The statement: "Unless an image is resampled (see Resampling), the amount of image data remains constant as you change either the pixel dimension or resolution." is misleading.
If the resample option is OFF, you cannot change the pixel dimensions. This statement also implies that if you COULD change the pixel dimensions, somehow the amount of image data would remain constant, which of course is impossible.
I'm sure the INTENDED wording was: "Unless an image is resampled (see Resampling), the amount of image data remains constant as you change either the DOCUMENT SIZE or resolution."
I've been trying to get this corrected for years. Please -- you have no idea how much confusion this small statement causes among newcomers to digital imaging.
Yes, of course you could crop the image, but why would you think that doesn't change the image data? That's precisely what cropping does.
In any case, the documentation pertains to the Image Size panel, and is meant to clarify the function of the controls in that dialog box. Since cropping the image does change the image data, even in this unrelated case the statement would still be wrong.
May be in this context by "pixel dimension" they meant not the pixel count but the size of the printed pixels which is expressed with pixel density resolution (ppi). On first read I thought it is OK but now I see it is confusing.
The "Document Size" in your corrected statement is not good either because it suggests the size of the file on disk or in RAM when open. May be Print size is a better term.
Great minds think alike. We had just changed the text to read "print dimensions" instead of "pixel dimensions." Because ultimately, this concept is relevant primarily for print output.
You should see this change online later this week.
Thanks, Mark! (Or as my dad used to say, "Great minds run in the same gutter...")
I've taught a few seminars on "resolution" and what various terms really mean, and although the basic concepts in digital images are pretty simple, there is so much mis-use of the terminology that most people are seriously confused. This seemingly tiny misprint was adding measurably to the chaos!
I've found that the last line of defense for a non-technical person's ongoing sanity is to stick to a couple of really basic principles:
1. Use the term "pixels" to refer to the virtual dots of color in the digital image file itself; virtual dots are just color numbers, and have no size.
2. Use the term "dots per inch" to refer ONLY to the physical spots of color from an output device, either the screen or the printed page; these have real and very different sizes.
3. When confusion arises, remember that the pixel dimensions of an image are the only reality. What you see on screen or on the page is an interpretation of those pixels, in the form of glowing or inked dots.
4. (And this is really key to understanding the whole thing) You can't judge an image's "resolution" without taking into account the viewing distance.
A "printed dot" on a billboard may be the size of your thumb, but from 300 feet away the billboard may be just as sharp as an HD TV in your livingroom or a webcam on your desktop monitor. And they may all have exactly the same pixel dimensions. So, in the end, it's best for non-tech people to avoid the term "resolution," and think of it as simply "apparent image detail."
IAC, I'm very glad to hear the docs are finally being corrected on this little point. "Print dimensions" does the trick (the units shown are all print-related except the one for "percent").
I perfectly understand pixels and resolution but to be honest I have never paid attention to the terms and this thread made me think about them. When I first read your post ashtangakasha, my first reaction was that you may be confused but then after checking the terms and some wiki pages I realized that certain terms mean certain things and you of course have a point. However, I have to say that I find even the established terms not very appropriate.
I think in the Image Size window if instead of Pixel Dimensions and Document Size, they use Pixel Count and Print Size, it would be much more clear to new users what these are for. I’ve seen a lot of people creating images without any plans to be printed, talk about problems with the size of their web images even after they have boosted the resolution to 300 ppi and above. If the Document Size section in Image Size is changed to Print Size this will ring the bell for those users.
As Mark pointed out, they are changing the labeling to "print dimensions" instead of "pixel dimensions." This is in the lower half of the Image Size dialog box.
Your suggestion to call the upper two fields "Pixel Count" would cause some confusion, since these values aren't really the pixel count. "Pixel count" would refer to the total number of pixels in an image, while the width and height are quite different. The pixel count is the pixel width multiplied by the pixel height. The two numbers are correctly labeled already, since that is how the "dimensions" of something are defined.
Whether or not an image is to be printed, the pixel dimensions alone don't determine the size of the output (screen or paper). The pixel dimensions only determine how many dots will be printed or displayed, but the size of the dots determines the final output size. If an image is 72 pixels wide, and the screen is 72 dots per inch, then displaying it at 100% will produce an image one inch wide. The same image, printed on a printer with, say, 216 dots per inch, will come out 1/3 of an inch wide. (Of course, the printer driver may be set to "fill the page" or otherwise scale the output, but that's another story.)
Your mention of "boosting the resolution to 300 ppi" may be related to your friends' problems. When you increase the size of an image, you're not adding more detail -- you're just using more dots to print the same original number of pixels. Although smoothing is usually done, the result will still appear less sharp. If an image isn't big enough, and is stretched (by the printer driver, by "enlarging" it, or by resampling it in the Image Size window), the stretched version will always show increased blurriness. The "DPI" or "PPI" setting doesn't increase the resolution of an image at all -- it just tells the printer driver how big you would like the image to come out. But in most systems, this DPI or PPI setting is ignored during printing, and Photoshop controls how much the final output will be stretched or reduced.
Finally, I think calling the Document Size section Print Size might be helpful if the user is actually printing the image, but as you say, often the user is just preparing an image for the web. Since "Document" is used consistently throughout Photoshop to refer to the current image, and since the fields are called Width and Height, and since the units of measurement are all physical output linear units (inches, cm, mm, etc., or percent), the meaning is pretty unambiguous.
If the image you're starting with isn't big enough for the desired output (e.g., 100 pixels wide for a full-width web page banner, which should be about 700 pixels wide), then you can stretch the image by changing the pixel dimensions and checking the Resample Image box. Or, if you know the DPI of the screen, you can set that field and enter a measurement -- but that's the hard way. Note that "pixels" are not available as a Document Size unit of measure -- that's because pixels have no size. Only output device dots have size. Pixels are just color numbers until they are printed or displayed.
I hope this helps clarify things a bit more,
as I said I perfectly understand everything you said and I wasn't asking for explanation about how things work but thanks anyway, it will be useful for those who don't know this. I was simply suggesting what I believe to be better terms for the sections in the Image Size window, although I may not be the best for this as English is not my native language. Next to "Pixel Dimensions" in the Image Size window the file size is shown which is the result of the pixel count calculation and this made me come up with this term. "Pixel Count" may not be the best suggestion for you but to me the word dimension implies size, and pixels as you said have no size until displayed or printed. And I don't think people will confuse it in the context of the Image Size window but when used in the help documents, instructions, and general information. When they hear "pixel dimensions" people who don't know the meaning in programs like Photoshop may think this is about the size of the displayed or printed pixels and not their numbers. May be the section can be called just "Pixels"
"Document size" doesn't sound very appropriate to me. On practice these settings are used for printing. Although this section can be used for calculating the physical size of an image on a display device if the screen pixel density (ppi) of the device is known, and vice versa to calculate the ppi if the physical size and number of pixels of the image are known, its typical use is mostly for printing purposes. May be the section can be called "Physical size". "Document size" when used in help documents, instruction, and general information can be easily confused for file size. From the document information popup menu, at the bottom left on the document window, one can choose Document size which means entire different thing.
I'm in so much agreement here. The Image Size Dialog is at best confusing, and at worst misleading. Part of it is terminology. Here is a revised Image Size Dialog I ginned up a few months back. Changes noted in blue. The term "resolution" I think should be relegated to some non-quantitative level. That is, don't tie it to a specific number. Instead use "Pixel Dimension" to describe the number of pixels in an image and "Pixel Density" to describe the size pixels will be when printed. Also, I think default setting in the dialog should be Resample unchecked. IMHO.
I like your suggested changes Charles. I think though the "Document Size" should be displayed in two states. When Resample Image is unchecked "Document Size" should be displayed inside the Pixel Dimensions section because the Print Size section won't affect it. And when the Resample Image is checked "Document Size" should be displayed outside the Pixel Dimensions section to indicate that all input fields will affect it. And I think a better place in that state should be either next to Resample Image or under the buttons like in the New document dialog to save space.
I completely agree that "resolution" nowadays has a lot of meanings that makes it very ambiguous. There are camera resolutions, screen resolutions, printer resolution, and image resolution could mean both pixel dimensions and pixel density and even it could mean pixel count (thanks to the desperate efforts of the marketing departments of camera manufacturers). "Pixel Density" is much better.
I also think that may be "Print and Physical Size" instead of just "Print Size" can cover the small percentage of cases when this section can be used to calculate the physical size of the image on certain screens, for example when designing for a specific iphone or pad apps.
Charles, I like your refinement of the Image Size dialog. It's subtle, but definitely helpful. Sadly, this dialog is extremely important and useful, but so many people are completely flummoxed by it.
My only suggestion for your version would be to change the "Print Size" caption to "Output Size". Perhaps a tooltip over "Output Size" could help clarify that this applies to screens (e.g., 72 DPI, 96 DPI, etc.) as well as to printers.
Visually separating the typical conflation of Document Size, Pixel Dimensions, and Pixel Density is brilliant, and would help keep these concepts "clean."
The changes for the New Document dialog also seem to be a helpful improvement. If the documentation AND the GUI are especially rigorous in using these terms with fanatical consistency, users who study the functionality and the explanations will get a lot farther down the road to understanding what these terms really mean and (more importantly) how they interact.
...My only suggestion for your version would be to change the "Print Size" caption to "Output Size". Perhaps a tooltip over "Output Size" could help clarify that this applies to screens (e.g., 72 DPI, 96 DPI, etc.) as well as to printers...
I did not realize that you could define PPI (output size) for screens. Screens are rarely 72 or 96 PPI and screen designers never have a need to define PPI or physical dimensions. Physical dimensions in inches or centimeters are meaningless. Screen designers only work in pixels.
As suggested, "print size" makes perfect sense.
Actually, physical dimensions for screen dots really do have a meaning.
Since all digital displays have some specific number of dots per inch (actually "dot clusters" or "cells), there is a very real physical size issue when designing for a display. For example, lots of software (Adobe included) assumes that a screen's "nominal" dots per inch is 72, but many recent monitors are closer to 100 dpi. The phones and tablets offer another significant range -- the Retina display (640x960 on a 3.5" diagonal yields a dpi of 326!) along with various other screen densities on other phones and tablets.
If you want to predict the viewability of a design on all these displays, you have to take into account the physical size of a screen dot.
Much of the confusion about pixels and "dots" comes from the origin of the term "pixel," which of course is a slangified variation on "picture element." Pixels are also called "pels" in some parts of the industry. So technically/historically, a pixel refers to the smallest unit of image display on a physical device, and all such pixels or pels have a real physical size.
Unfortunately, digital imaging professionals have now also taken the term to refer to the smallest unit of image data in a digital image file. This has caused a lot of confusion, but there seems to be no other appropriate word (that I know of).
Nowadays, it seems important to distinguish between the virtual pixels inside a digital file (which are of course just lists of numbers representing colors) and the physical pixels in a real-world output device (screen but not paper).
I personally advocate using the term "pixel" to refer to the virtual dots, and using the term "dot" to refer to the dots on your screen (which have always been called pixels). I guess it just seems more intuitive for people I work with to think of "real dots" and "virtual pixels."
Since the issues are the same between physical screen dots and physical ink dots on paper, I would love for them to be referred to by the same name, and that's why I encourage "dots" per inch (dots and inches both being physical). "Pixels" per inch does have an industry standard meaning with respect to hardware displays, but not with respect to ink on paper, where we talk about DPI and also (another historical term) Lines per Inch.
Adding to the confusion is the fact that "pixels per inch" has no meaning if you're referring to the digital image file itself (outside the context of printing or displaying) since the virtual dots are literally of no size (and the image could end up unchanged on a postcard or a billboard). So again, I think it's clearer to reserve the term "pixel" for the virtual, and the term "dot" for the physical. The only people this will really annoy are (hopefully) those who already clearly understand the issues. Those who are still trying to grok the true nature of digital images will not know the historical meaning (before it was coopted by programmers).
charles badland wrote:
and, a new New Document dialog:
How is your proposal particularly different from the current File - New dialog? Just the terminology?
Your wanting to move Pixel Aspect Ratio somewhere else might have merit for someone who doesn't need it, but what would the people who use it do? You don't need to see it - just don't show the Advanced settings.
Yep. Mostly terminology change, terms that (I think) make more sense for New Document dialog.
Plus it is a bit simplified through elimination of the disclosure triangle that revealed only two additional "advanced" options. (Seems kind of silly.)
I always want to see what my Color Profile is when creating a new document. But I never use the pixel aspect ratio. I wonder if it is feasible to move it to have that be somewhere else?
Marian Driscoll wrote:
...Screen designers only work in pixels....
They do but they may need to be aware of the physical size of the display in order to decide the pixels needed. For example interactive displays like bank machines, tickets consoles, even digital billboards, etc. - you need to now the physical size in order to make a functional design. The screen designers when editing image size will never uncheck resample image but they still may use the "Print size" section to calculate the needed pixels. Some iphones have abve 300 ppi. If the designer only is interested in the pixel dimensions and doesn't care about the physical size, if using 80 ppi monitor display, s/he may design buttons that will be too small to be pressed individually by a finger on the iphone.
My only suggestion for your version would be to change the "Print Size" caption to "Output Size". Perhaps a tooltip over "Output Size" could help clarify that this applies to screens (e.g., 72 DPI, 96 DPI, etc.) as well as to printers.
But there is no way of knowing the resolution or size of the output display device. That is why web designers deal in pixel dimensions, not PPI. Physical size on various monitors is neither predictable nor controllable.
charles badland wrote:
But there is no way of knowing the resolution or size of the output display device...
screen designers are not always web designers Check my previous post for some examples where the screen size and the ppi of the device is usually provided to the designer.
ashtangakasha and emil,
If I'm designing a web page for an iPhone 3 (480x320) and an iPhone 4 (960x640). They both have the same physical screen size. What PPI should I set for my document?
If I'm designing video graphics for 720p to display on everyone's TVs, what PPI should I set?
If I'm designing for a kiosk (bank machine, ticket POS), what PPI should I set?
This discussion has gone nutty. People that I thought had a grasp on this topic now seem to be thoroughly confused.
Marian Driscoll wrote:
... If I'm designing a web page for an iPhone 3 (480x320) and an iPhone 4 (960x640). They both have the same physical screen size. What PPI should I set for my document?...
You don't need to set any PPI for the document and they will not make any difference - you and everyone here knows that . What I'm talking about is when creating the design one should be aware of of what difference that PPI will make to the design between the two different iPhones. Because there will be a difference and while for most web design these things are out of control and thus irrelevant in certain cases they are not out of control. Yes, I have been part of designing a bank machine interface and that information (size and ppi) was provided to the design team. And yes this information is helpful and can influence design decisions. Photoshop can display correct physical sizes when choosing View > Print size if the monitor ppi is entered for Screen Resolution in the Units & Rulers section of the preferences. When I have a screen design job I often have two windows of the same image open side by side using Window > Arrange > New Window for.... One of the windows is at 100% zoom and the other is showing the Print (physical) size of the image on the eventual device. Being aware of the physical size does make a difference in making design decisions even though it could be a range of sizes that the design may need to cover.
emil emil wrote:
...Photoshop can display correct physical sizes when choosing View > Print size...
Don't you mean "View > Document size..."?
Photoshop is at odds with itself. If Adobe names a section "Document size" instead of "Print size" in the "Image size" dialog, then it should match the terminology in the "View" menu.
This is why Charles' suggestion to change "Document size" to "Print size" makes perfect sense.
Marian Driscoll wrote:
... This is why Charles' suggestion to change "Document size" to "Print size" makes perfect sense.
Yes I like "Print size" but my point is that screen designers may also use this "Print size" section to make Photoshop display the physical size of the image if information about it is available. It is not a typical use but still some people may use it for that and that's what made me propose in one of my earlier posts "Print and Physical size" for the name of that section.
Marian Driscoll wrote:... Don't you mean "View > Document size..."? ...
This is the menu. Again, to me more appropriate name would be "Print & Physical size" for that menu. The example in my previous post with multiple windows of the same image to use this feature for checking the physical size of the destination is an example that it could be used for non-printing purposes.
Your preferred approach to designing for various displays is, of course, entirely up to you. But the physical size of the device, the physical size of the dots on the device, and the typical viewing distance are all significant. Perhaps in the projects you work with, the significance isn't often a big deal, but in many situations it is important to know what your end-user will be experiencing. That's why there are so many tools appearing in the design profession for accomodating the proliferation of screen sizes and their respective pixel densities (i.e., the pixel dimensions of the display).
Here's an example.
Suppose I design, with pixel accuracy, a picture that fits exactly on a low-cost smart phone with a display that's 480 dots wide and 640 dots high, and (for the sake of argument) 4.5" diagonal. Suppose that picture contains some digitized text (not scalable fonts, but rasterized text). That image will be displayed "perfectly" (one dot per pixel) on the targeted screen.
Now imagine the same picture with rasterized text on a newer 4.5" display that's 960 dots wide and 1280 dots high. The original picture is 1/4 the area of the newer phone, but the new phone's screen is the same physical size, so the image will take up only 1/4 of the screen.
More importantly, it will also be only 1/4 as tall and 1/4 as wide. Rasterized text that might be readable with the bigger dots on the 480x640 screen may well be unreadable from the same viewing distance, because the dots are 1/4 as big and the picture appears to have shrunk. (If you get a magnifying glass and look at the new screen from half as far away, of course, the smaller image will contain exactly the same pixels as the old phone's screen.)
Whatever software is driving the 960x1280 screen may have "stretched" the picture to fill the whole screen, but that doesn't make the pixels involved any clearer. It only makes the pixels fatter. With the usual anti-aliasing, the fatter pixels will get "blended" so the stair-steps won't be as obvious, but the resulting image is necessarily blurry. It's like watching an old DVD on a new HD TV -- it may fill the screen, but it's way blurrier than HD.
However, if the designer creates two pictures, one optimized for 480x640, and another optimized for 960x1280, the end-user will get a better experience on a better display, which is what you'd want to happen. This is why Adobe has gone to such great lengths to accomodate this situation in CS6, especially in InDesign and Dreamweaver.
Alternatively, you could design just for the 960x1280 display, and take your chances when the driver for the lower-pixel-density screen "shrinks" the picture by stripping out the pixels it doesn't have room to display (which is often better than just displaying 1/4 of the picture and scrolling). But the end result will be a picture with detail removed, probably not as gracefully as you'd like. If a column of pixels being removed happened to align with the verticals in some small letters, the text could be rendered unreadable.
I'm afraid if you try to reconcile all the things Emil Emil is saying with all the things I'm saying, you will encounter contradictions, because I don't share some of Emil's conclusions (e.g., "You don't need to set any PPI for the document and they will not make any difference.") Perhaps due to language issues, some of his statements don't appear correct to me, but I'm not going to try to merge all the different assertions into one 20-page essay! If you'd like to have a private dialog to clear up some of this, I'd be happy to correspond off the forum.
.... I don't share some of Emil's conclusions (e.g., "You don't need to set any PPI for the document and they will not make any difference.") Perhaps due to language issues, some of his statements don't appear correct to me,...
What I meant is that the ppi information in the metadata of a digital file will not make any difference how the image containing such information is displayed on screen. "Save for web and devices" in Photoshop gives an option to remove the metadata when saving images for the web. For displaying on web browsers and other various image viewers this information is simply ignored. This information is used when printing the file and in image editing programs like Photoshop this information will be used to display the physical size of the image on the ruler and also when using viewing functions like Print Size as I demonstrated earlier.
However for screen designer the ppi information is useful for the purpose of finding the physical size and also to be aware of the pixel density. For finding one of these three properties: pixel dimensions, physical size, and pixel density (ppi), only two of them are needed, Photoshop will show the third one when two of them are entered in the Image Size dialog.
I'm unable to follow your smartphone scenario. Neither Android nor iOS developer documentation really discusses DPI/PPI. Their guidelines are for pixel dimensions.
I appreciate that we have ways to simulate physical size on our desktop screens, but as noted earlier, we can rarely even guess the viewing distance for the end product. Our perception of scale varies greatly between holding our phone at arm's length vs. holding it to our face.
Yes, I completely agree with you that the declared DPI in the metadata is nothing more than an annotation, which is usually ignored. And I also agree that the Image Size panel is indeed useful for calculating pixel dimensions or output size based on DPI (even if the designer ends up not changing anything). Sorry for any confusion.
The DPI/PPI is really implicit in the combination of the physical screen size with the pixel dimensions of the screen. You can get the DPI by dividing the screen's wider pixel dimension by the screen's wider physical length. (Since screen sizes are usually given as a diagonal, the length & width require a quick trip to Pythagorus. Another important factor is the "most likely" viewing scenario. That is, if the screen in question isn't a phone, and is very large, then having lots of pixels doesn't necessarily mean that text on the screen will be unreadably small. Unless the user is viewing the screen from a great (relative) distance.
In practical terms, although we can't know the viewing distance, it's a safe bet that a smart phone won't be viewed from more than arm's length. With the iPhone Retina display, however, whole web pages can be viewed quite comfortably if close-up reading glasses are used. But the small print on those pages will be unreadable at, say 12 inches. On the other hand, I have a stopwatch app which I use for radio program recording sessions, and the numbers are as big as possible so they can be readable from across the room. On a phone, they're still too small, but on a tablet they're great.
How important this is just depends on what you're actually doing with the display. A measurement app which shows a ruler on the screen had better be adjustable so that the DPI will match real world inches. A phone dialing app need not be very concerned, as long as the text is "sufficiently" large. Apple and others have taken some pains to accomodate the big discrepancy in DPI between the earlier iPhones and the Retina model(s) by providing both size limits and automatic scaling.
And of course, vector graphics, which usually includes fonts, can be scaled automatically without any loss of clarity at all.
But remember, being concerned with DPI is only one way of being sure your design will look good to your type of user on various displays. The DPI setting inside the image file (as set in the Image Size panel) doesn't, by itself, have anything to do with how the image is displayed. As Emil pointed out (and I initially misunderstood him), the DPI you set in Photoshop doesn't change the image in any way -- it just sticks the DPI number into the image metadata. If you turn on the Resample Image checkbox, and then change the pixel dimensions or the combination of DPI and output size, the image will get resampled and become actually bigger or smaller as shown by the new pixel dimensions. As Emil and others have pointed out, sometimes we use the Image Size dialog box just as a calculator to check the "effective DPI" (or the effective print size), even though we may not plan on changing the size of the image.
Everyone has some good points here. The conversation actually highlights some of the resolution terminology inconsistencies I've been grumbling about for years. But one thing, just to avoid confusion with printing terminology, is to use PPI when talking about pixel size. DPI is historically associated with the size (or frequency) of ink drops on paper... and I think it adds to the existing confusion when you use "DPI" to describe "PPI" concerns. (Many "Ds" can be used to print one "P")
That's a very good point, Charles, and in principle a agree. I usually have a more orderly presentation about this stuff, and include some discussion about droplets, print cell dither patterns, rasterization engines, etc. I agree that in that context "dot" can be misleading, but I find that most nontech people I talk to have no awareness or interest in the droplet, and don't really need to know about it, so, if possible, I don't bring it up.
I've always used "dots" for screen printing, which is a completely different beast, but now that "lines" has became more or less arcane, I've recently been encouraging nontech people to think just of dots, for both displays and print, even though displays really use cells of 3-4 dots. The screen image is more analogous to print than to the virtual datapoints in a digital file, so I've been trying to reserve "pixel" for the data elements. I'm reluctant to use "pixels" for printing, with nontech folks, just because it leads to the misunderstanding about the a data pixel having a size. Although clearly the other kind of pixels DO have size if we're talking about sensor elements or display elements. In truth, we really need another word for the dimensionless points of color that make up the "reality" of a digital image. Datsels?
Maybe I'm crazy, but when I try to stick to what's technically correct in one field, I always run into someone from another related field where the jargon is technically NOT correct. "Resolution" is a good example of that, especially if there are a couple of scientists or engineers within earshot: Display resolution, Optical resolution, Angular resolution, Sensor resolution, Image resolution, Color resolution, etc.
There's a great essay on what pixels really are, by the famous Alvy Ray Smith, one of the pioneers of digital imaging (PARC, LucasFilm, Pixar, inventor of the Alpha channel, etc.). "Pixels are not little squares!" is one of his favorite cries. You can download that (and other papers) here: http://alvyray.com/Memos/MemosCG.htm#PixelIsNotSquare. It's technical, delightful, fascinating, and more than a little illuminating.
"View Print Size" only works accurately in Photoshop when the User has set their Screen Resolution correctly in the Prefs. as was mentioned previously.
Two ways to find the resolution of your monitor:
Open a new 500 pixel x 500 pixel file in Photoshop, and set it to 100% (view actual pixels).
Take a ruler and measure the image on screen in inches.
Divide 500 by the measurement.
This is the resolution of your screen in ppi.
(For example, a NEC 2690 displays 88 pixels per inch.)
However, if you have Acrobat Pro, fire it up because it does the calculation for you automatically.
Thanks for the link. Very interesting article and site. It is interesting to think of a pixel as just being a point sample... and the square.. I guess almost an artifact derived from the point sample.
I work mostly with inkjet printing and the square pixel model has always worked well for me. His article is a bit dated (1995), but he comments:
There is a new technique for printing ink-on-paper that uses stochastic patterns
within each little square. Hence the separations do not have to be rotated
relative one another in order to “show through”. The little square model is a decent
model in this case.
When I have the time, I want to go over his other pdfs. I'm sure most will be over my head, but I hope to learn something...
Allen, I appreciate that some feel it is important to visualize device screen size via PPI tweaks to their 'print size' display but the whole idea is still not so useful in the grand scheme of things. Even in the print realm, many of us cannot show 'print size' for a portrait tabloid sheet. Designers are already accustomed to understanding scale. Large format print designers do not work on 8 foot computer displays.
My entry into this discussion regarding the 'print size' label suggestion was based on the majority of new users confused by it as evidenced by posts in this forum and the wealth of folk on the internet claming that web graphics need to be 72ppi. I also take issue with the inconsistency between the "Image->Image Size->Document size : View->Print Size' options. I was simply agreeing with Charles' suggestion. While it does not fully address your and Emil's concerns, it would be a huge improvement over the current situation.
Marian Driscoll wrote:
.. While it does not fully address your and Emil's concerns, it would be a huge improvement over the current situation.
I fully agree with this. In fact I was the first one who suggested in this thread just the name "Print size". Later I also suggested "Print and physical size" as a term that will cover all eventual uses of this section which may not be limited all the time to printing. And I was also thinking about the Adobe people who may be reluctant to change it just to "Print Size" for this reason and leave it as the current "Document Size" which doesn't look appropriate for everyone here. And I also think that "Print Size" will be most helpful because the more advanced users will know it can also be used for displaying physical size while the beginners my get confused by the addition of "Physical" in the term. Just "Print size" will make it very obvious that this settings will not affect anything if the image is not going to be printed.
@Allen, I read some parts of the article in your link and while it is interesting, if your purpose is to keep things simple in your communications with beginners, I would advice to stick with popular established notions about things like shape of pixels. Raster images are coded to describe a virtual 2D grid on x,y coordinates with pixels. Pixels as coded in the file by describing their color and coordinates and nothing is specified about the shape. Software engineers can design any interpolation method to render the pixels in any shape they like but the most practical interpolation is a square shape. And this makes it fair enough to think about the shape of the pixels as squares. This can be demonstrated by zooming in a digital image on screen or printing it with low resolution (pixel density). What other shape would make more sense?
Regarding ppi vs dpi, when talking about monitors, after the LCD monitors replaced CRTs most people use the term "monitor pixels" and not "dots" to describe the smallest picture element (cells). And this also comes from the fact that lcd panel technology is based on a very similar idea to the virtual 2D raster image grid. It is simply fixed size cells on a grid. The CRT monitors used RGB guns to shoot light at a point to render a monitor pixel and at the beginning the term dpi was used to describe the resolution of CRT monitors. At early 90s all monitors were with fixed resolution of 72 dpi until Apple created the first multi-resolution monitor I think in around 1994. Later the LCDs started to replace CRTs and during the transition period people were using both terms dpi and ppi to describe monitor resolution. Now CRT monitors are no longer in production and the term ppi is widely accepted and considered correct when describing monitor resolution.