Timeline for Dot vs. Pixel - rudimentary explanation
Current License: CC BY-SA 4.0
23 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 18, 2021 at 10:34 | comment | added | ComicSansMS | @Wolff The more crucial point of that paper is that the point samples are always perceived through a reconstruction filter and squares are a really bad model for how this works in practice. So if for instance, you are worried about how a sharp border between two colors will look like in print, the squares model will not be a good predictor. I guess most artists have a good intuition for the different reconstruction filters for print, screen display, etc. The paper attempts to formalize that intuition in a more mathematical way. | |
| Nov 16, 2021 at 17:40 | comment | added | Wolff | @ComicSansMS, thanks for the link! That's interesting to me personally (not least because I dabble a bit with WebGL at the moment and it helps to think like this). But I would argue that for most graphic designers it makes more sense to regard pixels as squares. They look like squares and behave like squares, both in our graphical applications, on screen and on print. Graphic designers are in the end mostly concerned about how things are perceived. Background knowledge is always good of course, but I can't really see how this "points in a grid" model could affect a design decision. | |
| Nov 16, 2021 at 13:19 | comment | added | ComicSansMS | Minor nitpick: Pixels and dots are not really little squares, even though graphics software sometimes confusingly likes to display them as such. The truth is a little more complicated. | |
| Nov 16, 2021 at 10:02 | comment | added | Peter Cordes | Ah right, ink bleed on paper, forgot about that part of the physical world :P I think I'd read about that before at some point. Thanks. Agreed that's more explanation than belongs in an answer, but perhaps something like "(which are larger than pixels so they can be round and have controlled spacing, and for physical paper/ink reasons)" could be snuck in somewhere if you want, or just here in comments is fine. | |
| Nov 16, 2021 at 9:32 | comment | added | Wolff | This frequency is different on different papers. The more uncoated and coarse the paper is, the bigger the pattern. The ideal is to have completely round halftone circles like before the digital age where the pattern was analogue. A high DPI is needed partly to be able to make the halftone circles seem round to the human eye and partly to be able to make the size of the circles vary in small increments. Hope this clears things up a bit. 😀 | |
| Nov 16, 2021 at 9:32 | comment | added | Wolff | @PeterCordes, I think adding this information to the answer would be too much. It's a big subject. But in short: You are on the right track. A checkerboard pattern of tiny 2400 PPI dots would not be durable and you can't be sure that every single dot would be transferred to the paper. Besides that there is the issue called dot gain. When the ink hits the paper it bleeds out a bit. So a 50% checkerboard pattern would just become all black. The halftone pattern needs to have a certain frequency (measured in LPI - lines per inch) so this doesn't happen. | |
| Nov 16, 2021 at 8:13 | comment | added | Peter Cordes | If there's any simple explanation for that (or a link that explain it), it would be cool to throw that into your answer when you introduce the concept of halftone dots. If there's some standard connection between their size and the PPI of the printing machine, that would somewhat relate those two resolutions perhaps? Or that's probably going to be more confusing to bring up for people that aren't already thinking about half-tone dot pitch as some kind of resolution. | |
| Nov 16, 2021 at 8:10 | comment | added | Peter Cordes | This may be getting off topic, but the obvious (to this casual outsider) followup question is why the "halftone dots" are so big (so many pixels), much higher than the 2400PPI resolution would appear to allow. Is making them at such a coarser scale necessary for physical durability of the print plate over a run, or is it more about being able to finely vary the spacing to get different visual intensities without making the pattern irregular? e.g. exactly half could be done with a checkerboard pattern of 2400PPI dots, but other regular patterns that tile the plane are different avg. density. | |
| Nov 16, 2021 at 0:43 | vote | accept | Wes | ||
| Nov 15, 2021 at 21:29 | comment | added | Wolff | @U.Windl, yeah it's confusing that "dots" are used to create ... well "halftone dots". Would be easier if we could call the latter "circles" or something to able to differentiate. In offset print a dot isn't really a "drop" of ink. It's a square area like you see in the preview on screen. In my example a 1/2400 inch x 1/2400 inch square. I think I lost the thread here. Do you still have a question? 😅 | |
| Nov 15, 2021 at 21:13 | comment | added | U. Windl | @Wolff About the "typo": Actually it was intentional: The text said (the equivalent of) a dot is the smallest drop of ink on the paper, but I contrasted it with a "color dot" that consists of multiple single-color-dots (e.g.: CMYK). I got the impression that print resolution is so "low" because of that. | |
| Nov 15, 2021 at 16:34 | comment | added | Wolff | Let us continue this discussion in chat. | |
| Nov 15, 2021 at 16:17 | comment | added | Vikas | @Wolff yes I've seen those imperfect colors on cheap printed stickers, perhaps on matchboxes sometimes. So Pantone prints real color? Suppose if there's Pantone brown/pink/skyblue color, it wouldn't mix some colors but actually uses real colors and prints things? | |
| Nov 15, 2021 at 16:09 | comment | added | Wolff | @Vikas, I simply just think that our language is lacking some precision here. The word "pixel" is both used for the tiny physical square on your monitor and for the abstract idea of a square "picture element". | |
| Nov 15, 2021 at 16:06 | comment | added | Wolff | Yes @Vikas, ordinary printers can only print solid colors. (Some special printers might exist which I don't know about.). If you look at printed material with a magnifying glass, you will most likely see a pattern of some sort. This is one of the reasons for sometimes choosing to print with spot colors (like Pantone) because then you can get solid, smooth colors without pattern. But only in that specific color. See my comparison between the brown colors here. | |
| Nov 15, 2021 at 16:03 | comment | added | Vikas | Also, since pixel is not a physical thing, what do we call one dot in a screen (laptop)? Casually I would always say screen consists of pixel and pixels become dead in old screen. | |
| Nov 15, 2021 at 16:02 | comment | added | Vikas | So a printer really doesn't print any other real color apart from C, M, Y and K? All other colors are mere illusion? Earlier I had believed that printer mixes C, M, Y and K in proper ratio to create another color. But what you've described makes my belief incorrect. | |
| Nov 15, 2021 at 15:58 | comment | added | Wolff | @U.Windl, about the relationship between dots and pixels: I've tried to show how a print PDF can have images with multiple resolutions and that pixels can even be rotated. So the print PDF can have many different sizes of pixels. All these pixel grids don't align to each other or to the grid of dots or even to the grid of the halftone. It's hard to simplify it because it's not that simple. The original pixels are simulated by halftone screening which is made of dots. | |
| Nov 15, 2021 at 15:53 | comment | added | Wolff | @U.Windl, there might be a typo in your comment, but dots can't be said to consist of dots. They are the smallest unit. They do not really "exist" on screen. They are displayed only for preview purposes as pixels in pure black or the color of the ink used for print, but they are to be seen as "printing instructions" for the device. They do not as such "have" a color. In offset printing they end up as a tiny square on the printing plate that can hold ink. The plate itself does not "know" which ink will be applied. | |
| Nov 15, 2021 at 15:19 | history | edited | Wolff | CC BY-SA 4.0 |
added 6 characters in body
|
| Nov 15, 2021 at 7:13 | comment | added | U. Windl | What's still unclear is the color relation: A pixel is definitely colored, while a dot on paper or even on screen may consist of multiple dots to represent one colored pixel. So IMHO there is no 1:1 relationship between on pixel and on dot; it's more like a 1:n (one pixel represented by multiple dots). | |
| Nov 15, 2021 at 4:00 | comment | added | slebetman | This is very close to the perfect answer. This should be the accepted answer | |
| Nov 14, 2021 at 16:30 | history | answered | Wolff | CC BY-SA 4.0 |