PPI/DPI setting... why do you want it?

Respectfully Robert, you've gotta admit that you are doing a whole lot of assuming here, and asking us all to go along with you in your very philosophical "I believe they want"s. If we ask the gang at Apple do you think we'll ever actually get an answer? This has become my new koan (Think: If a tree falls in the woods and no one is there to hear it...).

Ask over on that other forum, and I bet you'll have an answer from one of the decision makers within a few hours (during business hours). There is the biggest difference I can see so far. One team seems to be listening and asking for input. The other seems tone deaf, arrogant and ill-informed. Silence has never been a particularly respectful seeming policy in answer to paying customers with what they feel are legitimate complaints and questions.

In answer to the question I think a lot of cameras
have 72 DPI as a default because they mostly assume
the final image is headed to the web. It's just the
DPI that makes the most sense for most users, because
if users do print consumer printers simply cannot be
as persnickety about the DPI as a professional
printer.

You may be correct as to why camera makers use that number. If so, their reasoning is wrong. The Web doesn't care what the dpi is at all, while print cares very much. The only numbers that matter on the Web are how many pixels tall by how many pixels wide. If one image is 900x600 pixels at 300 dpi, and another is 900x600 pixels at 72 dpi, they will appear at the same size on a web page: 900x600 pixels. It's too bad they have it backwards: Dpi as no meaning on the Web, and all the meaning in print. If camera makers want to assign a dpi, it should be a print dpi.

Rather than focus my comments on the sameness of pixel dimensions vs. dpi/ppi as many people have done very clearly and correctly already, I'd like to suggest that from a workflow perspective including dpi/ppi may mean less work, and fewer headaches, for some people. I personally would like to be able to specify a ppi/dpi when I crop an image, and also when I export an image.


If I have a jpeg image from my Nikon D200, the pixel dimensions are 2592 x 3872, easily large enough to produce 8x10in images at 300dpi without resampling. At the end of a long day I may overlook my pixel dimensions, and end up cropping the version down to 1564x1956 pixels, which is an 8x10 aspect ratio, but far below the 300dpi required by my print shop for 8x10 prints.

I feel it would be useful for Aperture to resample the version right away to 300 ppi/dpi, showing me any artifacting or pixilation caused by cropping the image so much. This means I don't waste time exporting the image at full size, then resizing/resampling it in photoshop only to realize that it would've been better to crop the image differently, or not at all. Part of my workflow is determining what photos meet my needs, and adding a ppi/dpi setting to the crop tool could save me some time in this case.

It was illustrated very well by Robert, how unimportant to the actual image the dpi/ppi is. At the very least, I would say dpi/ppi is a convenience to help us humans work better together, and to help get images closer to the real world output we want. For myself, this convenience, which is essentially non-destructive to the image itself, can save me phone calls, emails, uploads, downloads, meetings and headaches.

The same way that far more destructive options are within my control at export, such as choosing a new image format, color depth, a watermark, a new color profile or less harmful options like the inclusion of metadata; it seems to me, that the ability to select the dpi/ppi on export should be an option for those of us who would like it.

powerbook G4 (late 2005) Mac OS X (10.4.3)

As has been beautifully explained, 72 dpi is a legacy resolution setting that is no more appropriate today than using a TRC gamma setting of 1.8 for the display in the Mac OS. But that’s another story.

The question that should be looked at here is the one where a poster asked “Why was this set to 72 dpi and not able to be changed? Design or rush to market?”

This is a question I ask ALL THE TIME when using Aperture and even asked the product manger to explain to me. There are two possibilities with a lot of these issues. One is it’s a bug. The second it was designed this way on purpose. The 2nd options is NOT a good one! There are far too many bad design issues in Aperture and this resolution tag (that’s all it is, a tag) stuck at 72 is lame! It’s not like we’re in Photoshop 1.0 history void here. There have been imaging applications on the market now for 16 years or longer. They’ve had functionality like curves, info palettes and the ability to enter any resolution tag for the number of pixels you have in an image. WHY is what is supposed to be the most modern take on imaging from Apple so poorly thought out? That’s the question we need to ask Apple (and Adobe in light of Lightroom).

PowerBook, G5 Mac OS X (10.4.3)

Well I have to say that I agree 100% with Robert (and Apple and Adobe).

Because it may seem well and good to be able to change the PPI, but then a very ugly problem rears its head:

Do you resample?

Well that brings up issues of "Why are you changing the PPI?" and "Do you want to match an absolute physical size (300 PPI for an 8" x 10" print), or just change the existing PPI to 300 PPI keeping the same number of pixels?" How much confusion does Photoshop's Resize dialog generate for people?

Those are very, very confusing issues for people who don't care about them (i.e., non press people).

As Robert said so eloquently, those are strictly pre-press issues. They have 0% to do with the typical photographer.

If you're publishing your photos in a book, Congratulations. If you decide to do some of the pre-press stuff yourself, use Photoshop for that functionality. It doesn't belong in Aperture or Lightroom.

Changing the resolution tag and resampling are different. We’re talking about simply providing a tag in the file that tells some applications and printers “given this number of pixels on the long and short axis, divide them up by X”. Adding or removing pixels (resample) is another issue.

Please read all the posts.

No resampling is involved in the proposal; in fact, no change to the image data itself. We are only talking about changing one value that only exists in the metadata, from "72dpi" to another value like "300dpi". This will cause a more useful interpretation of image size in everything from Photoshop to Preview to prepress.

Also please review the post where I showed that dpi does not matter for the Web, but does matter for print, therefore if any dpi should be the default, it should be a reasonable print value; and because the cameras aren't doing the right thing, it would be nice for our software to.

Let's rephrase that previous paragraph:
- In the current situation, no one gains, and some lose.
- If the proposed feature is added, some will gain, and no one loses.

It's hard to argue against that!

I was reading this thread and I was wondering what the dpi of my images were coming from my D70. So I took an image that had been imported into Aperture, and opened it in Photoshop CS2. Image size said it was 72 dpi. That really worried me because I was working to put a book together and 72 dpi for printing probably would not look very good.

But then I opened the same image file (but NOT imported into the Aperture library) in Photoshop CS2, and it said it was 300 dpi under image size. Same source, one imported into Aperature, the other as is... both looked at by Photoshop...different dpi.

So now I am very confused.
1. If I order a book thru Aperture, will they try to print at 72 dpi or will they adjust to the best dpi for their printer?
2. As a photographer that does final hard print as the goal, print quality is king... not presentation on the web!
3. During import does Aperture automatically size the images to 72 dpi? Is this a user selectable option. Not sure I care as long as the final output (printable jpeg files are 300)

Thanks
steven

Thanks Ian,

I was reading that some of the books coming out of Aperture did not look any better than iPhoto books so this made me wonder about the process of going from 72 dpi to something more acceptable for printing. Do you know where Aperture books are printed and do you understand their process for handling the files?

Thanks
Steven

Please read all the posts.

Please read my post. I said:

Well that brings up issues of "Why are you changing the PPI?" and "Do you
want to match an absolute physical size (300 PPI for an 8" x 10" print), or
just change the existing PPI to 300 PPI keeping the same number of pixels?"

Clearly, I understand the difference between setting the PPI alone, and resampling. The way I typically change the PPI is in Photoshop's Resize dialog, where you can turn off resampling. This just changes the PPI.

If people are given the option to change the PPI, will they then ask for resampling? It's a slippery slope. And IMHO will lead to feature-itis. I just don't think it belongs in Aperture.

Your math is a little off.

ppi is a measure of pixel density, (ie, pixels per distance).

Yes, if you have the total number of pixels and the total real size (= distance) you can calculate pixel density.

The important equation here is not:

ppi * size = pixels

It's:

f(ppi, pixels, size) = c

That is, in calculations regarding size, pixel count, and pixel density, there are only two degrees of freedom. Given any two, the third can be trivially calculated. And this is important.

The file interchange formats like jpeg & tiff are communications mechanisms. When you communicate an image to someone else, like a printer, you are attempting to tell the printer what you want him to do, (ie, how to render the image). And your file format includes, essentially, the ppi & pixel count info. When you tell him what size to print, you're giving him the third piece of info.

What should he do when these three pieces of data don't match?

Assuming he's even willing to calculate the third, (most printers probably aren't set up to do this automatically), which two should he assume are correct in order to calculate the third? Or should you be forced to hand him a fourth piece of info, which is a statement about which of the first three should be considered flexible?

Remember, your job is to provide him with a set of pixels. And his job is to get those pixels rendered onto paper. Any variation in any of those three pieces of data will mean that some pixels are not printed, or that some are printed more times than others. In any case, he'd be changing your set of pixels. While that may be a convenient service for you, many artists would find this terribly inconvenient and would rather the printer simply reject requests that contain incompatible specifications rather than trying to guess how to fix them.

It gets more complicated when we include the possibility of vector based diagram descriptions or infinitely flexible/stretchable images which don't suffer from pixellation.

I've been in the Design/Advertising/Publishing sphere for over 30 years and Mike Wood and Robert Olding have got the goods on this subject...

Here's a few other things to tinker with:

Some years ago Agfa published info regarding resolution that goes something like this - the eyes of a person with excellent near sight can distinguish dots on the printed page (that's printed dots not screen pixels ok - can't print pixels) up to 266DPI. The rest of us might need a magnifiying glass to do so but after this point the image will appear as 'continuous tone' just like a photo printed from film. That (supposedly) produced the standard irrespective of whether Agfa really worked this out or some scientist or optometrist.

Then the pre-press world (those ladies and gents that take the designers work and get it ready for printing) decided a few more dots would cover any contingencies and we end up with a de facto standard of 300DPI in the print game. Nothing to do with pixels.

Raster Image Processors - RIP(s) - are used to change the pixels of digital images to dots that are printed on the page (we're talking photos/illustrations here - think tonal). Think of a RIP as a printer that processes the digital info (pixels) into analogue dots for the printing process.

Others have discussed the next level where LPI comes into play so let's leave that at "read the entire thread".

Increasingly now this data goes straight to a plate (film is dying out) which is then placed in the press and away it goes.

By this time of course the physical size of the work has already been determined by the designer/finished artist as already discussed in this thread.

I have many photographers supply dvds of images for motorsports each fortnight and each has their DSLR set to a different PPI - 72, 180, or 300. Now I've taken a lot of photos with my own Nikon D70 but I'm not a professional commercial photographer so this is just a guess - you should be able to adjust the PPI your cameras record at somewhere in the menu so if you are supplying work for print output it only makes sense to ask if the designer/Art Director prefers a specific resolution.

I resize images constantly and go to sleep at night wishing I had a buck for every time it's necessary but generally I try to keep images physical size at or near the intended output size BUT until I see the pics I'm not going to have any idea what that size is going to be (99.9% of the time). So perhaps the photographer is better to adjust the PPI camera setting to 300 and leave the re-sizing to those who are going to use the work.

Aperture has cut the time I take to process a few thousand images by half first go. When I say 'process' I mean import, select up to 20 images for an article, refine the selection to the dozen or so required, make colour corrections, change mode to TIFF, output to the required directory ready for placement and final sizing in InDesign.

The Adjustments HUD is a godsend. On my Powerbook Aperture runs fast, outputs to expectations. Now that I have better understanding of the structure of file management applied in Aperture I am finding it increasingly easy to use - bring on some of the bells and whistles mentioned in these forums!

Adobe Bridge offers nothing of consequence as a competitive environment, I'm not even sure why anyone would want to consider it here other than because "it came with CS so I already own it".

Finally, Aperture is just as good a tool for Art Directors and Designers (who need to work with large numbers of images) as it is for Professional Photographers and if those critics of Apple think the company goes to the trouble of running up web environments such as user forums to "NOT" read them you are being a little bullish, to say the least. Just enjoy them, keep offering up useful bits if you've got them and let's go forward.

Danb

Some years ago Agfa published info regarding
resolution that goes something like this - the eyes
of a person with excellent near sight can distinguish
dots on the printed page (that's printed dots not
screen pixels ok - can't print pixels) up to 266DPI.

There’s a huge difference between a halftone dot and stochastic screening. So we should define which you’re referring to. Output to a device like an ink jet (say an Epson) is using stochastic screening, you CAN on some images clearly see the difference between 266 and 360!

Raster Image Processors - RIP(s) - are used to change
the pixels of digital images to dots that are printed
on the page

A RIP takes vector data (not raster data if you will) and produces a raster (some use the term bitmap but that’s a file that is either black or white). It rasterizes math into pixels/dots. This could be sent to a device that produces a halftone dot or uses a totally different kind of screening.

I have many photographers supply dvds of images for
motorsports each fortnight and each has their DSLR
set to a different PPI - 72, 180, or 300. Now I've
taken a lot of photos with my own Nikon D70 but I'm
not a professional commercial photographer so this is
just a guess - you should be able to adjust the PPI
your cameras record at somewhere in the menu so if
you are supplying work for print output it only makes
sense to ask if the designer/Art Director prefers a
specific resolution.

That be nice but don’t count on it. There’s no reason we can’t have a setting on the camera that produces a desired DPI setting (it’s really PPI) and written as metadata in the file. But it’s hard enough to get the camera manufactures to provide the right color space EXIF data so I’m not holding my breath.

Aperture has cut the time I take to process a few
thousand images by half first go.

The could cut more by simply allowing us to insert a new or different resolution tag in rendered images.

Adobe Bridge offers nothing of consequence as a
competitive environment, I'm not even sure why anyone
would want to consider it here other than because "it
came with CS so I already own it".

Oh yes but Lightroom is a different story.

If people are given the option to change the PPI,
will they then ask for resampling? It's a slippery
slope. And IMHO will lead to feature-itis. I just
don't think it belongs in Aperture.

I understand this concern, but consider this: In Aperture 2.0, 3.0, etc, Apple will add other features, and it is likely that compared to this proposed features, some of those other new features will lead to steeper slippery slopes of feature creep than the "horror" of resampling could. If you sincerely want to prevent future feature bloat, preventing the addition of a resolution setting is not the more effective place to expend effort. For this particular feature, the benefit would outweigh the future risk.