This topic (dpi and web graphics) recently hijacked another thread (Node 31800, about half way down the first page), but here's where we can continue the discussion, if necessary...
Re-post from previous thread:
This is a subject that makes me cry and reminds me of how bad I am with numbers. I had both my boss and a co-worker sit me down not that long ago and help me understand the difference. I still don't fully grasp it. I live in a dpi print based world. But apparently the web doesn't care about dpi, only pixels.
DPI and PPI ExplainedDisplay, Printing, dpi and ppiUnderstanding ResolutionThe Difference Between DPI and PPI
My understanding - and I'm not a web designer - is that when preparing graphics for the web you should always be working in 72 dpi (or ppi). That way you are dealing with the images at the size they will display on screen. And, when creating tables and such you thus already know the pixel dimensions of your images without having to do any math. Deviating from that standard can only cause confusion IMHO.
But... a question pertaining to topics raised in the other thread - does it affect image loading speed if the graphic is at a higher dpi (same pixel dimension)? Not that I'd do that but it was discussed earlier.
This might be an silly question. But, if you use the "Save for web" feature in photoshop/illustrator, does it default the DPI?
I've never really realized there was a difference between pixel density and DPI. Maybe I need someone to sit down and explain it to me.
It defaults to 72 dpi (I think) but if your image is at a higher dpi than that, it may make the image proportionally larger to maintain the pixel dimensions. Not sure tho.
"Save for Web" defaults to 72 dpi AND respects the size it is at in the document.
i rest my case.
Do you know this site: www.scantips.com/no72dpi.html (Say no to 72 dpi)
And maybe I can help some people with my article: Misunderstandings about dpi - www.dpiphoto.eu/dpi.htm
continuing from the other thread...
"no, it only exists in photoshop."
Ah, that makes sense.
All that setting the image to 72dpi is makes sure that your 100% = 1 pixel of screen for every pixel of image. I assume your video editing software when it sees a 72dpi just then doesn't bother resizing it to ensure that every pixel you created in photoshop is brought over as a pixel in your video image.
It's a setting that you need to have, but really isn't related to print dpi what-so-ever. Just an old habit of photoshop's. ;o)
And now back to this thread...
"My understanding - and I’m not a web designer - is that when preparing graphics for the web you should always be working in 72 dpi (or ppi). That way you are dealing with the images at the size they will display on screen."
The dpi really doens't matter. It's just a particular setting in photoshop to make it easy to know when you are truly seeing all the pixels (100%). The image, itself, retains no dpi info online.
"But… a question pertaining to topics raised in the other thread - does it affect image loading speed if the graphic is at a higher dpi (same pixel dimension)?"
Not at all. The dpi setting isn't used at all by the web browser. All it cares is what the actual image is in pixels.
"Maybe I need someone to sit down and explain it to me."
I'm trying but probably not fully getting the point across. Toss out some more questions and I'll try to clarify! ;o)
i can't believe i'm still in this. i think we have a disconnect between three worlds: 1) strict technical language, 2) misleading naming conventions in software, and 3) actual user experience.
i think the question of loading time relates not to "dpi per se" but rather to file size.
a 300x300 image at 300ppi (photoshop language) is a much larger file than
a 300x300 image at 72ppi (photoshop language).
if i take the 72ppi version into video, it comes up at 300x300 pixels on the screen,
but if i take the 300ppi version into video, it's much larger.
I suppose I should add that there has been a longtime wish to have DPI settings online. At one point, the PNG format had a proposed multi-resolution option. The idea that the PNG could serve up the low-res version for on screen viewing, but if you wanted to print it, you could send a higher resolution image (at the same physical dimensions) for printing on paper. It was a great idea, but still hasn't been adopted.
Patty, if you look at the examples I posted and check the image properties, you'll see that both of the graphics (1 @ 72dpi and 1 @ 300dpi, both 300 x 300px) are the exact same file size. Bizarre, but true.
you’ll see that both of the graphics (1 @ 72dpi and 1 @ 300dpi, both 300 x 300px) are the exact same file size. Bizarre, but true.
I think that's where I'm lost, how can you control the DPI aside from the resolution?
Well, to use the examples set prior.
If I have a 300x300 pixel image with a resolution of 72ppi and I have another 300x300 pixel image with a resolution to 300ppi.
In this example, how am I controlling the DPI? All I've changing is the pixel resolution. A 300x300 pixel image at 300ppi is equal to a 1250x1250 pixel image at 72ppi, right?
The point I think Darrel is trying to drive home to all of us is that the web doesn't care about resolution, only pixel dimensions. As in my example linked above, two files with different dpi display exactly the same if the pixel dimensions are set the same. It's funny that all of us are clinging so tightly to the "300dpi means larger dimensions" thing, and I can only guess it's because that model is true for print.
Andrew, the crazy thing here is that there's no need to "control the DPI." We're just so used to having to control it exactly for print that we simply don't want to let go of that control with web graphics, even though, according to Darrel's comments and the proof in my examples, we don't need to worry about it at all.
"A 300x300 pixel image at 300ppi is equal to a 1250x1250 pixel image at 72ppi, right?"
It CAN equal 1250x1250, if you resample it to do so, but apparently it doesn't necessarily equal 1250 x 1250; that is, you can simply "tell it not to" by changing the pixel dimensions.
but in what environment were your test images finally rendered to us ?
yes granted, on the web, they are the same.
but at what point in the process was one a larger file than another ?
and at what point was the larger one reduced to match the smaller,
and how many different ways are there to do that,
and which are lossless and which are lossy ?
in photoshop and aftereffects, there is no question that dpi (ppi) is directly linked to
file size and relatively linked to image quality.
"I think that’s where I’m lost, how can you control the DPI aside from the resolution?"
You don't. It's just a bit of meta information that says 'hey, when you print this image, print it at this size on paper'. It just controls the SIZE of the pixels when printed.
On screen, it's irrelevant.
"The point I think Darrel is trying to drive home to all of us is that the web doesn’t care about resolution, only pixel dimensions."
"in photoshop and aftereffects, there is no question that dpi (ppi) is directly linked to
file size and relatively linked image quality."
Nope. File size is determined by the number of pixels in it. I think that's where you're confused.
These images would be the same file size:
100x100 @ 100dpi
100x100 @ 1000dpi
printed from photoshop, one would be 10 times smaller, but they both have the same number of pixels and take up the same amount of memory on the computer. One would print 1" square, the other .1" square
Now, these images WOULD be different file sizes:
1" x 1" @ 100dpi
1" x 1" 2 1000dpi
They'd print the same size, but the first would be 100x100 and the second would be 1000x1000. The second would be 100 times larger in file size.
To clarify (hopefully!) the reason you see a file size difference, Ch (I assume) is that aftereffects retains a physical size to your image. Film, for instance, doesn't have a resolution. I assume you can output your afteraffects file to film at any resolution you like. Either way, it has to fit the physical size and aspect ration of a film frame.
So, if you retain a physical 'printed' size to an image, and increase or decrease the dpi, yes, it will increase or decrease the file size, as you are adding or removing pixels.
However, if all you care about are the pixel dimensions (as one is on screen) then increasing or decreasing the dpi won't change anything on screen. It's still the same pixels.
this just making me crazier. i insist we have a language gap here.
>>“in photoshop and aftereffects, there is no question that dpi (ppi)
>>is directly linked to file size and relatively linked image quality.”
>>Nope. File size is determined by the number of pixels in it.
>>I think that’s where you’re confused.
your response doesn't address the point. the operative detail is: pi. per inch.
the amount of information per inch can vary. more information is more information.
more information is a bigger file. higher resolution means more information. higher resolution means bigger files. i am dealing with this is a big way these days as my world transitions from NTSC to HD.
I still know where you're coming from Chris; when I tested Darrel's theory and proved his results I was shocked. So here's what I did.
I started with a 300dpi image that was 1600x1600 pixels, let's say the file size was 1MB. Throw that on the web and, sure enough, it fills the screen and is slow to download.
Change the resolution (without resampling) to 72dpi (thus keeping the image at 1600x1600px). This is where I've always been mistaken: I thought this would reduce the file size, but it doesn't. But...
Resize/resample that 72dpi graphic down to 300x300px (that is, change the pixel dimensions) and the file size drops, let's say to 75K, and now it displays at the size I want (300x300px).
Now, here's the weird part, go back to the 300dpi/1600x1600px image, make sure "Resample Image" is checked, and simply change the pixel dimensions to 300x300px, leaving the dpi at 300. This is the crazy part: the image is now the right dimensions for the web, and the file size is the same as the twice worked-over version above, only 75K.
>>I started with a 300dpi image that was 1600x1600 pixels, let’s say the file size was 1MB. >>Throw that on the web and, sure enough, it fills the screen and is slow to download.
>>Change the resolution (without resampling) to 72dpi (thus keeping the image at >>1600x1600px) and the file size drops somewhat (let’s say 200K).
my point exactly.
>>Now, here’s the weird part, go back to the 300dpi/1600x1600px image, make sure >>“Resample Image” is checked, and simply change the pixel dimensions to 300x300px, >>leaving the dpi at 300. This is the crazy part: the image is now the right dimensions for >>the web, and the file size is the same as the twice worked-over version above, 75K.
when you downrez and then uprez again you lose information. you cannot recreate the information that was in the original. the resulting file can be compressed much more easily than the same size original.
i've also noticed that photoshop defines file size in inconsistent ways that i don't really understand. file size and upload time may also be linked to compression codecs. i'm not an expert there either.
>>“DPI has nothing to do with file size.”
i think this statement may be technically true but functionally misleading,
because in photoshop ppi is linked to image dimensions. it depends how you are using the information.
in my video work the final product is lossless, uncompressed.
i constantly see changes in file size relating to resolution.
anyone at adobe care to weigh in on this ? i'm officially out of my league.
i just know what i have to deal with on a daily basis. maybe it's all a dream. i love you all.
Frankly I don't care if the web "sees" dpi. Photoshop cares. So it seems to me that if I'm creating graphics/pix for the web, in Photoshop, I should use 72 dpi. If I need the same image for a print project, I use 300dpi. While both images might be the same size, in inches, they will be very different in pixel dimensions.
If anybody is confused you can see the diff b y going into View in P-shop - view at "print size" and "actual pixels". If you're using 72 dpi those two pix will appear the same on screen. I realize that you can upload graphics at other than 72 dpi as long as it is the same pixel dimensions but why would you do that?
This is making my head spin. I'm sorry I started it!
I agree Patty, I feel like my entire understanding of the world I live in is being questioned! I totally agree with what you've said, but I find it fascinating how wrong I've been about this question as far as web graphics go. And the idea that:
"DPI has nothing to do with file size."
That makes me shiver to say it, but, despite what I thought I knew, it seems to be true.
The actual files for bitmapped graphics, if you look at them (for example by opening one up in a hex editor and comparing against the JPEG spec), simply describe pixels. A JPEG file, for example, is defined in terms of a rectangle that is X pixels wide and Y pixels high, and consists of a compressed summary of X times Y pixels. The information about pixels per inch is purely optional, as a way of telling printers how big to print the thing. It's not really part of the data at all.
Video may be somewhat different because the compression algorithms are more complicated and may want to take into account how large it will appear to be.
Frankly I don’t care if the web “sees” dpi. Photoshop cares. So it seems to me that if I’m creating graphics/pix for the web, in Photoshop, I should use 72 dpi.
Photoshop cares if you're going to print it, but not if it's going on the web. It assumes 72 dpi for web use as a sort of arbitrary convention - that number describes the original 1984 Macintosh screen. Computers now generally have smaller pixels (a 20-inch iMac, for example, at 1680x1050, is around 98 ppi). The main thing, from Photoshop's point of view, is that a 72-dpi image will either have small enough pixel dimensions to work on the web or it will have large enough spatial dimensions to alert you to the problem.
The whole thing is easiest if you acknowledge that print and web are different beasts. Print works in units of area, so dpi is absolutely essential, whereas the web works in pixels, so dpi is totally irrelevant - what matters there is the number of pixels.
Chris, note that I edited the first statement you quoted. Changing from 300dpi to 72dpi does not change the file size unless the pixel dimensions are also changed.
"more information is a bigger file. higher resolution means more information."
YES! But we're talking *on screen* resolution. Which means 1 pixel of image information = 1 pixel of the screen.
Changing the DPI of an image doesn't necessarily change the amount of information. 300x300 is and always will be 300x300 regardless of the DPI setting you give it. If you RESAMPLE while changing DPI, then that WILL change the file size, because you add and/or subtract pixels. But that's a result of the resampling specifically.
"Frankly I don’t care if the web “sees” dpi. Photoshop cares."
Photshop doesn't care either. YOU may care because it's nice to know that 100% in photoshop means 1 pixel of image = 1 pixel on screen if the image is told to print at 72dpi.
"So it seems to me that if I’m creating graphics/pix for the web, in Photoshop, I should use 72 dpi. If I need the same image for a print project, I use 300dpi. While both images might be the same size, in inches, they will be very different in pixel dimensions."
That is true. If you are dealling with real inches in print, that's when DPI matters. (even then, Photoshop doesn't care, but that's another thread...)
"“DPI has nothing to do with file size.”
That makes me shiver to say it, but, despite what I thought I knew, it seems to be true."
Well, here's a better way to say it:
DPI, by itself, has nothing to do with file size.
If you combine DPI with wanting output at a specific pixel dimension, THAT's when file size changes because you are adding and subtracting data from the image to accomodate the specific DPI at a specific print size.
Or maybe that's just more confusing. ;o)
Ch...I too would love to hear from someone at Adobe. I do believe you that DPI makes a difference in your work. I'm sure it has something to do with After effects assuming some sort of 'virtual' frame size in real measurments.
As far as I know, Darrel is right. The dpi is totally unconnected from file size. The only thing that has influence on file size, is pixel dimensions. So only when you resample images in photoshop when changing the dpi-settings, file size will change.
AAARRRGGGGHHHHH. why do we keep citing this data in isolation when obviously the relationship between ppi and file size comes into play RELATIVE to size on the screen ???????? call it inches or pixels. it's data per display size.
every photoshop user deals with this everyday.
Darrel finally said it best: "DPI, by itself, has nothing to do with file size."
operative term: BY ITSELF. but we don't interact with this information "by itself".
in the world of production it is totally relevant data.
to me this is like saying the numbers 3 and 9 on a postage stamp, by themselves, have nothing to do with the value of the stamp. right. but relative to currency and the change in my pocket and my desire to send the letter, it is relevant data.
Chris, I tried to read all of this and I feel your pain.
I hate to potentially add confusion, but just like scaleable text, scalable images are one of the next big things to sprout on the web thanks to CSS and JS. Check this link:http://www.xs4all.nl/~sbpoley/webmatters/resize.html
"AAARRRGGGGHHHHH. why do people keep citing this data in isolation when obviously the relationship between ppi and file size comes into play RELATIVE to size on the screen ???????? call it inches or pixels. it’s data per display size."
Chris, I had to read this several times to figure out how you could think there is any relationship between ppi and file size. I think there may be a cultural gap here that's hindering communication between people who just use images and people who muck about with file formats. A Photoshop user might think of pixels as just the things visible on the screen, and might think the underlying data are something different. This would seem like a sensible assumption, because it corresponds to the reality of preparing image files for printing. But actually file formats such as JPEG and GIF consist of "abstract" pixels that ideally correspond to the pixels on your screen. There's a file header that contains various information about the file (including, in JPEG/JFIF format, 4 bytes that define the ppi values); everything after that file header is just a compressed or uncompressed array of pixels. When you open up a JPEG in Photoshop and then select "Image Size", the display tells you the dimensions in pixels at the top and the dimensions in your units of choice (inches, centimeters, etc.) below. The pixels in the upper half of the Image Size dialog are the image data - no matter what dpi the Resolution box shows, the amount of image data is equal to the pixel dimensions.
When Photoshop (or some other image viewer) displays an image at 100%, that means that one abstract file pixel is equal to one screen pixel.* When a Photoshop file is sent to a printer, though, it doesn't work in pixels - it works in units of area. So when the file gets printed, it is converted into different information based on the dpi setting in the file. So you are correct that the amount of data depends on ppi in the world of (print) production, but Darrel is also correct in saying that changing the ppi of a file has no effect on the file size unless you also change the pixel dimensions.
Web browsers will display the image at 100% unless forced to do otherwise, like in the example from Patty's site that started this thread. So for web use, the size on screen depends solely on the pixel dimensions, and ppi has no effect. This is very different from how images are prepared for print in Photoshop, but it's the reason why, in the alternate-size image example that Dan posted, it's necessary to have separate files for each size.
* Obviously, if the image is being displayed at something other than 100%, either some pixels of the file data are being dropped out or additional pixels are being extrapolated. This doesn't affect the file size, because the number of pixels actually stored in the file remains the same.
i give up. it's not about the abstract, it's about the practical reality.
ppi RELATIVE TO IMAGE SIZE affects file size. in my life, in my dimension, in my universe. or am i just dreaming about this computer in front of me ? sometimes i think so.
we could take this all the way into the abstract and say that the pure data doesn't really exist at all, it's just modulated electrons; pure energy. so what ? what matters is what they mean to us in context. i have to deal with the interdependence if PPI and IMAGE SIZE.
i never deal with them in isolation. i deal with their practical relationship, with FILE SIZE the third point in the triangle. call it a game, call it a naming convention, call it a day.
i need a drink.
(i changed "people" to "we" because it's all about... us.)
I think Chris is frustrated because you guys are implying that if I were to scan a six inch square image it would make no difference in file size whether the image was going to be scanned at 72 ppi or 600 ppi or 1200 ppi.
I don't think that is what you guys are saying, but I think that is what some people are hearing.
"ppi RELATIVE TO IMAGE SIZE affects file size."
I don't think anyone has disagreed with this. I probably shouldn't have used the word "abstract", except that I wanted to make the point that the pixels you see on screen aren't necessarily the pixels in the file.
Here's a simple summary:
file size = pixel dimensions
on screen, image size = pixel dimensions / display ppi [nominally 72]
in print, image size = pixel dimensions / image ppi [whatever the ppi setting in the file is]
I think we've all been saying this, but in different ways and with different emphases that made it look like disagreement.
"I think Chris is frustrated because you guys are implying that if I were to scan a six inch square image it would make no difference in file size whether the image was going to be scanned at 72 ppi or 600 ppi or 1200 ppi."
Yes, that may be the problem. Darrel and I have been talking about pixels not inches, but people who think in "paper" terms think in inches, not in pixels.
yes, the scanning reference is apt. that's a practical example, no matter what the engineering book says. the equivalent example exists as i import psd images into other platforms.
all technical definitions aside (ha ha), here's the simplest example i can give:
i have a photoshop image 1280 x 760 @ 300ppi. i place it unaltered in illustrator.
it comes in natively at a certain size on the ai doc.
i have a copy of the same file. i convert it to 72ppi without changing pixel dimensions
(1280 x 760). i place it in the illustrator doc. it comes in much larger than the higher rez file because the same amount of visual "pixel" information is spread over many more inches, because there is less information defined per inch. if i want it to be the same "placed size" natively, i have to make it a smaller file. less information, lower resolution
of the original visual information.
the exact same thing occurs when i bring images into aftereffects. so video is more like print ? - it is in effect a print medium with a preset resolution...
i just have to think of dots, pixels, data PER INCH as a defining aspect of resolution and file size, because my output medium handles it that way. lets call it a language problem.
if it's some arbitrary definition within the arcane world of adobe / video, so be it.
functionally, for my work, the video screen remains 72dpi, and i prepare elements according how they look in the working environment. and the higher rez images are just
bigger files, no matter who is asking.
you wanna see a grown man cry ? i can't wait to meet you guys... hugs all around ;-}
This discussion really accentuates the different point of view of this matter. Personally, I've grown up with the "pixel" outlook on computer images, and hence I often think that the resize dialog in more advanced image editing software, like Photoshop, is overtly complicated. "Inches? DPI? Why would I care about those arbitrary values? I want control over the actual number of pixels, because that is what is relevant for how much information is stored in the picture." As others have commented, for web-use (today), the numbers of actual pixels in the picture is what actually matters. If you are thinking in terms of inches and DPI, you are making the matter unnecessarily complicated for yourself. Unfortunately, Photoshop and the like tends to enforce the concept of DPI (as the discussion about 72 DPI shows).
I don't know if I can add much to the discussion apart from what has already been said, except to note that the term "resolution" sometimes is used in two very different meanings, which might be a source of great confusion. Sometimes it refers to the number of dots/points per inch, DPI/PPI, and sometimes to the raw number of pixels without any notion about physical length (for example when talking about computer screens, or digital cameras).
Personally, I’ve grown up with the “pixel” outlook on computer images, and hence I often think that the resize dialog in more advanced image editing software, like Photoshop, is overtly complicated. “Inches? DPI? Why would I care about those arbitrary values? I want control over the actual number of pixels, because that is what is relevant for how much information is stored in the picture.
I agree. I think of the pixels first, and the "real world" image dimension measurements (ppi/inches/cm) as extra info in the file which other applications (or Photoshop's Print dialog) may or may not look at when I reuse that image in other contexts. Which is how the info is actually stored. I don't think that GIFs don't contain any real world size info at all.
Yes, it's handy if I know all my images in a print project have been tagged as "300ppi" so when I scale them to 150% in Quark, I know that they will come out at 200ppi (plus you can't scale smaller than 10% which is a problem if your images are tagged as 72ppi). But it is not a property of the image itself.
even in various contexts, "resolution" is a term that necessarily implies unit size and amount of data, relative to eachother. whether you call them pixels per inch or dots per inch or grains per cm or blocks per mile or bits per second...
only in relation to data density per unit can you discuss the relative resolution of different documents. a higher data density within the same unit is more information, and (if that information requires storage without compression) its a larger file. more information is just... more information.
the photoshop convention may be "unreal" but it works well enough across platforms,
and provides clear enough standards that i can manipulate and anticipate results.
to say "dpi has no relation to file size" in the abstract sounds more unreal to me than the
mythical pixels per inch because, as bluestreak wrote, when i scan a 6 inch doc at 300dpi
i get a larger document than when i scan a 6 inch doc at 72dpi.
the word resolution refers to different standards of measurement at different stages in the process; input, manipulation, output. at various stages it may entail file size, image quality, or interpolation, but not necessarly all at the same time.
and because screengrabs and video grabs still come in at 72ppi that remains a functional
convention for working with screen resolution. is it true ? not really but it works for me in many situations until a better uniform standard is in place.
the truth is... the electrons are having a big party in there and bumping into eachother
at unimaginable speeds, playing drinking games, singing kareoke, and making snow angels in the documents. if you listen closely to a really hi-rez document, you can actually hear them singing. i have a lawnmower. can you swim ?
There's the difference right there. Maybe you upstart web kids should have a bit more understanding of "old media", or print. And take a broader view. Photoshop deals with DPI and inches because that information is EXTREMELY relevant for print media. I may use the same image in print (300 dpi) in a pdf or comp (150 or 200 dpi) and for a web graphic or jpeg to email for approval (72 dpi). Images need to be scaleable both in terms of real size (inches) and resolution (DPI or PPI). If I had to make those calculations in pixel dimensions I'd go nuts.
If you are thinking in terms of inches and DPI, you are making the matter unnecessarily complicated for yourself.
Only if you are talking about the web. But I DESIGN for the web, do not program for it. I create my files in photoshop. And yes, once I have a grid figured out I resize various elements using their pixel dimensions but it still feels very unnatural to me, coming out of a different design tradition, not to think in terms of inches, at 72 DPI.
So I second what Chris says - it's not about the abstract but about practical reality.
And I will add one more thing - if you take the same file, same file size (in pixel dimensions) and place it in a design program such as Quark or InDesign, the clarity with which it will display will be very different if you place it at 300 dpi and scale it up or at 72 dpi, full size - in inches.
Maybe you upstart web kids should have a bit more understanding of “old media”, or print. And take a broader view.
I don’t know if I should regard “upstart web kids” as an invective or not, but in any case, I maybe should clarify that my whole message was about images on the web (which is the subject of the thread), and about resizing them, as opposed to images to be printed. Of course DPI is very much relevant for printed matter; it would be silly to imply otherwise.
I must admit I am not fully following you here, but I don’t know if it is because you are saying something blatantly obvious, or something very subtle. In your example, do you mean that both of the pictures are to be printed in the same size (in inches)? Which, then, would you say, has most clarity?
Another reason why 72ppi could confuse - when I had an old CRT monitor, that was about 72 ppi, but on my current laptop 100ppi would seem to be closer to the resolution on the screen - so don't expect your 72 by 72 pixel image to show up as one inch across on every web user's screen - for many, maybe even most people, it'll appear smaller.
The one place where the dpi resolution will be relevant is if you're putting an image file online and expect people to download it and print it out by itself, i.e. not as part of a HTML page - then their printer will probably respect the dpi number and size it accordingly.
Kellie, I have two monitors (one CRT one iMac) at right angles to each other and the same image at 100% or actual size or on the web looks drastically different. So when I do design for the web I make sure to test it out for size on both monitors.
If you are measuring image size in inches, sure. Though the term there is DPI. Not PPI.
But we're talking on-screen graphic here.
PPI is simply a measurement of screen density. It has nothing to do with your image, really.
"you wanna see a grown man cry ? i can’t wait to meet you guys… hugs all around"
And beers. Hugs and beers!
"There’s the difference right there. Maybe you upstart web kids should have a bit more understanding of “old media”, or print. And take a broader view. Photoshop deals with DPI and inches because that information is EXTREMELY relevant for print media."
All the confusion here is really software's fault. They still assume the world is 1984 and we're all using Macintosh computers that truly have monitors that display 72ppi. All the technology has changed, but the software never gave up that assumption.
Patty, you are right DPI is important when printing. It's *only* applicable when printing.
"to say “dpi has no relation to file size” in the abstract sounds more unreal to me than the
mythical pixels per inch because, as bluestreak wrote, when i scan a 6 inch doc at 300dpi
i get a larger document than when i scan a 6 inch doc at 72dpi."
And thats because changing the DPI pegged against physical dimensions inscreased the actual number of pixels.
DPI doesn't DIRECTLY change the file size. Scanning software uses a DPI setting combined with a physical measurement setting (inches) to calculate the pixel dimensions for you. Changing the DPI in your scanning software is just telling the software to scan in more pixels of data. You are reinterpolating the image, just as you can when resizing in photoshop.
I agree with what you are saying, Ch...Software has one way of describing all of this which is somewhat antiquated. It's just the way we've always done it, so it's what we're used to.
It is good to understand the issues of screen vs. print image settings, though...regardless of what Photoshop wants to call it all.
This is not to be rude to anyone, but I honestly don't understand why this is so difficult to comprehend.
When I started design (largely self taught) I picked up the 300dpi number for print and the 72dpi for web but could see that there was a problem with the latter:
OK, for print I want this image to be 2.5" square, so if I need (asks printer...) 300 pixels per inch, that image that needs to be ... 300 x 2.5 = 750 pixels square.
Now for the web, I also want that image to be 2.5" square, so that means it should be .... 72 x 2.5 = 170 pixels square ... saves it at 72dpi ... checks it on own screen... OKish. Hang on, when the CEO sees it projected on the screen in the the boardroom it's going to be much bigger. And in fact it's smaller if I change the resolution on my Mac. (back in the CRT days). So what's changed - is it displaying the same number of pixels? Yes. So the dpi must have changed. So that 72 isn't a constant. (measures a few screens with ruler and checks display settings...) 75, 80, 85... tries saving at each of those DPI values, plus 300 for good measure... no effect whatsoever) right, let's forget that 72!
The operative element for web is wysiwyg - whether it's 72 dpi or 100 dpi (depends on your monitor) you are working in screen resolution.
To answer Johan, the image will LOOK better in Quark or InDy if it's 72 dpi at actual size rather than 300 dpi (same px dims) enlarged to fit. Even tho the file size is the same.
To confuse you webbies further, in print there's a distinction btw high res and low res. While the printer gets high res files, it is common to use low-res scans when making a layout to keep the Quark or InDy files from getting unmanageably large. If the scans have the same name, the printer can replace them automatically while preserving sizing, cropping, etc. So yes, to reiterate, DPI is very important.
To answer Johan, the image will LOOK better in Quark or InDy if it’s 72 dpi at actual size rather than 300 dpi (same px dims) enlarged to fit. Even tho the file size is the same.
Aha, you are talking about the appearance on the monitor, not in the resulting print. I tried what you described, and there is indeed a clear difference. However, that is only an artefact of the software's optimization algorithm. In theory, the result would be identical, and if you turn on High Quality Display (instead of Typical Display) in InDesign, you will indeed see that the pictures look the same.
ppi RELATIVE TO IMAGE SIZE affects file size.
If you are measuring image size in inches, sure. Though the term there is DPI. Not PPI.
Technically we should always talk about DPI (dots per inch) for printers, and PPI (pixels per inch) for displays. And perhaps less importantly, SPI (samples per inch) for scanners. All of these are input/output devices. Image files themselves have none of these.
The two meanings of the word "screen" -- display and halftone -- may have helped in the confusion...
"If you are measuring image size in inches, sure. Though the term there is DPI. Not PPI."
No! That's what is perpetuating the confusion. Both are measurements "per inch." If you want to only measure in pixels only that's fine. I love measuring that way and do measure that way. But don't refer to it as PPI.