The Flea's Knees - Handmade Subpixel Type Family with 3px x-height

miha's picture

It has x-height of 3 pixels, it is italic, inspired by old masters and most important – made by hand. You must not look too closely, because colors get visible. It's work in progress. Here:


I am also writing a program which will set text automatically. These horizontal lines in the picture is actually information about glyph unicode value and positioning.

It can be used in favicons and for … well, one reason is enough :-)

Scroll down for additional weights of roman, bold & bold italic.

The title was "First Handmade Subpixel Type Family, Ever*".
[* Moderator's note: Miha's enthusiastic use of the claim "first ever" has been seriously challenged by StoneCypher. Read on for the drama and sub-pixel intrigue, if you dare!]

PS: Ken Perlin's work is interesting too.

Comments

StoneCypher's picture

Bloodtype: Surprisingly, I'm actually just one person. The pluralization is unneeded.

Bloodtype's picture

If I may quote your website,

http://fullof.bs/files/2009/08/DesignersWhoThinkTheyMatter.png

"Designers usually use Mac's, they're so stupid. But programmers that can also design (like me) know better. Who would hire a designer when you can hire a programmer slash designer that can do it all!"

This suggests that you hold designers generally in low esteem, as well as design itself. Miha's exuberance in finishing a wonderful typeface comes from an artistic temperament that you most likely don't have. You may think more logically in zeroes and ones, rather than laterally and creatively. This is very useful for computer programming but not so much for design. I don't know if it has been suggested to you that you might have borderline Asperger's syndrome, but this is something worth asking your doctor about.

I think your talents lie more in the programming area than in design. Most people looking at your website would conclude that it could do with a designer's touch.

Your's in good faith.

Bloodtype's picture

But can't you see the joy in the title; this is an artist expressing their pleasure in a job well. Miha isn't submitting a patent, just sharing her work with (hopefully) like-minded fellows. All the people you've mentioned have had their work out there. Miha hasn't ripped them off. Believe me, I know what it's like when an idea I've had gets nicked before I've had a chance to put it out there. Miha hasn't deprived anyone else of income or kudos, and that is the important thing. The wording of the title of his/her offering is relatively unimportant.

StoneCypher's picture

Bloodtype:

"This suggests that you hold designers generally in low esteem"

It's an understandable mistake to make. I didn't make that comic, nor did I write that text, actually. It comes from a website called http://businessguysonbusinesstrips.com/ . My attention to that comic was the claim that developers were stupid, by someone with limited grammar skills.

I have a great respect for designers, because they fulfill one of the few roles in the software development process I cannot fulfill myself. The reason my relationships with my designers are so good is that I take very good care of my staff and respect them for their work.

"Miha’s exuberance in finishing a wonderful typeface comes from an artistic temperament that you most likely don’t have."

You're welcome to your opinion. I would point out that you don't actually know anything about me other than how I react to false claims by a typographer.

"You may think more logically in zeroes and ones, rather than laterally and creatively."

I may. Or I may not. Speculation doesn't do you much good here, and I'm actually closer to being an artist than I am to a developer, though neither label applies particularly well.

"I don’t know if it has been suggested to you that you might have borderline Asperger’s syndrome"

Please be serious. An Asperger's victim doesn't communicate well with others, and would never have written any of this in the first place, no matter how they felt. I don't think you should probably attempt to make future psychological assessments of people you don't know. It's not very compelling. But that's just my opinion.

Incidentally, I very much hope there are no Asperger's victims reading this. If there are, you just made them all very sad, by pretending that a guy who's angry must have their disease, and that their disease causes angry behavior. It does not, and it's inappropriate to use psychological disorders to shame people in public, because the people who actually have those disorders can become very hurt by that.

"I think your talents lie more in the programming area than in design."

You know as little about my talents as you do about my mental state, good sir. I am disinterested in future speculation on the matter, though I've no doubt that won't stop you from trying. Typically after Asperger's, the armchair psychiatrist goes for obsessive compulsive disorder, depression, schizophrenia or schizoaffective disorder, psychosis, paranoia (as was used earlier by someone else who is confused about what they understand) or psychopathy.

It's a hilariously varied, unrelated and inapplicable bunch, but I'd like you to have a broad palette to choose from when feigning the ability to diagnose people as a medical doctor.

If you're looking for something more exotic, Porphyria is nice this time of year.

"Most people looking at your website would conclude that it could do with a designer’s touch."

That would be because the CSS is currently broken by a malfunctioning SEO plugin that I'm too lazy to fix. A swing and a miss; the design was actually made by one of the most successful video game artists in the Nintendo gaming community, and I get compliments on the skin at the rate of several a day from strangers.

I tried looking for your web page to offer return commentary, but Google has no idea who you are, other than a largely obscured linkedin page.

My opinion, given the success rate of your guesses, is that you should probably stop.

StoneCypher's picture

"But can’t you see the joy in the title; this is an artist expressing their pleasure in a job well."

Joy does not excuse falsehood. If I'm happy about the new video game I just wrote, should I call it the first video game ever?

"Miha hasn’t ripped them off."

We disagree. My opinion is that claiming the first ever is an enormous ripoff to the dozens of people who beat him to it by decades.

I would make an example by claiming rights to some of your work, but I can't find any.

"Believe me, I know what it’s like when an idea I’ve had gets nicked before I’ve had a chance to put it out there."

That's a very different situation. Being beaten to the punch is a bear, sure, but that isn't what's happened here at all. Miha would need a time machine and 30 years of time gasoline for this viewpoint to be appropriate.

"Miha hasn’t deprived anyone else of income or kudos"

The former, no. The latter, we disagree: Google now attributes this significant technical success of yester-year to Miha. That is depriving the appropriate people of kudos.

"The wording of the title of his/her offering is relatively unimportant."

The fact that someone looking into the history of subpixel fonts now thinks that Miha came up with this in 2009 is contrary to this viewpoint.

However, this is a matter of opinion, so I cannot say that you are wrong, only that we disagree strenuously.

I appreciate that you've become civil.

StoneCypher's picture

And for future reference, programmers don't think in ones and zeros any more than artists think in terms of the crystallographic refraction index of their paint.

John Haugeland is http://fullof.bs/

Bloodtype's picture

John, if someone googles "first subpixel font" or whatever, they'll only find Miha's claim to it (plus your numerous counter-claims), not the definitive answer. You'd be a fool to think you could find proof of anything in a quick web search. As far as Asperger's Syndrome goes, I do have some knowledge of it from my brother. I am certainly no psychologist; I suffer from depression and I hate it when people say they're depressed when what they really mean is temporarily sad. Suffering from mental illness has given me some perspective and a keen interest in the whole spectrum, I certainly wasn't trying to shame you.

"I would make an example by claiming rights to some of your work, but I can't find any."

Whatever anyone has said about you, I think you must agree that the above statement is both aggressive and deeply unpleasant. Just because I haven't as yet made my work available on the Worldwide Interweb, it doesn't mean I haven't done any, and your rank arrogance in suggesting that you would, "make an example by claiming rights" to it is hopefully beneath you.

"If I'm happy about the new video game I just wrote, should I call it the first video game ever?"

Hell yeah! In that moment in time. It may be cheeky, it may be flawed. These words weren't carved in stone, they were genuine feelings expressed by someone whom it would be difficult to label a rip-off merchant. And if a google search proves him wrong then so be it, or if he contends it doesn't, then argue away forever, but maybe focus your energy on other projects, as I hope to do mine.

Anyway, that was my best attempt at proper writing.
Nice chatting with you,
Owen

StoneCypher's picture

"You’d be a fool to think you could find proof of anything in a quick web search."

Actually, I was able to find quite a bit of proof with quick google searches some time, and I documented such here. Perhaps you should catch up on the thread before making further sweeping generalizations.

Patents exist for the sole reason of establishing order of primacy proof.

It is amusing to see you say "you'd be a fool to think you could do that thing you just did a couple of days ago", though.

"As far as Asperger’s Syndrome goes, I do have some knowledge of it from my brother."

I am not interested, and I confess to having a very difficult time believing someone who has this terrible affliction in their family would make such a crass and grossly inappropriate claim in public. Do not follow this line of reasoning any further. Your behavior is highly inappropriate.

"I certainly wasn't trying to shame you."

Riiiiiiight. You just speculated on my mental health in public based on a dislike for the things I said, and told me to go to a doctor and ask whether I had mental health problems because I chose to stand up and say "this dishonesty isn't right."

You undermine yourself by saying such transparent things.

"“I would make an example by claiming rights to some of your work, but I can’t find any.”

Whatever anyone has said about you, I think you must agree that the above statement is both aggressive and deeply unpleasant."

I most certainly do not. It's a simple thing to say: I'd try to show you how it feels for someone to make a claim to your work inappropriately, but I spent about 30 seconds looking for some and gave up because it was boring to search for things I couldn't find under a relatively common name. There's nothing aggressive or unpleasant about it. I think you should read the things you're saying to me before making so much as a single more comment about my being unpleasant; your behavior is far worse than that about which you complain (which seems to be de rigeur here.)

Certainly someone who speculates about people's mental health in public should feel it relatively difficult to claim that someone else who said "I stopped googling" is being unpleasant.

"Just because I haven’t as yet made my work available on the Worldwide Interweb, it doesn’t mean I haven’t done any"

(blinks) Who said you hadn't done any? I just said I couldn't find some to use as an example. It's not really fair for you to get angry at me for things I didn't say.

"and your rank arrogance"

Ah yes, I'm the aggressive one. For saying I couldn't find something on Google which apparently isn't on the web at all. Which empowers you to call me "rank arrogan[t]", because apparently either arrogance has an aroma or a position in a heirarchy.

"in suggesting that you would, “make an example by claiming rights” to it is hopefully beneath you."

I can't even figure out what you're trying to say. It should be beneath me to describe as a thought experiment that hypothetically I made claim to your work while saying I was pretending it to show you why it was hurtful?

... I just don't see what you're angry about here.

"“If I’m happy about the new video game I just wrote, should I call it the first video game ever?”

Hell yeah!"

At this point it becomes clear that I should not be listening to you. You are advising me to be dishonest on grounds of pride.

No thanks.

In that moment in time. It may be cheeky, it may be flawed.

No. It's called a "blatant lie". If you are unfamiliar with the difference between cheeky and flawed, and lying openly about your accomplishments to garner notoriety, then I'm just not sure what to tell you.

It is never appropriate to claim a world first when you know for a fact it isn't true. Lying has nothing to do with being proud or being cheeky. It has to do with having a serious character failure and an infamiliarity with basic adult behavior.

The reason the word "liar" hurts so much is that you're supposed to be horrified at the idea of being one. It's not an invitation to mince it into carefully re-framed spin that makes it look like repeated false claims to world firsts are just someone taking pride in their work.

They aren't, and it's a sentiment which is toxic to civilization. It's so toxic in fact that it's illegal in most of the modern world for a journalist to knowingly lie.

Any further advice you give on what makes it okay to make statements which one knows to be false will make it fundamentally impossible to listen to you, I'm afraid. If your goal is to talk, great, keep it up. If your goal is to discuss and be listened to, find something to say which isn't pretending it's okay to lie in public because it makes you feel good. It isn't, and merely reciting the idea makes most listener's stomachs queasy.

These words weren’t carved in stone, they were genuine feelings expressed by someone whom it would be difficult to label a rip-off merchant.

Making claims to world firsts isn't a "genuine feeling", it's a falsehood. This isn't a therapy session. It's a known to be false claim to accomplishment which is being published and re-published, and no matter how many times you try to find some way to slice and dice that until it looks like something different, that remains the fundamental nature of this issue.

whom it would be difficult to label a rip-off merchant.

It'd also be difficult to label them a basketball player. Luckily I didn't use either phrase. Please refrain from further straw men.

"but maybe focus your energy on other projects"

This isn't a project. This is a promise to an aggrieved friend.

Bendy's picture

Can we bring this thread back on topic?

aluminum's picture

My cat's breath smells like cat food.

Zara Evens's picture

Another sub-pixel typeface could have been created in the time it took John to craft all his posts. John, why don't you post some of your work? Let's talk about that instead of sitting here wasting time on this battle of the alpha males.

If this childish bickering doesn't come to a halt, I will close the thread to additional comments.

typerror's picture

I do not have a dog in this hunt, but have found it enlightening , disheartening and in a perverted sense, fun.

I would rather see this resolved than close it because of "childish bickering!"

Sure John can be a &*)%#$@! (so can I) but WHAT IS the definitive answer? He has shown himself to be quite knowledgeable, albeit a bit gruff.

As to the alpha male crap... just say "Can't we all just get along," instead of we do not believe in free speech! I hate PC crap.

And, no, I do not like it either, but it seems things are being left unresolved!

Michael

p.s. Zara, if I read between the lines of your post you are being confrontational and demeaning!

p.s.s. the title is a pretty ballsy claim.

Zara Evens's picture

OK, well it seems the community has spoken and they would like for this kind of behavior to continue, so I will stay out of entirely.

Please contact your moderators when issues and conflicts arise - not me.
http://typophile.com/moderators

Thomas Phinney's picture

Zara et aliae: John/StoneCypher *did* post links to his work, as well as the work of others. After a ton of reading over the past couple of evenings (because I think sub-pixel anti-aliasing is interesting), I am of the opinion the links pretty thoroughly back up his claims, even though some relate to automated processes. Whether he's been insensitive, confrontational, or a jerk is an orthogonal issue to whether he's right. I'm utterly convinced that with regards to the facts that he is indeed right, and has been pretty clear. His demeanor does not constitute justifiable grounds for ignoring facts.

I still think Miha's fonts are cool. But they're not the first ever of their kind, is all. Possible the first done so small?

Cheers,

T

typerror's picture

Thank you Thomas!

I just sent a prayer to Zara asking her to reach out to a "knowing source" and you of course have appeared as the knight in shining armour. Well, at least at our age it is a bit dinged and dull, but you still stepped up to the plate.

I did not care who was right, I just wanted some knowledgeable person to step forward.

As I told Zara... I have learned more about something I knew nothing about than in any other post on this site. Cool. But I cannot see doing a 4 point Chancery:>)

Michael

muji's picture

Thomas

I'm glad you trawled through the long list as I gave up when I didn't immediately see a definitive 'answer'.

I understand that Miha has conceded that the claim 'first handmade subpixel type' would be inaccurate but that he believes the prior existence of a 'type family' has not been proven.

I have looked at the example John gave me http://www.mrob.com/pub/xapple2/index.html which contained two images showing a pixel font with upper and lower case, in 'thick' and 'thin'.

If a typeface simply with more than one weight can by definition be called a family, then it's a family. In my eyes, however, the normal expectation of a 'family' for a regular text font is to see at least italic versions of roman and bold as well. As I personally would have difficulty in describing just any two weights a family I looked up a dozen or so definitions and the combination roman-italic-bold-bold italic was generally given as the basic example of a font family which matches the family Miha has designed.

So, among all John's links was there an example of 'handmade subpixel type' we can all see that shows a typeface with a family of weights rather than just 'thick' and 'thin'? This would solve the dispute very simply. Without one I can understand why Miha is sticking to his 'first' as his type is based on a traditional type family.

If Miha's claim is proven to be inaccurate then perhaps the 3 pixel x-height is new? I'm sure John will let us know if he thinks it's not.

aluminum's picture

I do believe StoneCypher works for Fox News. He's done an excellent job at derailing a conversation and directing it towards a heated and strongly worded debate on an issue that really isn't germane to the original subject matter. Well done!

miha's picture

Thomas, do you really think John Haugeland is right? I've asked for a handmade subpixel type family already in my first reply to him, but really couldn't find any. Is there a specific link that convinced you? The link to time machine looked promising on the first sight, but there is only greyscale antialiasing.

The criticism for my title could be that it is just nitpicking – heck, the difference between first whole family and first typeface with one style is small, it's just a small step forward with evolutionary nature. Thinking more if this title really matters I came to conclusion that it doesn't, typeface is already special, it would be more important to make it better than wasting my energy here. Besides, there is a possibility that we can't know the truth because such typeface isn't/wasn't documented on internet. I didn't change the title because it would look like that I believe John Haugeland is right.

Now, I have to say I have so enough reading this thread. There was very little discussion, impoliteness by various users (yeah, haven't ever seen that in such extend here!), moderators commenting only after a few days, and one moderator stoped moderating. Thomas – I respect your opinion and I will be glad to change the title as you think it is the best, even if I will disagree with you.

StoneCypher's picture

"Another sub-pixel typeface could have been created in the time it took John to craft all his posts. John, why don’t you post some of your work?"

I did. Maybe you should actually take the time to read the thread soon.

"Let’s talk about that instead of sitting here wasting time on this battle of the alpha males."

Yeah, an alpha male is actually someone who gets into physical fights in order to secure breeding rights, not someone who stands up against an ethical wrong.

Typerror: I appreciate the show of support.

Thomas Phinney: Depends on how you look at it, if you'll pardon the pun. I have seen smaller fonts, but they aren't complete, and they aren't families. I've seen fonts that are taller but thinner (and made one such.) I've seen fonts which are shorter but wider. I have not seen subpixel fonts with this small a text footprint, which is the metric I usually use personally. I have actually seen (and personally written) binary fonts smaller than this - I made a complete IBM high ascii codepage in 3x4 pixel - but it's whole pixel, not subpixel, and also not antialiased (it was made to make a game called Rogue on the old Sega VMU, which had a 48x32 monochrome discale screen.)

Aluminum: Cute. You're welcome for that detailed answer I gave you to your question, by the way.

Miha: You're just looking for excuses at this point. That font does have bold and italic variations. You'd know that if you had just looked once at the rest of the examples; that's from the original list, and the original link shows the rest of that family, which not only has bold and italic, but also outline, emboss and ridge. So do half a dozen other links from that list. Stop claiming knowledge on the topic when you won't even read the links given to you.

I’ve asked for a handmade subpixel type family already in my first reply to him, but really couldn’t find any.

That's because you won't read the links given to you. I even pointed one out explicitly from the list before you went onto this new tangent as being families at various sizes, weights, slantings and so forth. It's the one that comes from archive.org. There are half a dozen families on that page alone. I've also reminded you about this twice; it's clear that the reason "you can't find any" is because you refuse to read the people who are showing you where they are.

but there is only greyscale antialiasing.

Right, which is why the second paragraph reads as this exactly opposite statement:

Airwindows formally thanks Microsoft for devising ClearType, and thereby figuring ways to add strange color casts to type on LCD screens, while simultaneously reinventing a technique they may well have first heard of from me.

Maybe you should look again? Those are not grayscale anti-aliased (well, all subpixel fonts are actually grayscale anti-alised; it's images imitating subpixel fonts which have colors. But I already explained that to you, and it appears you weren't listening.)

Willful ignorance is not a good excuse for false claims.

I didn’t change the title because it would look like that I believe John Haugeland is right.

Yep. That's exactly why I tried to be harsh with you: it's the only way to reach people who are digging in their heels, refusing to believe the obvious truth, taking only cursory examples out of single points full of lists of evidence, missing the text that comes up by the second paragraph even when the link list explicitly points it out beforehand in prophylaxis, and _still_ manages to think it isn't true.

Unfortunately, Aluminium has already used the Fox News joke.

heck, the difference between first whole family and first typeface with one style is small, it’s just a small step forward with evolutionary nature.

Well, it would be, if the first time this was invented this hadn't been deployed as a family. As you've already been notified repeatedly, of course, the earliest example I've been able to find is the Apple ][, wherein bold, italic and bold-italic were part of the very first family, so there's no evolution involved, that's just how it was done from day one.

Thinking more if this title really matters I came to conclusion that it doesn’t, typeface is already special

Good, then you should have no problem removing the false claim, since it doesn't matter to you, and resting solely on the special-ness of the typeface.

Muji: "So, among all John’s links was there an example of ’handmade subpixel type’ we can all see that shows a typeface with a family of weights rather than just ’thick’ and ’thin’?"

Yes, the very first example has normal, bold, italic, bold-italic, underline, bold-underline, and cursive, bold-cursive, underline-cursive and bold-underline-cursive (which may or may not be a separate font depending on how you look at it, as it only replaces the letter faces and the digits.)

Also, several of the gameboy advance and DS examples I gave have entire families, and as I pointed out, I converted about a dozen of Microsoft's fonts at various weights, slants and sizes (mind you, I had to go back and hand-pixel them, because ClearType's hinting falls apart below eight pixel em-height.)

Furthermore, the archive.org site has about three times as much work as I did, and beat me to the punch by some unknown amount of time, but at least seven years and probably substantially more than that.

perhaps the 3 pixel x-height is new?

Nope. Indeed I've seen subpixel fonts with x-heights of two. At this size, I often find it more productive to discuss em-heights; the small glyph height at this size often has to vary (for example, most hand-drawn small fonts of em-height 4 have an x-height of 2, but allow lower case S, Z and so on to go up to 3.) Also, it's occasionally important to discuss raw glyph height, because at this size sometimes a letter with a descender is up-shifted positionally to make due.

I looked up a dozen or so definitions and the combination roman-italic-bold-bold italic was generally given as the basic example of a font family

Well, I wasn't really going to bring this up, because a lot of people are going "waaaah argument", but I guess I kind of have to now.

A font family has nothing to do with its faces; a font family is a group of fonts meant to work well together. An example of a font family in the modern era with which most are generally familiar is candara, consolas, corbel, calibri, constantia and the ugly one. The misunderstanding that a family means its faces comes from the older days of machine typesetting where a font was just a font, and had one glyph per index and that was it; at that time, bold/italic/bold italic would just be shipped as separate fonts, usually with b/i/bi on the end of their names (you can still see that on the system bitmap fonts in Vista boxes), and at that time it seemed natural, as they were a set of fonts meant to work well together, to refer to them as a family.

What Miha made is actually a single font with several faces, not a family at all. Citation should be easy for people here to find, since they're so many type enthusiasts, but since a person doesn't make claims without evidence, you can find such discussions in Meggs (aka "Typographic Design: Form and Communication", doesn't show up until the yellow 2006 edition, so if you have the cyan 2002 you won't find it) as well as in one of the iF yearbooks, where I initially learned this (not sure, think it's 2005.)

But I wasn't going to argue that, because what Miha meant was clear, so I figured I'd just work with what Miha thought the terminology meant. 'Cause you know, that isn't a first either.

John Haugeland is http://fullof.bs/

Thomas Phinney's picture

"A font family has nothing to do with its faces; a font family is a group of fonts meant to work well together. An example of a font family in the modern era with which most are generally familiar is candara, consolas, corbel, calibri, constantia and the ugly one. The misunderstanding that a family means its faces.... What Miha made is actually a single font with several faces, not a family at all."

You're welcome to use the terminology however you like, but that is not how most people, experts or laypeople, use the terms. I even did a survey on this topic a few months back.

Cheers,

T

William Leverette's picture

I can furnish definitive evidence that a type family must consist of at least four members ;)

StoneCypher's picture

Thomas Phinney: That usually means one of four things. I was wondering whether you could distinguish between them, or offer an alternative.

1) I'm misremembering.
2) I'm citing a source correctly, but that source is full of crap (eg wikipedia).
3) I'm citing an obsolete or archaic source (books that have since been superceded), or a source affected by language differences (eg british english, translation)
4) It's a matter of divided opinion, of which I am unaware.

Are any of those four cases what's happening here?

I should also point out for the record that Adobe uses the nomenclature I gave. There was an old technology on the NeXT called "multiple master fonts", which were awesome: they were hand-hinted explicitly scalable fonts, so the artists could set algorithms for how the fonts thickened, and so the end user could just drag a slider each to set the boldness and the italicness of the font. (I believe the technology failed because they were preposterously difficult even for professionals to make.) Adobe used the nomenclature "derived faces", and described sets of "multiple master fonts" as a family. So, for example, the MMF family that came with the NeXT was [jensen, miriam, kepler, something-forgotten], and you could buy other families - four or five fonts each, generally - for $80, which back then was an absolute steal.

Also, all the old typesetting books I inherited from my grandfather use the term as I had believed and presented.

Thank you for clarification on the matter.

John Haugeland is http://fullof.bs/

Thomas Phinney's picture

@Miha:

Yes, I really think he's right. A regular and a bold isn't much of a family, but it's a family, so even that version of your claim isn't technically true. Of course, when you wrote the original post you only had one face, right, so you're kind of retconning the whole thing when you retitle your post, no?

Not that I think this is the biggest deal in the known universe. But then again, dang, anybody who is seriously into type must be obsessed by minutiae. Comes with the territory, right?

@Stonecypher (John):

What I'm saying is that Corbel and Constantia, for example, are two distinct font families, and that you're in a very tiny minority (like <1%) to define them otherwise. If you can find any typesetting book that has two fonts with unrelated designs, by different people, with completely different names, which calls these two fonts part of the same family, I'd love to see it.

I'm rather familiar with Adobe's multiple master fonts, having done my MS thesis on multiple master technology, and working in Adobe's type group for over 11 years. I never heard anybody use the term "derived faces" the way you do. (There is a font production The set of multiple master fonts that were described as a family would be an upright and an italic, or those plus some expert set or something. It's not like Adobe would lump Minion and Myriad (not "miriam") together and claim they were one family.

I'm not sure where you think Adobe uses the nomenclature you gave as far as the definition of "family." For example in the glossary of type terminology on the Adobe web site, the definition of both "family" and "font family" reads: "A collection of faces that were designed and intended to be used together. For example, the Garamond family consists of roman and italic styles, as well as regular, semi-bold, and bold weights. Each of the style and weight combinations is called a face."

So, the example certainly doesn't support your idea. You can choose to read "designed and intended to be used together" as broader than I believe the writer of that sentence intended, I suppose.

Cheers,

T

William Leverette's picture

Right, which is why the second paragraph reads as this exactly opposite statement:

Airwindows formally thanks Microsoft for devising ClearType, and thereby figuring ways to add strange color casts to type on LCD screens, while simultaneously reinventing a technique they may well have first heard of from me.

John, please revisit the linked web archive page from above in order to peruse it and correctly parse this quote (if not the irony of it). A Smoothtype-assisted 2x supersampling AA technique – laboriously upsampled by hand – more accurately describes this author’s work.

While on the subject of Smoothtype, please allow me to briefly digress into my own nostalgia: Ah, I remember that hazy evening when I installed it, Kaliedascope and Bill Bart's BBX Mercury theme on my sister’s brand new 3rd-gen Grape iMac...That was some fine blurred-out Segourney-Weaver-CRT-smashing shit right there. So much so that, later on, I just had to replicate the hat trick on her college-bound 2nd-gen iBook – what with that super-sexy subpixel feature finally turned on! You see, the font smoothing algorithm introduced in OS 8.5 used a type of hinted antialiasing sans subpixel rendering that looked mighty chintzy to me next to just about any supersampling technique then available.

So yes, Smoothtype did feature the option of subpixel accuracy for setting antialiased text. But that’s not at all what your quoted source was seeking. In that very quote, no less, he scoffs at MicroSoft’s über crispy ClearType. Now, to demonstrate the civility, respect, and effective communication of posting an inline image to illustrate my point (as Mija rather genially requested above), an image from Airwindows:


See what he did there? We’re peering at a completely different species of “subpixel”.

Kindly,
William Leverette

muji's picture

I totally agree with Thomas on which fonts can comprise a font family.

A type family is based on the unique characteristics of a single typeface, eg, Arial (Regular), and contains fonts which vary the weight and/or width of that original weight. Even 'super' families such as ITC Stone (which has four subgroups - serif, sans serif, informal and humanistic) are still derived from the basic shape and character a single typeface. A "group of fonts meant to work well together" is just that. I don't think it even has a name unless it is 'sympathetic fonts'.

The simple visual demonstration of a type family is Univers (see below).

Adobe sells Univers as a 'type family': http://store4.adobe.com/cfusion/store/html/index.cfm?event=displayFontPa...

Adobe sells separate fonts grouped together as 'collections', eg:
http://store4.adobe.com/cfusion/store/html/index.cfm?&store=OLS-US&event...

Strike throughs and underlines aren't additional weights in a font family because the weight of type they underline or strike through already exists. They are an extension of an existing weight. (Has the confusion arisen because software programs behind the scenes treat each extension or added style/effect as a separate weight, or are they created by automatically stylising an existing weight?) Programs like Word then muddy the waters further by allowing a weight like Arial Italic to be both a separate font AND a 'style' of Arial Regular!

miha's picture

"Yes, I really think he’s right. A regular and a bold isn’t much of a family, but it’s a family, so even that version of your claim isn’t technically true." (Thomas)

"I can furnish definitive evidence that a type family must consist of at least four members ;)" (William)

I am not going to discuss what is type family and what isn't. I said I was going to change the title according to Thomas' opinion, and so I did.

"See what he did there? We’re peering at a completely different species of 'subpixel'." My thoughts exactly. Posted as an answer to subpixel type, and by such knowledgeable person – what were the intentions? I said it was irrelevant, but didn't write further induction.

Thanks for commenting, William and David.

I hope that the discussion continues on the subject of subpixel design.

(PS: I had to use improper characters in the title)

muji's picture

I hope John will now let sleeping dogs lie on the subject of the title.

In deference to Miha I'm not going to say anything further on what is considered a type family but I would beg John to look a bit deeper into it himself before he decides to persist with his current understanding.

Thomas Phinney's picture

On the meaning of "family" in typefaces/fonts, see also:
http://www.thomasphinney.com/2009/04/font-terms-survey-results/

On the differing meanings/intent of "subpixel" rendering and font design, I concur that many of StoneCypher's links and discussions have to do with, well, something other than the LCD-oriented ClearType-style rendering we're looking at in Miha's work.

I also think that not being "first" in one sense or another isn't always that big a deal. It's still a really nice piece of work, being done with different intent and probably with different implications than the earlier work. I'm reminded of the discovery of Viking visits to the new world pre-dating Columbus. Fascinating and cool, but didn't really change the importance of Columbus at all.

BTW, Miha, have you considered using Photofont technology as a delivery vehicle for this? Seems like you could use something like that....

Cheers,

T

muji's picture

Thomas

I note that nearly 30 percent of your respondents consider bold and italic not to be separate fonts. Do you think it's the curse of the B and I buttons in some programs? They know to press them creates bold and italic but do they also realise that in most cases a real bold or italic font is called up rather than the chosen weight is actually emboldened or italicised?

StoneCypher's picture

Thomas Phinney: I may be misunderstanding and/or misremembering, then. Thank you for clarifying, and I retract my statements on font families.

William Leverette: I appreciate the sentiment, but I believe that you are mistaken; notice the text at the top which says "Now with subpixel smoothing for LCD screens!" That is unambiguous regarding whether the man had subpixel 3-column hinting, so right there you have a somewhat clearer answer. His hand-made fonts are also subpixel, though they're color bitmap like Miha's rather than positional subpixel map like some other examples.

Smoothtype is just a seperate thing he makes. I'm talking about the hand-made fonts, not his plugin for the standard system fonts. It is not uncommon for system-level font enthusiasts to attempt multiple technologies.

He makes fonts, and he makes a program that attempts to smooth out things not using his fonts. (He actually makes other stuff too; poke around his site for a lot of fun ideas.)

Either way, I'm not sure why we're belaboring this particular site; it isn't among the earliest several, it just gives the examples of full families of fonts that Miha was asking for.

Miha:

My thoughts exactly. Posted as an answer to subpixel type, and by such knowledgeable person – what were the intentions?

If I were wrong there, it would be a single mistake among many other things. Everyone is human (except Dick Cheney). Mistakes are not evidence of foul play. As Lord Bryce said, apparently possibly regarding me, "never attribute to malice what is adequately explained by stupidity."

I notice that you've updated the title, above and beyond what the moderators chose. Thank you for that; I very much appreciate your show of good faith. I retract whatever I said about honesty; in the end, you made it right.

Thomas Phinney's picture

Great, now you two can kiss and make up! Or if that isn't acceptable, you could both kiss me, and I could pass it on.... :)

@Muji: Yes, a noticeable chunk of people think of "a font" as including bold and italic variants. But among the general public, a high percentage also don't realize that bold and italic are commonly distinct human-made designs, rather than special effects simply created by the computer. Of course, Windows and Mac OS don't help matters by doing faux bold and faux italic when the real fonts don't exist.

Cheers,

T

miha's picture

So, how to continue? With further additions, of course:


There are some glyphs that are not yet ready, but generally speaking I added:

  • superscript and subscript numerals in both lining and old-style forms (quiet unusual for (sub)pixel font, isn't it?;))
  • predefined fractions which are in Adobe Latin 4 – OK, some may use some help
  • arrows + some triangles, also in Adobe Latin 4
  • some currency glyphs
  • I just started adding accented characters
  • case sensitive forms (dashes, parentheses etc)

How is a web designer going to use the typeface?
I also thought about Photofont, but it doesn't support type families (or does its use become unnecesary complicated). Furthermore, it doesn't support both lining & old-style figures and other feautures. To be fair, this is not supported by CSS either (!).
Using the sIFR-like method is in my opinion the best. I think I will take parts of sIFR and make it appropriate for my typeface. I am doing text layout program in Flash anyway.
Using sIFR-like plugin for webpages has the following benefits:

  • it is automatic and text is easy to change and style (with html and css)
  • if user scales the webpage, the typeface should be changed to a bigger system font – instead of smoothing the type and breaking a subpixel effect like in static images
  • users on iPhone see the regular text (if rotated, subpixel effects on iPhone don't work)
  • it is also easier to "install" than sIFR because typeface is already saved in proper format
  • it suggests the idea that The Flea's Knees should be used only in extremely small quantites (like in headlines, but just the opposite size)

Another option is use in predesigned pictures.

Personally I think that such small text should be used very rarely. I made it because I like the optical illusion and making something in a "new" way.

And now a question … has anybody brave tried doing a small subpixel font or maybe just some glyphs? I would be glad to see some other examples. I can only study the work of Ken Perlin – subpixel typefaces from decades ago don't work anymore on LCD screens, and there are no 1x magnifications of fonts made for game devices.

gingerbeardman's picture

This is interesting: textorize produces beautifully subpixel antialiased text graphics on OS X, through a Rubycocoa Ruby script.

http://mir.aculo.us/2009/09/29/textorize-pristine-font-rendering-for-the...

site is currently down for me, so i used the google cache

StoneCypher's picture

Gingerbeard Man: here's the Win98 (?) tool I threw together a billion years ago to dump them. It looks kind of broken in Vista, due to changing border sizes, but it's usable; in XP mode in Win7 it works exactly as expected. It's going to dump a horizontal PNG with Microsoft's cleartype filter. I confess to a strong preference for Microsoft's filter over Apple's.

The two pulldowns at the bottom are the character range to dump (you can do limited ranges, or it's good up into unicode.) It isn't smart enough to skip unimplemented characters.

http://sc.tri-bit.com/outgoing/DumpFont.exe

Please note that both my tool and the Ruby tool are fundamentally flawed: they're generating images, which are not equivalent to a subpixel font because they lock the characters into one of the three primary physical positions. I re-render each character individually to keep from generating weird pseudo-kerning effects. (The ruby script might too; I haven't looked.) However, real SPAA requires a live font engine with knowledge of the screen and access to a vector font; image dumps can never be a real subpixel font (it could be faked with three subpixel image sets and some complex positioning, I guess.)

Obviously, as an executable from the internet, scan it seventy five bujillion times with antivirus.

John Haugeland is http://fullof.bs/

Dan Gayle's picture

Wow. I'm sorry I wasn't paying attention to this. A LOT of great resources here for people who have to make web icons, i.e., me :)

StoneCypher's picture

Dan: icons shouldn't be subpixel anti-aliased. They'd be visibly improved only to about 30% of customers, and visually impaired to nearly 70%.

The market is not homogenous enough for rendering strategies to be enfored at the image level.

Thomas Phinney's picture

Years ago I was thinking that cell phones and the like might be best served for their main system font by a custom bitmapped sub-pixel anti-aliased font with three different versions of each glyph, one for each possible subpixel x-starting-position.

Cheers,

T

miha's picture

In case if you are asking when I am going to finish my typeface: I am working on it! Special characters and feautures are nice, but they do take their time. I also see that accompaining website would be appropriate (not on its very own domain though).

"icons shouldn’t be subpixel anti-aliased. They’d be visibly improved only to about 30% of customers, and visually impaired to nearly 70%."
I think they should be. Nothing wrong if they aren't, but why not take the advantage of the hardware.
Is there any source for the 30% claim? I don't think it's true. I think the majority of web users use LCD screens with RGB order of subpixels, on which the subpixel information is displayed correctly. According to Microsoft: "a very small number have a striping order of BGR [on which subpixel design isn't displayed properly]."
Of course there are users browsing on rotated mobile devices and other display technologies, but they are in minority. I can't check my typeface on CRT display, but Steve Gibson's opinion is "sub-pixel rendering is better than nothing on CRT's", so it can't look bad in absolute terms.

This is just … web design. Sites aren't displayed optimally for a small number of visitors in a vaste range of cases: for example because of having javascript turned off, no Flash or just too old plugin, outdated browser, (old) mobile device, setting of non-native monitor resolution, zooming in and displaying blurred pictures etc.

"three different versions of each glyph, one for each possible subpixel x-starting-position"
I won't deny that this thought didn't cross my mind :-)

miha's picture

Matt – "This is interesting: textorize produces beautifully subpixel antialiased text graphics on OS X, through a Rubycocoa Ruby script."

Nice! This is exactly the perfect kind of "response" to a discussion about subpixel rendering.

The use for my typeface is very limited, but its purpuse could be more of an educational one: if people learn basics about subpixel rendering they may also quickly notice that the quality of rasterisers differs, and consequently response in a negative way to improve the situation. Look at this test for comparisons in text sizes (made by Karsten Lücke).

StoneCypher's picture

Is there any source for the 30% claim? I don’t think it’s true.

Yeah. Try doing some research, Miha. You really shouldn't be challenging other people's honesty at this point.

The reason I didn't look up the numbers for you is because the last time I did your research for you, you complained at length about how there was too much of it, and never got around to reading it.

It's amazingly inappropriate for you to call someone a liar based on guesswork.

I think the majority of web users use LCD screens with RGB order of subpixels,

I'm sure you do.

According to Microsoft: “a very small number have a striping order of BGR [on which subpixel design isn’t displayed properly].”

Yeah, it turns out that there are a lot more web-enabled devices than the ones that run ClearType.

Of course there are users browsing on rotated mobile devices and other display technologies, but they are in minority.

Ah, made-up data. The correct number is easily gotten from Google statistics. Two cellular phone browsers are now in the top 10 browsers on earth, and both are usually used landscape.

Please stop refuting other people's claims based on your imagination, Miha. That isn't how this works.

but Steve Gibson’s opinion is “sub-pixel rendering is better than nothing on CRT’s

Yes, for text. Not for icons. The reason it's better for text is that the 3/9 filter that you don't use approximates the anti-aliasing that you don't use. Steve's opinion is based on two things you don't actually do, and applies to monochrome text rather than colored images.

so it can’t look bad in absolute terms.

Yes. It can. In fact it does.

Sites aren’t displayed optimally for a small number of visitors in a vaste range of cases:

The number is 70%, Miha, not "a small number of visitors" based on things you imagine.

Making tiny improvements for a minority of customers while ignoring the wrongness to a significant number of customers which is much more unpleasant than leaving things as was just isn't that smart.

aluminum's picture

StoneCypher, you seem to have knowledge. To bad it's wrapped inside of sarcastic rants.

Anyways, if anyone has a specific source for that 30% statistic, it'd certainly be interesting.

I do see how sub-pixel rendering pre-system level is more of a limited-use experiment as you are at the mercy of the particular display technology and the the orientation of it.

As stated, the use of many hand-held screens being used horizontally is an issue. There's also problems with different pixel setups with screens (such as the OLPC screen which uses an interesting sub-pixel matrix).

Please correct me if I am wrong, but this is really only useful specifically when there either a white and/or black background as well.

Still, fun experiment!

Thomas Phinney's picture

It's an interesting problem. When I was thinking about having different glyphs for each x-position, I wasn't considering rotatable screens at the time... now I have two, my iPhone and my main jumbo LCD monitor at home. So you'd also need either three more for "rotated" mode, or at least one plain grayscale glyph for rotated mode. But what do you do about spacing, if you were doing sub-pixel incremented spacing before? You'd have to change the spacing and potentially have reflow as part of rotating the screen. That might not be a big deal as you'd be potentially changing line length and such anyway... hmmm.

Cheers,

T

lopez's picture

Stonecypher, I am very curious about the thirty percent statistic. I cannot seem to be able to find a source for that. Would you mind providing a link? I don't think miha was challenging you, it seems to be a genuine question, one that I have too.

Frankly, I'm also having trouble believing this statistic. Is that including every display ever made still working?

Thank you in advance!

dberlow's picture

SS> Two cellular phone browsers are now in the top 10 browsers on earth, and both are usually used landscape.

And don't both have really high resolution so that the orientation of the stripes is insignificant?

Cheers!

Bert Docks's picture

I think it's a wonderful font, far more legible than others which are scaled down through Cleartype or similar post process filtering. I'm deeply saddened that someone should come along and, while providing some interesting information, miss the point so completely and go about ruining someone's day by trying to discredit their excellent work because they objected to a little excited (and deservedly so) wording.

I for one will use be using this font in all sorts of places it wasn't designed for. Keep up the great work, Miha. And don't feed the clearly jealous troll.

miha's picture

Thanks Paul, I feel in the similar way.

The process of designing this font has slowed down. Well, it is a font with many glyphs and features, especially for a pixel font, but I have to admit my excitement also become smaller (it actually happens every time I learn about a new, exciting, specific, geeky area of knowledge).

Btw, I am glad this post made so many views, I really like Typophile, and I see it as some kind of contribution.

neumann's picture

This is a VERY nicely detailed piece of pixel by pixel handiwork.

However.... I think that all the attention to colored pixels is unnecessary and actually distracting. I have yet to see a cleartype or sub-pixel implemention where I am not noticeably distracted by the color fringing.

Look at the image below. It is one of Miha's lovely specimens converted to grayscale by selecting only the green color channel in photoshop. It could be refined by hand from here to wring a teeny bit more detail, but it is a pretty good automatic conversion.

While it can not perfectly match the detail of the colored pixel version, it is preferably to me for the lack of color fringing.

neumann's picture

here is a side by side.

miha's picture

I have yet to see a cleartype or sub-pixel implemention where I am not noticeably distracted by the color fringing.
I am not very distracted by colors on my 17″ laptop with 133 PPI, which is higher than an usual display. On the other hand, I can still clearly see colors (if I don't sit too far away) on an average display. It distracts me, but grayscale rendering is even more blurry and that distracts me even more.
I also don’t like aliased fonts (OK, there are exceptions). Aliased font like Georgia wasn’t even made for LCD display and I don’t think the old master had this in mind when he designed it for the CRT screens. However, I think I could get used to aliased fonts with very carefully antialiased curves. For example, Mana.

While it can not perfectly match the detail of the colored pixel version, it is preferably to me for the lack of color fringing.
Well, it is your opinion … but this is partly my fault. Color fringing is actually too pronounced and if I continue with this typeface, I will find a better balance between fringing and blurriness. There is also something else: When I was doing this typeface, I started with colors which were produced by Windows ClearType GDI rendering. I changed them just a little and didn't do a research to find out what are actually the best colors. Read it in this thread, where I criticized that the lightness of red/blue colors is programatically symmetrical, but not perceptually.

I think that all the attention to colored pixels is unnecessary and actually distracting.
I mainly share your opinion for regular and bold weight, but not for italic and bold italic. I think subpixels can really help in displaying a nice italic font without too much blurriness. Well, an example of my font with different levels of color fringing would be nice (I hope I can manage to post it soon – made with filters)

PS: Thank you for your comments! At last, some argumented criticism :-)

Syndicate content Syndicate content