People steling my photos on Facebook

[thecoalman]

Actually DPI is a hardware characteristic of a display device, whether it is a monitor,

For monitors this goes back to Apple. They selected 72 DPI because it was closest to the monitor they were producing, it was actually 72 DPI. It doesn't apply to any monitor other than that particular monitor but they still use it. This is where lot of the confusion arises and why the term DPI is still used for this is beyond me. The OS would have to know what the physical dimensions of the monitor was to apply this setting correctly.

That's an awful example! :-) If the monitor is 22 inches wide, it would be 1920/22 = 87.

I was trying to keep it simple.

Welllll... You can take a screen shot of the image as displayed, but that is not necessarily the image that can be downloaded. I'm sure you know that, and it's just imprecise wording not a misunderstanding.

For all intents and purposes it would be. For example if I were to save a .png and upload it to the web and then take a screenshot of that .png I can save it and the checksum is same as original which doesn't happen unless it's the same exact data. This will only work with lossless images like.png and you'll need to save the image with same software and settings used to generate the origianl.

[thecoalman]

That's an awful example! :-) If the monitor is 22 inches wide, it would be 1920/22 = 87.

Ahh I see what you mean, that 18 in my sentence is typo. It should read:

To determine the DPI your monitor displays you need the physical size and the display setting. For example if you monitor is 18 inches wide and you are using a 1920X1080:

1920/18 = 87 DPI

My first calculation I used 20 inches which comes out to 96 which is the deafault setting in windows so to avoid confusion I changed it to 18 and mistakenly edited the 20 to 22 if that makes any sense.

[GadgetRick]

It is true monitors (typically) aren't 72dpi, that is my fault as that's what they used to be back in the day. It's a holdover from those times. I'm getting old...

I'm not sure we're talking about the same thing here. I have a LOT of experience in the printing industry (in a past life) and I am also a photographer. I used the first version of PageMaker, a Mac and a LaserWriter to produce a newsletter and have been doing things like this for quite some time now. So I have some experience in what I'm talking about as well.

Also a good point about displaying a pdf--or any vector-based document. It's all rasterized for the sake of the monitor or to print (unless printed on a plotter which draws with pens rather than ink). I was keeping it simple.

My point is, if you take a screenshot of an image, and you then try to print that image, it will be at whatever resolution the monitor is displaying that image unless you do something else to it afterwards. If you do a screenshot of ANY image, then try and print it larger than what it's being displayed at, you will notice a difference in the quality of the print. You aren't going to be able to print a poster-sized print without having major pixelation.

As an example, I went to my website and displayed an image at its largest size and took a shot of the image (i.e. not the whole page). File size was 1.2MB. Then I displayed the same image at the smallest size and took a shot of the image. File size was 213KB. This illustrates what I'm talking about. The only thing that changed was the amount of pixels I was capturing (on the screen). The image itself looks the same (although one is larger and one is smaller) but they are not the same file size and, thusly, they have a different amount of data included.

Now, if I were to take the smaller image and try to print it at the size it was captured at...no problem. However, if I tried to enlarge it to, say, 16x20, the resulting print wouldn't look all that good.

If there is something about capturing parts of the screen with a normal screen capture I'm missing, I'm happy to learn.

Now, back on the original topic, I totally agree that for the situation which was presented by the OP--this whole resolution/print size discussing is irrelevant since the photos will look just fine on a web page or on FB.

[GadgetRick]

Originally Posted by thecoalman

For monitors this goes back to Apple. They selected 72 DPI because it was closest to the monitor they were producing, it was actually 72 DPI. It doesn't apply to any monitor other than that particular monitor but they still use it. This is where lot of the confusion arises and why the term DPI is still used for this is beyond me. The OS would have to know what the physical dimensions of the monitor was to apply this setting correctly.

And this explains why I mentioned 72dpi...I'm getting old and have done most of this kind of work on a Mac.

[Floyd_Davidson]

Originally Posted by thecoalman

For monitors this goes back to Apple. They selected 72 DPI because it was closest to the monitor they were producing, it was actually 72 DPI. It doesn't apply to any monitor other than that particular monitor but they still use it. This is where lot of the confusion arises and why the term DPI is still used for this is beyond me. The OS would have to know what the physical dimensions of the monitor was to apply this setting correctly.

It predates Apple too. As noted, it's just an application of standard typesetting numbers for text fonts to the then new class of display hardware known as a monitor.

For all intents and purposes it would be. For example if I were to save a .png and upload it to the web and then take a screenshot of that .png I can save it and the checksum is same as original which doesn't happen unless it's the same exact data. This will only work with lossless images like.png and you'll need to save the image with same software and settings used to generate the origianl.

Nahhhhh. It requires merely that the image not be resampled for viewing. Most of course are, and we never know it. That's why a good image editor determines exactly what the display device uses, and allows some method for "100%" viewing. That way you know specifically that it is one pixel of image per one pixel of display, with no resampling.

[thecoalman]

Originally Posted by GadgetRick

My point is, if you take a screenshot of an image, and you then try to print that image, it will be at whatever resolution the monitor is displaying that image unless you do something else to it afterwards.

I probably should of clarified this before, if the image is larger than the available resolution of the monitor then a simple screen capture won't get whole image. Generally speaking images on the web are not larger than the screen resolution unless they are meant to be downloaded as a file.

The main point is by default the browser displays an image at whatever it's pixel dimensions are. If you take the image I posted above as example I'm going to capture the full 248*276 image exactly as it would be rendered by file.

If you're not capturing the entire image there is few possibilities as to why you are not getting full image. If the pixel dimensions of the image being displayed within an HTML document is larger than the available screen resolution you'll get scrollbars which of course would prevent you from getting full screenshot of image but you could still take multiple screenshots and stitch them together to get full image identical to the file.

Another possibility is the the HTML document can specify different dimensions than the native size and will do on the fly resize.

Last but not least a lot of browsers will scale an image down so it fits within the window when the image file is directly loaded into the browser.

If there is something about capturing parts of the screen with a normal screen capture I'm missing, I'm happy to learn.

A bigger monitor!

[Floyd_Davidson]

Originally Posted by GadgetRick

It is true monitors (typically) aren't 72dpi, that is my fault as that's what they used to be back in the day. It's a holdover from those times. I'm getting old...

Then I must already be old. :-)

I'm not sure we're talking about the same thing here. I have a LOT of experience in the printing industry (in a past life) and I am also a photographer. I used the first version of PageMaker, a Mac and a LaserWriter to produce a newsletter and have been doing things like this for quite some time now. So I have some experience in what I'm talking about as well.

The ancestry for those and virtually all other text formatting on computers dates back to 1964 and an undergrad at MIT named Jerome "Jerry" Salzer, who wrote a text formatting backend to an editor that he also wrote. "Typeset" was the editor, "Runoff" was the backend, and it was written for the CTSS operating system.

From that IBM evolved "FORMAT" for the IBM System/360, MIT (probably Salzer) ported the concepts to MULTICS, and from that Bell Labs deveoloped "roff", "nroff' and then "ditroff" for UNIX in the early 1970's. Certainly aware of all of those, but perhaps not really using it to derive his code, Donald Knuth also began working on device and processor/OS independent typesetting software in the middle 1970's (because getting the second edition of "The Art Of Computer Programming" typeset appropriately was a nightmare), and released TeX in 1978.

(Incidentally, I learned typesetting using nroff, and still use TeX with regularity.)

None of those programs were designed to output text to a computer monitor, but they all used the standard typesetting font parameters, such as the "point" that is 72 points per inch, as a font dimension. And when CRT monitors began to appear in the 1970's it was in conjunction with these typesetting programs. IBM for example came out with the the Displaywriter System, a dedicated word processor. Xerox PARC developed the Xerox Alto. Both had "portrait" oriented monitors, intended specifically to emulate the look a "Letter" sized paper page.

Oddly, IBM chose to design their PC with a monitor and keyboard that they thought would prevent competition with the Displaywriter, while Apple, Apollo, Sun Microsystems and others made efforts to incorporate graphic features (as well as others) from the Xerox Alto (which Xerox totally ignored for years).

All of which is to say that Apple was in no way shape or form doing anything original that relates to 72 DPI monitors! What they did (with that and many other technologies) was see the marketable aspects of the technology, and they ran with it when others didn't.

Also a good point about displaying a pdf--or any vector-based document. It's all rasterized for the sake of the monitor or to print (unless printed on a plotter which draws with pens rather than ink). I was keeping it simple.

My point is, if you take a screenshot of an image, and you then try to print that image, it will be at whatever resolution the monitor is displaying that image unless you do something else to it afterwards. If you do a screenshot of ANY image, then try and print it larger than what it's being displayed at, you will notice a difference in the quality of the print. You aren't going to be able to print a poster-sized print without having major pixelation.

Okay, now we're all getting close to the same sheet of music!

The point is that "whatever resolution the monitor is displaying" is a dimensionless pixel count that is transfered to the screenshot. It is not something related to DPI.

As an example, I went to my website and displayed an image at its largest size and took a shot of the image (i.e. not the whole page). File size was 1.2MB. Then I displayed the same image at the smallest size and took a shot of the image. File size was 213KB. This illustrates what I'm talking about. The only thing that changed was the amount of pixels I was capturing (on the screen). The image itself looks the same (although one is larger and one is smaller) but they are not the same file size and, thusly, they have a different amount of data included.

And that has no direct relationship to a meta data tag for DPI. It's simply a pixel count.

Now, if I were to take the smaller image and try to print it at the size it was captured at...no problem. However, if I tried to enlarge it to, say, 16x20, the resulting print wouldn't look all that good.

You probably can't literally do that! Instead you must resample the existing image and make an entirely new image, with more pixels, because photo printers don't print at 72 DPI. Resampling to print at 300+ DPI is what makes the print less than good looking.

If there is something about capturing parts of the screen with a normal screen capture I'm missing, I'm happy to learn.

I think you've got the mechanics down pat. The original problem was the description you gave, using DPI incorrectly. Doing that causes some of us to jump very quickly just because, whether you have a small misunderstanding or a large one, it is something that many people are very confused about and adding to their confusion is not good.

[GadgetRick]

Excellent discussion here! I love it when I can discuss things with people who know what the heck they're talking about rather than the typical Internet person who thinks they know what they're talking about.

It's also quite nice talking (typing) with people who understand the history. I'm pretty good with it but couldn't have gone all the way back to the IBM involvement (although close ). So I certainly have learned something I didn't know yesterday.

Anyway, back to your regularly-scheduled thread...