Print at 1440 or 1200 only if the print is to be viewed under magnification or for specialized printing like lenticular prints

For that last line, keep in mind that most drivers run at either 600 or 720 PPI so sending a higher resolution than that may produce mixed results.  And some printers (particularly Canon large format) will reduce the driver resolution to as low as 200 PPI if you create large media (paper) sizes: that can be turned off in the driver settings while others (Epson) may require selecting "Finest Details" to get the 720 PPI resolution.
Logged
Binartem, Inc.
ddisoftware, Inc.

Rand47

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1882
Re: Resolution - again (and again)
« Reply #2 on: December 10, 2020, 09:30:20 am »

It may be worth noting that “some” Epson Printers (e.g. SC P75xx/95xx) should receive 300/600 and not 360/720.

Rand
Logged
Rand Scott Adams

langier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1503
    • Celebrating Rural America, the Balkans and beyond
Re: Resolution - again (and again)
« Reply #3 on: December 10, 2020, 10:30:39 am »

Are you looking at the prints with a microscope or on the wall at normal viewing distances? The numbers all seem to be overkill here.

It may have changed but there's people on Lula who have forgotten more than I'll ever learn, but from what I remember and have been told the "effective" res on most inkjet printers is the equivalent of a 60-90 LPI halftone but the "microwaving" and other tech makes it appear much better than reality.

Regarding printing an image sent at those high dpi sizes, 360, 720, etc. Files with that density seem way on the high side. Usually, 300 dpi as a general rule of thumb will suffice for nearly all inkjet printers for most substrates. However, for many images that can be just a starting point.

Do an experiment for your self. Take a file sized your normal print size and print it as is and able on the back the details. Now take that file and size it for your print at 720 dpi, another at 360 dpi, another at 300 dpi, 240 dpi, 180 dpi, 150 dpi, even 100 dpi. Print them all and notice both the time it takes to send to your printer, complete the printing cycle, the file sizes, etc. Mark the res/file size on the back.

When all the prints are in hand, put them on a table or on the wall after shuffling them all. No fair peeking at the back! take a look at normal viewing distances and have others do the same. More than likely, very few will be able to tell the difference between each print let alone be able to take each print from lowest-to-highest file dpi and put it in order. On the wall at normal viewing distances even the lowest dpi input you'll probably pick out, maybe the next lowest and then it will become more difficult.

When you print onto canvas, each file will probably be even harder to distinguish from one-another due to the weave of the material. Canvas is very, very forgiving of the input resolution I've found.

The math will tell you one thing, the print on the wall tells you another and if you push the envelope and ignore the numbers you may be quite surprised!

I'll share with you a couple examples, one from the mid 1990s outdoor advertising campaign and a more recent from a museum show I produced a few year ago.

The 1990s image started as a 35mm negative that was scanned as by a Kodak Pro PhotoCD lab at 6000x4000 pixels. The composition wasn't wide enough so the image was cloned for the format which got us our aspect ratio of 24:7. Seems like a weird size... However at the time, it took a lot of patience to make this thing work. The file got sent to the billboard company and they printed it. 48x14 was the largest of the billboard size signs they produced. I drove by it on the freeway about 50-60 dpi and that image looked great!

More recently my colleague who is a wiz with any camera he's handed had some lovely work shot with his iPhone. The two of use took on the challenge to photograph several California Missions for a prominent museum for a few years since they had nothing scheduled and needed to fill the space. Though many of the photos were originally shot with 10-12mp up to 36mp images and most of the prints delivered were nominally 16x24 prints, we still needed some large images to enhance the show's design. Several of the images we chose were taken with his iPhone.  With good craft not only did they print well at the 16x24 inch size which made most of the show uniform in size (making it easy to frame and hang), but we took those same iPhone images and printed them up to 24x36 (the math stops most to a print of 10x12) and placed side-by-side with the rest, looked great. We took both the Executive Director and Curator on a walk-through after the show was hung and before the opening to discuss our work. When we were finished and pointed out the photos that were shot with the iPhone (probably a 4), they were both blown away! We did the work ourselves simply using Photoshop and printed them on my Epson 9900.

Since those shows and even before, I've taken many images that the math tells me to stop at some piddling print size and set aside the math, pushing how far I could blow-up the image while ignoring the math. You may be quite surprised at how large and how nice your image will look once you print it and hang it on the wall.

Bottom line, don't let the file resolution and the calculations get in the way of a good and large print with these wonderful printers. Push yourself out of the box and see how far you can go!

Logged
Larry Angier
ASMP, ACT, & many more! @sacred_icons
https://angier-fox.photoshelter.com

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20649
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Resolution - again (and again)
« Reply #4 on: December 10, 2020, 10:32:19 am »

Stick with this (and when you have enough resolution for your Epson):
https://www.digitalphotopro.com/technique/photography-workflow/the-right-resolution/
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

mearussi

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 787
Re: Resolution - again (and again)
« Reply #5 on: December 10, 2020, 12:08:22 pm »

The simplest thing to do is do your own tests and decide what you like. That's what I did and found 300ppi for original file resolution for glossy/luster or 180 ppi for canvas printed at either 600dpi (Canon) and 720dpi (Epson) work well enough unless you're looking at the print with a loupe. It also prints faster and uses less ink. BTW, highest dpi settings can actually make the print look worse, especially on matte paper, due to dot spread.
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Resolution - again (and again)
« Reply #6 on: December 10, 2020, 05:28:12 pm »

Stick with this (and when you have enough resolution for your Epson):
https://www.digitalphotopro.com/technique/photography-workflow/the-right-resolution/

The figures quoted in Jeff's article are for the print to resolve to the resolution of the human visual system, which is often taken to be 1 arc minute or 30 cycles per degree. Bob Atkins has an interesting article on SQF (subjective quality factor). It turns out that the visual system is most sensitive to a resolution of 6 cycles/degree (contrast sensitivity function). According to SQF an image will appear sharp to a human observer at this resolution. Thus, if 355 pixels/inch is needed to resolve to the limit of vision when one is viewing the print at 10 inches, one can often get away with printing at considerably less resolution and still have the print to appear sharp. As readers of the long out of print magazine Popular Photography might remember, they used SQF to determine what sized print would appear sharp using their lens testing data. The resolution testing program Imatest can report test results in terms of SQF. See here for details.

Cheers,

Bill

Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20649
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Resolution - again (and again)
« Reply #7 on: December 10, 2020, 06:41:04 pm »

As Bruce Fraser would say: The ideal viewing distance a photographer looks at a print is based on the length of his nose. 
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 218
Re: Resolution - again (and again)
« Reply #8 on: December 11, 2020, 02:44:14 am »

After reading tons of material, I still could not find a simple Thumb of Rules regarding print resolutions
Thanks
You’ve raised a lot of points and questions but let’s start with the following formula to see what you make of it. Viewing distance is the only variable necessary (working in inches)…

PPI = 1 / ((viewingDistance * .000291) / 2)

You might find it helpful to know that printing machines are often set to their highest resolutions regardless of photo file resolutions because they do not only print raster files but also print vector postscript files which require the line quality.
Logged

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 218
Re: Resolution - again (and again)
« Reply #9 on: December 11, 2020, 03:05:46 am »

The simplest thing to do is do your own tests and decide what you like. That's what I did and found 300ppi for original file resolution for glossy/luster or 180 ppi for canvas printed at either 600dpi (Canon) and 720dpi (Epson) work well enough unless you're looking at the print with a loupe. It also prints faster and uses less ink. BTW, highest dpi settings can actually make the print look worse, especially on matte paper, due to dot spread.
"Using less ink" tells me the incorrect profile is being used. Except for special "draft modes," prints using different resolutions should match as far as color and density. Otherwise, less or more ink would make a print either too light or too dark. Prints need the same amount of ink regardless of their resolution. When using most professional RIP software an operator can readily monitor, on screen, the RIP processing effects (mostly density) of different resolutions before actual printing. The different resolutions of prints, however, match in color and density.
« Last Edit: December 11, 2020, 03:12:54 am by Stephen Ray »
Logged

mearussi

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 787
Re: Resolution - again (and again)
« Reply #10 on: December 11, 2020, 06:53:50 am »

"Using less ink" tells me the incorrect profile is being used. Except for special "draft modes," prints using different resolutions should match as far as color and density. Otherwise, less or more ink would make a print either too light or too dark. Prints need the same amount of ink regardless of their resolution. When using most professional RIP software an operator can readily monitor, on screen, the RIP processing effects (mostly density) of different resolutions before actual printing. The different resolutions of prints, however, match in color and density.
This refers to just changing the dpi setting without making a brand new profile just for that one dpi setting. As for using a RIP, next time I have a spare $2,500 I'll consider buying one.
Logged

Binartem

  • Newbie
  • *
  • Offline Offline
  • Posts: 47
Re: Resolution - again (and again)
« Reply #11 on: December 11, 2020, 09:17:32 am »

It may be worth noting that “some” Epson Printers (e.g. SC P75xx/95xx) should receive 300/600 and not 360/720.

Rand

True.  I had forgotten about those.  All printers will report the preferred "native" resolution of the driver to the OS and as long as you go with a multiple of that number, (good printing software will do that for you) you should be fine.  What you don't want to do is print at 720/360 on a driver that runs at 600/300 or vice versa: always go with multiples of the driver "native" resolution.  As I said in my first post, I used to run tests and post super high res scans showing differences but I found most people got lost in the pixels.  Most printers even have a different line resolving power in the horizontal vs vertical and that can be shown in tests.  So the best advice I can give is to always print at either 600 or 720 PPI (depending on the printer) unless you have a good reason not to.  Some of those good reasons might include:

Large prints
If you are printing larger than about 13x19 and they are displayed in a way that you wouldn't expect close examination of small features, there is no benefit to printing higher than 300/360.  Similarly, there is little benefit (other than a few seconds of processing time) to "dumbing it down" to lower than 300/360 because your printer is likely to lay down the final dots (DPI) in the same general pattern anyway regardless of whether or not you "block them" on the input.  Simply put: you're not going to save on ink printing below 300/360 PPI.

Small prints
If you are printing small prints (particularly if they are viewed under magnification) or lenticular prints, there might be a benefit to going higher than 720/600.  Whether there will be a physical difference will depend on the driver, paper, and other factors.

Resolution
Keep in mind the images you are working with.  If you plan to print a single 24MP image at 30x20 inches, you are starting at 200 PPI (of image data) so there would be some benefit to upsampling to 300 or 360 PPI but you'd likely get no visible benefit from going to 600 or 720.
Logged
Binartem, Inc.
ddisoftware, Inc.
Pages: [1]   Go Up