Why can’t they be circles or other easily defined shapes?
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com
If the OP was by anybody else, my reply would be, “Ask Opal!”
Because images are generally coded in rectangular coordinates, not polar coordinates (you know, since the pixels are rectangular)… If there is a circular image, it’s just a rectangle with the corners rounded off.
I assume you know what I mean by rectangular/polar coords.
They don’t have to be rectangles. Early oscilloscopes and televisions, were, in fact, circles.
I think the move to rectangles might have actually been a marketing move, but that’s only a guess.
Good question. My WAG would be that because books were rectangular, other visual reading-type things like TV and computer monitors also became rectangular.
Also, homo sapiens as a species seems to enjoy square corners. Streets, houses, tables, chairs, etc.–I’m sure some psychologist somewhere has a theory about it (and probably a book in the works, too.)
“Why, sometimes I’ve believed as many as six impossible things before breakfast!” - the White Queen
There’s nothing theoretical about it… it’s a matter of efficiency and common sense. You can bind rectangular books by the edge, but circular books?
What kind of a mess would a city be if all the blocks were circular?
This stuff is just common sense. I mean, come on! The early oscilloscopes were circular, but their cases were even rectangular.
Ok so explain how that relates to computer graphics, again?
Say I have a gif that I want to put on my page. It is a circular button. The area between the circle and the edge of the graphic is just wasted. Why not define the image as a circle and cut off the extra? Reduce file size? Also, when wrapping text around an image it wraps around the square… it would just be better in general if you could define image shapes other than rectangles. It’s annoying. 
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com
Pixels are rectangular because they are formed by interrupting a stripe formed by a tiny beam of electrons being swept across the face of the CRT. (Not talking LCD display here.) The stripe has a straight edge top and bottom even if the beam is circular. When it’s interrupted, it yields small rectangles.
If the pixels are small enough it can be difficult to see the jagged edges when a circle is formed but they are there. A good circle could be formed if you could move the beam in a circle like an ocilloscope does. Monitors aren’t designed do that though.
Opal,
Karma is right on the coordinates. Even creating a “circle” in PSP or Photoshop, it’s not a true circle.
If you don’t choose anti-alias, It’s a terrible looking circle.
As for web pages, think about all the html coding you’d have to do to get a circular or non-square or rectangle for placement.
Coding for a 100 x 200 rectangle is a lot easier.
The coding would be just like an image map. No big deal. The web already has the method to define rectangles, circles, and polygons for image maps, why not for image dimensions?
And yes I realize pixels are rectangular, but that doesn’t explain why I can’t have an image shaped, say, like a + and the antialiasing is irrelevant because we already face that problem. You would still have to have a transparent edge around your image, but you could cut off the corners and save both file size and screen space.
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com
Pictures are defined as rectangles because it’s easier, works better, and really just as flexible, if you use it right.
A large transparency in an image will be compressed into only a few bytes, meaning that the file size will be just about the same. Smaller in fact, because non-rectangular images would require a more complex header to describe the shape.
Rectangles are easier to place, having nicely defined corners. How do you place a circle? By the center point? That could be awkward in some cases. Being that the screen in a rectangle, the rectangle is a natural shape for a picture, the picture can cover the whole screen without wasted space… a circular picture covering the whole rectangular screen would have to be significantly larger. (Sure, the same compression algorithm is useful here too) Imagine though, trying to fit a rectangular segment of a circular picture perfectly over the visible screen. You couldn’t just tell it to display at (0,0) you’d have to know how far into the circle the rectangular area started and display with negative offsets (-300, -400) which causes problems. You couldn’t use a center offset because there’s no pixel at the center of the screen, so your image would always be off by a bit.
The one format which works the best in all cases is a rectangle, being that it’s easiest to have one format which can mimick others if needed than it is to have many seperate formats (when do people start clamouring for oval pictures, or random concave polygons?) the rectangle is the natural choice.
And, there’s nothing you can’t do in a rectangle. If you have problems with text flowing to the rectangular boundaries of the image, drop in the picture, put a circle around the picture, flow the text within the confines of the circle, and remove the circle guide. Tada, text flowed into any shape.
If your program doesn’t do this, it’s a fault of the program, not the image shape.
I’m pretty sure Corel Draw (later versions) do this. I’ve had bad luck with Pagemaker, but that could just be its awkward interface.
You might want to look into .PNG files, which are basically .GIFs which compress better, are true color, and allow transparencies. Most web browsers support them now, though some (older Netscapes) break a bit on the transparencies. Unlike .JPGs, .PNGs are lossless, you’ll always get the exact same image, so they’re handy to use as an intermediate format.
I don’t have any definative answers to your questions, but I can tell you that graphics are heavily reliant on matrix multiplication for about everything. All the different things you can do to a picture like lighten it, sharpen it, change colors, resize it, distort it, all do the actual work via matrix multiplication and matrixes are XxY in size or rectangular.
“The truth does not make a good story; that’s why we have art.”
There are circular & oval computer images.
There are even circular desktop images & icons too.
But the edging of them is what sucks, you know that staircase effect.
Images are bitmaps, made of pixels. Normally you want a constant pixel size across the image, just because there’s not much point in an image format that has higher resolution in the center, for example. If you have constant pixel size, the logical thing is to make a grid pattern with them. And the most logical overall shape of a grid pattern is a rectangle. If you try to make arbitrary shapes with pixels arrayed in a grid pattern, you will end up with jagged edges.
Also, just defining the shape may take up a lot of memory, whereas the size and shape of a rectangle can be defined by just two integers.
You’ve completely missed the point.
I’m not talking about creating an image that has text in it. I’m talking about, say putting an image on a webpage, and aligning it “left” or “right” which means that text will wrap around the image, rather than the image sitting on the line like a word. Here is an example: http://fathom.org/opalcat/diary/031900.html
Now, if that image were a circle, it would be nice to allow the text to come into the edges, rather than stopping at the square point, you know?
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com
I think I’m missing some of your point, but in the case of the text wrapping around images in a web page, it’s a limitation of tha application. Browsers don’t typically support complex word wrappages because they are slower and take more information to represent. On the other hand, applications like Canvas or Photoshop can wrap text around the outside, wrap it to the inside, or bind it to a curve in pratically any complex shape you desire. That’s not a limitation of the computer, it’s a limitation of the software.
I thought, at first, you were asking why most computer graphics elements, like windows, icons, etc. tend toward the rectangular. If this is your question (or somehow related to your question) I offer this little tid-bit. The specification of any arbitrary rectangle can be specified with just two points. This defines the position and size. A circle could be represented with slightly less information, but it is just a special case. Since circles and ovals can both be represented with a rectangle (bounding box), this tends towards a common solution to both rectangles and ovals. Any other shape must be represented by a complex series of line segments and bezier curves.
Note: There is nothing specifically preventing non rectangular windows from being used. In fact, way back in 1984, one of the early example programs for the Macintosh demonstrated how windows and controls could be in any shape. I’ve seen similar programs for the PC. The reason that these kinds of windows are not popular is because they typically result in non optimal display of information and many users find them aesthetically disturbing in their rectangular viewport (i.e. monitor).
As for images, the file formats tend to favor rectangular representations and it’s only natural to map the images to rectangles on the screen since it’s a simple algorithm. However, I’ve seen non rectangular mappings, so it’s not impossible.
Probably more than your were looking for… or perhaps less… I don’t know… Hope it helps, anyway…
No. My question is “when I save a .jpg or a .gif etc why does it have to be a rectangle?”
I realize that current browsers dont’ support non-rectangular images, because it doesn’t exist. That is so completely irrelevant to the point I stagger at trying to express it. Basically, I want to be able to put non-rectangular objects in a webpage. That is what it boils down to. Why can’t I have an “L” shaped graphic, and put other stuff inside (other graphics, text, etc)? No, I have to use a tall, skinny rectangle and a wide, flat rectangle and fake it. And then I have to use tables and it gets all complicated.The L is acutally a pretty dang simple one to work around, but say I wanted to put a circular image in the very middle of my page and have the text surround it? As the world is now, I could have a circle inside an invisible square, and the text would leave open that square hole and look lame.
Now, I’m not saying it is the most natural thing, or the easiest, or the most convenient or the speediest, etc. I just want to know why it can’t be done. Regardless of how good an idea it would be to put a round image in my webpage, why is it not an option?
Dammit! I want my round .jpgs!
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com
No, browsers don’t support complex text wrapping because there is no such thing. The stuff you mean in Photoshop and such is a completely different concept. I’m just talking about putting an obstacle in a chunk of text and having the text use the available space. As it is, it can only use the available space outside a rectangle. I just happen to think it could be better.
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com
Let’s kill this post once and for all.
gifs and jpgs are rectangular because their encoding algorithm uses retangular gridding. You can’t just use part of the rectangular grid, you have to use the whole thing.
End of story!
I don’t understand why you feel such a need to “kill this post”. And yes, the current way things are done require rectangles. That is sort of my point. WHY can’t they come up with one that doesn’t? Or have something that says “unused parts of the grid don’t exist” etc?
I am aware that current image formats require rectangles. That is sort of the whole reason for my question. So far I haven’t seen any reason given why there couldn’t be one that was round.
–
I have over 2000 posts, dammit! Show some respect.
O p a l C a t
www.opalcat.com