When to assign Image Resolution?

Grampa Doug

At what point in the workflow do you assign the proper printing resolution? Does it matter?

arodney

Grampa Doug wrote:

At what point in the workflow do you assign the proper printing resolution? Does it matter?

Its simply a tag. You're telling whatever application how you wish to divide up the pixels should you output the document.

Lets say you have 1000 pixels along the long axis. That's what's key here. At 100 pixels per inch, you've told the application how to divide up those pixels should you wish to output the document and the result is a 10 inch print. If you had instead a tag of 300, that would tell the application you want a 3.3 sized output. So it doesn't really matter until you decide what size you wish when you print or output the document.

As to what point to do this, depends on what application you're referring to.

jfriend

Grampa Doug wrote:

At what point in the workflow do you assign the proper printing resolution? Does it matter?

Don't worry about it. I never set it. It is never used on Smugmug and is never used anywhere where you specify the output print size (which is everywhere that I print).

If you ever do set it, you have to be careful that you don't inadvertently cause a resampling of the pixels which can happen sometimes in apps like Photoshop if you don't do it right.

pathfinder

I never worry about "image size" unless I am ready to print to paper, and even there not so much.

I pay very high attention to the dimension of the image in pixels, short, and long. I need to be able to convert the long dimension in pixels, to pixels per inch ( in my head not in Photoshop )when I am ready to print, so I know that I have enough pixels to maintain an appropriate "pixels per inch resolution" in my final print's actual planned printed size.

When I KNOW EXACTLY what size I am finally going to print, then I can calculate how many pixels in total I need for a given print size at 300 ppi, or 240 ppi, or 180ppi, or even 140ppi. I have seen prints printed at all these different resolutions, and they are look remarkably similar on most modern 8 or 10 or 12 ink inkjet printers that print anywhere from 1 or 2 to 10 or 20 actual dots of ink per pixel.

So do not worry about "pixels per inch" designation in Photoshop, pay attention to total linear pixels until you are actually ready to hit PRINT. Then you do need to calculate if you have enough pixels to print your image with satisfactory quality, or if you need to consider uprezzing, or reshooting.

I wrote a little tutorial about this here

BinaryFx

For print, the resolution and the final print size are both important. Unless one has vector graphics, there will always be issues with sending too little - or too many pixels to the output device (yes, there can be such a thing as having too many pixels, although this is often a lesser evil than having too few).

From my prepress and page layout background, I assign resolution to a digital image based on the desired quality factor and screen ruling of the presumed output device. For a common example, for a 150 lpi halftone screen, the resolution would be 300 ppi. This would "size" the image to the ideal maximum physical size when the image is placed into a page layout or illustration program. If the layout artist then decides to enlarge the image by say 400% - it is known right away that the enlargement chosen is extreme and that quality will be lesser than if the image was left at 100% size or lesser.

The same principal applies to inkjet output from Photoshop, the image can be sized at the "ideal" max. print size and resolution for the original pixel count of the image. If you then wish to print larger, you know right away that there may be a quality cost.

With resampling turned off, resizing the image in Photoshop is simply setting a metadata tag defining print output size (some file formats; save methods and or software may not retain or reference this output size info).


Regards,

Stephen Marsh

http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/

arodney

BinaryFx wrote:

From my prepress and page layout background, I assign resolution to a digital image based on the desired quality factor and screen ruling of the presumed output device.

The same principal applies to inkjet output from Photoshop, the image can be sized at the "ideal" max.

These ink jets are not producing a halftone dot, nothing like it. With something like a modern Epson or similar ink jet, simply set the size you desire in Photoshop, let the PPI fall where it will as long as the resulting PPI isn't lower than 180 on the low end, 480 on the high end. Let the driver handle the output resolution and sizing. The PPI values can fall where they will, as long as we're within the 180-480 range. So 18x22 at 291dpi is fine. The DPI value in this case can be any odd value (it does not have to be an even divisible of some fixed value).

pathfinder

Grampa Doug wrote:

At what point in the workflow do you assign the proper printing resolution? Does it matter?

The original posters question is above and has been anwered around, but not directly. I think the answer is at "the beginning" and "at the end" just before printing.

I usually bring my image into Photoshop from Adobe Camera Raw or Lightroom at its native resolution from my camera unless I know I have a very large or very small images planned for use with the file. That is "the beginning" part.

If I know I want an image larger than 16 x 20 say, then I may uprez the image in ACR. If I am certain I have no need of a full size file, then I might choose one of the smaller file sizes to import the image in to photoshop from ACR, but this is rare, because I usually don't know that in advance. I can always downsize later, with a better result, that I can uprez data that is not there.

As has been discussed above by Stphen, Andrew, and several others, you will want enough pixels so that there is a general match between the image size and your final chosen output.

PrePress - printing to traditional commercial CMYK printers - requires a CMYK files ( usually ) and a discussion with the Printer ( a person ) to know what their machine's specific needs are.

For an inkjet printer, or for most online printing like that done by Smugmug ( which requires RGB files not CMYK files, even though the inkjet uses CMYK and more inks ), as long as you have enough pixels in total, to allow a final ppi of somewhere between 180 ppi and 360 ppi, you should be fine.

If you do your own printing on an inkjet, print a series of images all the same size at 140ppi, 180ppi, 240ppi, 300ppi, and 360ppi, and see if you can really see the difference. You may find a small difference, or you may not. Certainly for very large images, that will be viewed from more than 3 feet, 180ppi can look very nice. Bill boards are not printed at 140 ppi that is for sure!

"The end" is the final ppi chosen just before printing for a given print size. I am like Andrew, I print to an inkjet or Smugmug, and as long as I have enough pixels to be between ~200 - 350 ppi, I let the printing software drivers manage it.

It seems as if these discussions carry almost religious tones for some folks, but I also know that some fine professionals support the answers suggested here.

I have heard a few folks espousing 720 ppi for an Epson inkjet because its native printing is 360ppi, but I have never felt the need for that myself.

BinaryFx

arodney wrote:

These ink jets are not producing a halftone dot, nothing like it.

Apparently I missed adding "general" before principle. Agreed, the common quality factor ratios used for a halftone screen are different to stochastic screens in inkjets, and a halftone does not have the same exact requirements as FM screening.

The general principal is the same in that one can make a print too large for the pixel data in the file - so sizing the file to the expected output can help indicate the largest output size, if one wishes to print larger then it is known right away that there may be quality issues.

As Pathfinder indicated, some prefer to take this into account right before printing, rather than earlier in the workflow. I prefer to know as early as possible what is possible with the data, rather than finding out at the last minute that the output or design does not suit the data.


Stephen Marsh

http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/