mthome@bbn.com (Michael Thome) (10/21/88)
With respect to DPI levels of NeXT hardware: Given the actual DPI values of the display (*ROUGHLY* 100dpi) and the lzpr (400dpi or 300dpi) you should keep in mind that assuming the display postscript interpreter is any good at all (i.e. does some sort of anti-aliasing) and the display has a 2-bit grayscale, even the relatively low resolution level provided by the screen ought to look FAR better than it's dpi figure would suggest... maybe someone out there knows how to compute an "effective" dpi rating for a 2 bit screen? Is it, in fact, ~200 "e"dpi?
landman%hanami@Sun.COM (Howard A. Landman) (10/22/88)
In article <31144@bbn.COM> mthome@BBN.COM (Mike Thome) writes: >With respect to DPI levels of NeXT hardware: Given the actual DPI values >of the display (*ROUGHLY* 100dpi) and the lzpr (400dpi or 300dpi) you >should keep in mind that assuming the display postscript interpreter is >any good at all (i.e. does some sort of anti-aliasing) and the display >has a 2-bit grayscale, even the relatively low resolution level provided >by the screen ought to look FAR better than it's dpi figure would >suggest... maybe someone out there knows how to compute an "effective" >dpi rating for a 2 bit screen? Is it, in fact, ~200 "e"dpi? Here's a hard lower bound. Studies have shown that 2-bit grayscale gives you better "perceived" screen resolution than spending the same bits on extra 1-bit resolution. Since doing that would give you SQRT(2) more pixels in each direction, we know that the "effective" resolution of the 96-DPI NeXT screen must be AT LEAST 96 * 1.41 = 135 DPI, almost twice that of a Macintosh or Alto. If anyone familiar with those studies can provide a number saying how much better the grayscale was, then just multiply 135 by that factor (which should be slightly greater than 1.0). Howard A. Landman landman@hanami.sun.com UUCP: sun!hanami!landman
faustus@ic.Berkeley.EDU (Wayne A. Christopher) (10/22/88)
In article <74013@sun.uucp>, landman%hanami@Sun.COM (Howard A. Landman) writes: > ... Studies have shown that 2-bit grayscale > gives you better "perceived" screen resolution than spending the same > bits on extra 1-bit resolution. Are there any grayscale fonts available, or is it only useful for graphics? Wayne
casseres@Apple.COM (David Casseres) (10/26/88)
In article <74013@sun.uucp> landman@sun.UUCP (Howard A. Landman) writes: >Here's a hard lower bound. Studies have shown that 2-bit grayscale >gives you better "perceived" screen resolution than spending the same >bits on extra 1-bit resolution. Since doing that would give you >SQRT(2) more pixels in each direction, we know that the "effective" >resolution of the 96-DPI NeXT screen must be AT LEAST 96 * 1.41 = 135 DPI, >almost twice that of a Macintosh or Alto. We only know this if we know that the NeXT actually uses the 2-bit gray- scale display to do grayscale anti-aliasing. I have the impression that it does not; does anyone actually know? David Casseres
landman%hanami@Sun.COM (Howard A. Landman) (10/27/88)
>In article <74013@sun.uucp> landman@sun.UUCP (Howard A. Landman) writes: >>Here's a hard lower bound. Studies have shown that 2-bit grayscale >>gives you better "perceived" screen resolution than spending the same >>bits on extra 1-bit resolution. Since doing that would give you >>SQRT(2) more pixels in each direction, we know that the "effective" >>resolution of the 96-DPI NeXT screen must be AT LEAST 96 * 1.41 = 135 DPI, >>almost twice that of a Macintosh or Alto. In article <148@internal.Apple.COM> casseres@Apple.COM (David Casseres) writes: >We only know this if we know that the NeXT actually uses the 2-bit gray- >scale display to do grayscale anti-aliasing. I have the impression that >it does not; does anyone actually know? I too, would like the answer to this. However, assuming that NeXT *DOES* use fuzzy fonts and other antialiasing, I have a little more data. This comes from "The Media Lab: Inventing the Future at MIT" by Stewart Brand, p.170ff, which is Copyright (c) Stewart Brand 1987. Nicholas Negroponte is the director of the Media Lab. ------------------------------------------------------------------------------ The Media Lab has in its history a fine example of wrong exploiting - an excellent, simple idea, instantly applicable, available free to anyone, that has sat on the shelf in plain view of the world since 1971. Called Fuzzy Fonts, it is a cheap way to have much higher resolution print on computer and TV screens. Negroponte: "It's not subtle. When you see it, you gasp." He's right. Characters on the screen look just beautiful, like on paper, and you can read them even if they're tiny. On a TV set ordinary fonts are usually presented forty characters on a line, maximum sixty characters. Fuzzy Font characters are still easy to read at eighty characters per line, and you can go up to 100. "Having to look at stairstep characters and jaggies should be against the law," declares Negroponte. "Aliased fonts should be an OSHA violation." Aliased fonts are what you see on almost all computers. Each square pixel (picture element) on the screen is either black or white: one information bit per pixel. The problems come when you're representing a sloping line and you see a jagged edge instead, or the serif at the tail of a character is smaller than a pixel and it disappears entirely. "Anti-aliasing" smooths the jaggies by introducing a little gray in the right places. With two-bit pixels instead of one-bit you have the choice of black, white, or two shades of gray. That disappearing serif can be represented by a light gray pixel, and your eye reads it as a serif. It's cheap because doubling the resolution this way only doubles the cost, whereas doubling the resolution by doubling the number of pixels quadruples the cost - four jaggy one-bit pixels instead of one fuzzy two-bit pixel. Negroponte: "I personally have exposed tens of thousands of people to Fuzzy Fonts since Paula Mosaides - I remember her name because she was Greek - got us started with this back in 1971. The only semi-convincing argument against it I've heard is from people who claimed that the eye seeks out crisp edges, and if it encounters nothing but fuzzy edges it gets much more tired. That turned out to be wrong. Acuity is sharpness of the image; resolution is some measure of the finest level of detail that you can read. They are not the same at all. You can give up acuity and gain resolution. "Now, IBM ran an experiment where they presented the reader a page with a number of typographical errors in it, and the person was supposed to read the page and find the typographical errors. They would do it on paper, then do it on a screen with different errors but the same number of them. With a normal IBM or Macintosh screen they were something like 60 percent less efficient than on paper. Then they antialiased the fonts and put in the same kind of errors, and people came up with 98 percent of the efficiency of paper. So Fuzzy Fonts on a screen are the closest approximation to paper in terms of your ability to read them." Another test had people bringing a line on a screen to just touching a circle on the screen. If the line and the circle were anti-aliased with a little gray with two-bit pixels, the people were twice as precise. It is not just an aesthetic effect. Fuzzy Fonts may be the Media Lab's single most proven commercial idea. Why it had to wait until 1987 to become a product is a mystery. (Apple finally introduced Fuzzy Fonts on their second-generation Macintosh computers in spring, 1987, followed by IBM.) --------------------------------------------------------------------------- I'm puzzled by this last reference. Fuzzy Fonts on a Mac? Does anyone know more about this? At any rate, the above seems to indicate that two-bit anti-aliasing can get you about 2x resolution, so the NeXT is *capable* of an effective resolution of just under 200 DPI, *if* they implemented it. Howard A. Landman landman@hanami.sun.com UUCP: sun!hanami!landman
erics@eleazar.dartmouth.edu (Eric Schlegel) (10/27/88)
In article <74769@sun.uucp> landman@sun.UUCP (Howard A. Landman) writes: >I'm puzzled by this last reference. Fuzzy Fonts on a Mac? Does anyone >know more about this? Color QuickDraw and Font Manager support gray-scale (fuzzy) fonts on a mac II. I've never seen any fuzzy fonts actually created, however. ------ Eric Schlegel | DISCLAIMER: I'm just a poor college student, eric.schlegel@dartmouth.edu | which means I'm not responsible for what I | say and I can't pay you if you sue me anyway.
kent@lloyd.camex.uucp (Kent Borg) (10/28/88)
In article <74769@sun.uucp> landman@sun.UUCP (Howard A. Landman) writes: ... >mystery. (Apple finally introduced Fuzzy Fonts on their second-generation >Macintosh computers in spring, 1987, followed by IBM.) >--------------------------------------------------------------------------- >I'm puzzled by this last reference. Fuzzy Fonts on a Mac? Does anyone >know more about this? The Macintosh II supports color fonts (notice the color Apple symbol on the DA menu), including gray-scale needed for fuzzy fonts. I have never heard of anybody selling fuzzy fonts for the Mac, but it can be done. Kent Borg kent@lloyd.uucp or hscfvax!lloyd!kent
henry@utzoo.uucp (Henry Spencer) (11/01/88)
In article <74769@sun.uucp> landman@sun.UUCP (Howard A. Landman) writes: >Negroponte: "... The only semi-convincing >argument against it I've heard is from people who claimed that the eye >seeks out crisp edges, and if it encounters nothing but fuzzy edges it >gets much more tired. That turned out to be wrong..." This particular argument may be wrong, but I do believe I recall at least one fairly-authoritative person observing that there is little solid scientific evidence for the notion that fuzzy edges are just as good as higher-resolution sharp edges. (Anecdotes and testimonials are not evidence, or rather they are notoriously unreliable evidence, as witness the wilder claims for the Dvorak keyboard.) -- The dream *IS* alive... | Henry Spencer at U of Toronto Zoology but not at NASA. |uunet!attcan!utzoo!henry henry@zoo.toronto.edu
kent@lloyd.camex.uucp (Kent Borg) (11/01/88)
In article <241@lloyd.camex.uucp> kent@lloyd.UUCP (Kent Borg) writes: >The Macintosh II supports color fonts (notice the color Apple symbol >on the DA menu), including gray-scale needed for fuzzy fonts. I have I am told the color Apple is not a font. I got 2 pieces of mail telling me that (1) it is an icon that (2) is hardcoded from an MBar procedure. Not that I know what that means, but I feel better having confessed, and I feel more embarrassed that my first correction was to comp.sys.mac, but the erroneous article was here. If I take a deep breath maybe the confusion will subside... Kent Borg kent@lloyd.uucp or hscfvax!lloyd!kent