Sign up ×
Graphic Design Stack Exchange is a question and answer site for Graphic Design professionals, students, and enthusiasts. It's 100% free, no registration required.

Following on from this question: TTF and other "modern" font systems, and font size differences

Higher quality fonts contain hinting information, which in short better fits glyph boundaries to a raster grid.

It's commonly used on screen where it reduces anti-aliasing.

But is hinting information used at all by any print devices? If so, which kind (desktop laser/inkjet/imagesetters for litho etc) and when does it make a measurable difference?

I'm looking for direct references to it being used by print devices, and failing that, some empirical measurement (eg comparison of hinted font type /unhinted font type text/converted-to-vector type)

(Why don't I do it myself? I'm no longer in the industry so don't have the tools unfortunately).

share|improve this question
I don't think it REDUCES anti-aliasing on screen. It's merely a custom form of it for that particular font size. It's a good question, though. My understanding has always been that it's for low-resolution raster output (mainly screens, low-res laser printers in the past...) – DA01 Feb 9 '12 at 16:48
@DA01: It's merely a custom form of it for that particular font size. – While hinting may lead to this, it does not have to be. The most common way of hinting (as far as I can tell) is to include extra information in the font that tells the renderer where stems, baselines and similar are. This is not size-dependent (but less useful at higher sizes). – Wrzlprmft Jun 16 at 5:32

3 Answers 3

up vote 7 down vote accepted

Any printer driver worth its bytes pays attention to hinting (otherwise the other drivers would take it out behind the boot sector and beat the c*** out of it). Any RIP does, also. Hinting was originally developed for low-res printers (a 300-600 dpi laser printer is a low-resolution device), but used also for on-screen rendering. I found a good article from TUGboat that covers the subject well and simply.

To illustrate the point, here's a test done today using regular office copy paper on a standard non-Postscript office laser printer, directly from Illustrator. The font is Minion Pro Regular at 12, 9 and 6 point. At each size, the text block was copied and the copy converted to outlines. All six samples were set up on one sheet and scanned at 600 ppi:

12pt text:

12 pt text

12 pt outlined:

12 pt outlined

9 pt text:

9 pt text

9 pt outlined:

9 pt outlined

6 pt text:

6 pt text

6 pt outlined:

6 pt outlined

share|improve this answer
I'm not sure if I see a difference, but +1 for the effort! – DA01 Feb 13 '12 at 4:07
To me it's very obvious, but if you superimpose them (say, as layers in Photoshop) and turn the top one on an off, the many differences are immediately obvious. Look in particular at the serifs, and the vertical strokes (shape of the "legs" of the lowercase n, m and h in particular). – Alan Gilbertson Feb 13 '12 at 5:56
@go-me: 1) Why convert the text to outline? – To destroy the hinting information. 2) You won't find any font hinting on your OTF font – I have not tested that particular font, but yer, OTFs and TTFs may contain hinting information. 3) It doesn't lose any quality. – Of course it does; you have to rasterise at some point when printing. – Wrzlprmft Jun 16 at 5:43
@go-me: As for TTF it's on wiki. No hinting because they're like vector. That’s just wrong: TrueType fonts can contain hinting information, only that its not tons of bitmaps but more intelligent information like stem markers, baseline markers, relevant stem sizes and so on. If your claim were true, reducing a TTF font to outlines should also have no effect on screen rendering, which it cleary does – I just tested it (top: font; bottom: just vector outlines). – Wrzlprmft Jun 16 at 18:23
@go-me: The font not vectorized is being processed in real-time by your font engine. – Yes, and this process is called hinting, amongst others by the very links you gave. Thus it clearly contradicts your claim that there is “No hinting because they're like vector”. Moreover, I added a third variant in the middle by removing all hinting information stored in the TTF font and just having the font engine do the hinting. And again, it is different (note how the u is out of line). – Wrzlprmft Jun 16 at 20:17

Check this article at Typotheque: Click Here

Especially Mr. Bil'alk's response to Sebastian in the comments --> here

"...fonts are typically hinted up to 50 ppem (they are usually not needed in higher resolution), hinting effects will be visible in sizes smaller than 12pt at 300dpi, 6pt at 600dpi print, or 3pt at 1200dpi. Sometimes fonts are hinted to to much higher ppem, in which case the hinting will be visible in most text sizes in print."

Hinting is used for any rasterization processes - print or screen. The primary difference between screen and print is how noticeable the hinting, or lack of it, may be.

share|improve this answer
However, this doesn't actually provide concrete evidence that it's implemented on print devices. Would be great to have confirmation of this. – e100 Feb 9 '12 at 13:59
So, it seems that the answer is that it can be used for any output, though in practice, when it's used, it's primarily for small sizes on display screens. – DA01 Feb 9 '12 at 16:49
Actually, font hinting was originally implemented for low resolution printing, not for digital display. This is why they were first implemented in PostScript fonts. And since modern fonts evolved from PostScript fonts, they inherited hinting. So RIPs definitely look at hinting if it's supplied for the resolution in question. After all, hinting is by definition the fine tuning of the raster output of a font, so why wouldn't the Raster Image Processor read it when rasterizing the font? So the question is whether the font designer provided hinting for the dots-per-em being printed. – Lèse majesté Feb 10 '12 at 6:55
@Lèsemajesté - this would be better as an answer. – e100 Feb 10 '12 at 17:33
I think we all agree that the technology is there and it CAN be used for that. But what is being asked is if it's used at all these days for anything above maybe 300 dpi laser printers (which I'm not even sure exist anymore). – DA01 Feb 13 '12 at 4:05

TLDR: You might not want to worry too much about hinting-and-printing (maybe this is an opinion).

To answer to the question:

"But is hinting information used at all by any print devices? If so, which kind (desktop laser/inkjet/imagesetters for litho etc) and when does it make a measurable difference?":

PostScripts and printing

Simply put, font hinting is a screen preview of the postscript font with smoother edges. A postscript font itself looks like a bitmap before being rendered and that's how the rip likes it because it give a sharp clear edge (see image below) on plates. That's also really how fonts and vectors look like once printed: without any anti-aliasing.

Postscript fonts are self-contained. The hinting itself was adjusted by the typographers who created the fonts and rip has its own algorithms to adjust the font before "rasterizing" it; for example, using the font at 11.5pts instead of 12pts, the rip will make some maths to adjust the size and rendering based on the data it has, usually the 12pts data. Each platform or system actually needs to render fonts, so font are dynamic in a way. Postscript fonts deserve a lot of respect for this reason only, hours of work have been put into them in order to create such a perfect rendering.

How a postscript files looks like to the rip

The postscript font by itself contains 2 files per font; one for screen (display font) and one postscript for the rip system (for printing through postscript based systems).

Each postscript font comes with a file for each style also; that's also why they can be heavier than other fonts, they really contain more data. The italic for example may not work on some of them if they were not created with an italic version. The rip will only replicate a "fake italic" rendering or not at all. In theory, that font in italic simply doesn't exist to the rip. On a side note, that's why it's good to not use any artificial styling with postscript fonts (eg. using the "bold" button on a publishing software instead of using the real font "Helvetica bold".)

Example of postscript font scree/display and postscript

Postscript fonts would look ugly alone on a display because it appears "unprocessed", with big sharp pixels, similar to a bitmap (see first picture.) That's why there is an extra file in a postscript font package; it's what's called the "screen" or "display" font. Technically, the Rip doesn't need it but it's hard to verify a proof on screen without it. The hinting part on screen is handled by the screen font, and the rip uses the postscript one with its own data and the data contained in the postscript version of the font. The screen font is simply a kind of overlay low resolution font that temporarily replaces on screen the postscript font for visual reference only, it's not the one that is printed. It adds some kind of anti-aliases effect and makes the font looks smoother on screen. When you print a postscript font on a non-postscript printer (eg. most inkjets), that's why you see it in low resolution; it only uses the low resolution screen font to print.

In the end, all fonts and vectors are converted to a very high resolution bitmap when printed. In the case of postscripts, it's the Rip that does all the work of rendering the fonts with specific algorithms before "flattening" them. That's what ripping really does; it decodes the postscript mathematical data, read them and re-encode them. There's usually a part kept as raster and one for the vectors on most modern rips but ultimately all this get crush together when being output on a plate or a film. That's why I say it all get "rasterized" in the end.

TrueTypes and printing

For TryeTypes and OpenTypes, that hinting is managed by the font engine of the computers/system they're used on and the algorithms contained in the font. Then the information is sent to the Rip in a similar way vector information are transmitted. The results are more WYSIWYG in that way and the rip doesn't need to really process them any other way than a vector.

In a way, font hinting is similar to fonts as anti-aliasing is to images. Logically (Disclaimer: I haven't personally tested TrueTypes and Rip and films), the vectorized version of a TrueType vector font should be very close to what the Rip will really use to encode its own data. One way or another, the font will need to be rendered by one system or the other (yours or the printer's system), and then will be encoded to be sent as postscript.

THIS is how it really looks like at small size:

font hinting real example of effect printed with hinting and no hinting

Why care about hinting for printed projects? When does it make a measurable difference?

A lot of designers confuse what they see on screen and what is being really printed, AFTER being ripped. And this question is not about screen and web, it's about the relation of hinting in the print world.

Obviously it's very hard to verify as you cannot really compare the printed version with the screen one perfectly without scanning (putting back to digital) again that print... which might cause distortion that will change the results; the same kind of issue happen for pretty much anything that needs to be converted many times from a medium to digital to a medium, and usually the last version will suffer of some distortion compared to the original one. In this case, we're talking of distortions that may be at the level of 0.05 pts or 0.001 pts and less, even less than trapping (eg. 0.144pts for example, a "fat" standard trapping)... it doesn't take many conversions to lose accuracy in this case. Personally, I think "homemade" tests in this matter cannot give reliable results.

That's why I mentioned in another comment that this could only be verified properly if done with an expensive drum scan, and maybe use a film as medium (not paper and really not using an inkjet). Small home or office scanners' quality are way too low to be reliable in this case when compared to commercial scan or drum scans. Of course there might be a noticeable difference if you compare an on-screen 4pts size font with a 4pts print... but then again, when will you ever print at this size? Even on big font you will notice a small difference. Other factors will also make the thickness of that font change slightly, such as the dot gain for example and the type of stock or print method being used.

Comparing a vectorized font with an "active font" is like comparing an active scene seen with your own eyes with a picture of a scene. The hinting on screen is something dynamic, the print or vector is not dynamic in the same way. Obviously there will be a slight difference, you can't expect to see an "active font" being rendered on screen the same way as a vector, the same way you can't expect a vector and a raster image to look the same when zooming in. When you vectorize a font, it has the similar effect to tracing a low resolution in Illustrator; the anti-alias might or might not be rendered depending on the settings.

I have personally compared many media together because that's part of the quality control process as prepress technician; you use a film and a printed proof, and simply compare them together on a light table. I can tell you there's even a slight (very very tiny) difference when comparing a film and a laser print, and even a film and the final print. So this has more to do with the Rip process and algorhytm of each machine than the font itself. Technically speaking, they should be identical everywhere when not vectorized.

So why you should care or when... well technically I think you shouldn't. Even if there is a small difference, that difference is so tiny that your design should not suffer in any way because of that extra "conversion". And the other thing is... you don't have much control on the type of RIP being used, the fonts in it, the way it processes your texts/vectors and even the trapping at this point. These things too will change the way some elements look like once printed. Actually, trapping is what you should be worried about if you are really anxious about such small details and perfect rendering. One thing for sure, even a small 3-4 pts text or anything vector based will always look sharp and "hinting" has nothing to do with the readability (or lack of) at this point. The issue is in the choice of medium, the design, the trapping and these other factors.

Now this is based on my personal experience as a prepress who worked with different old and new RIPs, different media such as papers, films and plates, the different scans, and by having seen the differences between a film vs a laser print vs a preview from the Rips vs the final results. I'm not trying to appeal to authority (I'm simply answering) but the differences are so small from screen-to-print, they're not worth being anxious about them and frankly there isn't much you can do once your data are sent to the Rip. technically speaking, the Rip seems to treat vector based fonts that are not Postscript in the same way as any vector; so I guess the vectorized version of a font is very close to what you'll ultimately get on the rip. I personally remember doing most of my observations with Postscript fonts so I cannot confirm this besides pointing out the resources explaining how TrueTypes are processed.

If a project requires the "hinting" to be printed perfectly as on screen, then the font should be changed for a thinner one and the COLORS used (because of trapping) should be the main focus. If "hinting in printing" becomes an issue, that means the designer or printer did a bad job or chose a font that could be problematic for the project.

I think this answer has some good bits when speaking of print-and-hinting, I don't really see where are the mistakes. So I undeleted it even if it got a lot of downvotes. Maybe someone who has worked on Rip systems could point out what needs to be reviewed with some explanations and I'll gladly edit it. I unfortunately cannot provide a digital example as 1) it's expensive to scan on drum, 2) there's a cost to getting a film as well and 3) I no longer have access to the equipment either.

It's not a simple question since hinting is a display related concept and printing is something totally different. So obviously, it gets very technical and maybe someone from Agfa, Unisource, Scitex or Kodak could be better at answering this with more precision and real technical data.

PS: Postscript fonts can be vector or bitmap as source says below.

Font hinting (also known as instructing) is the use of mathematical instructions to adjust the display of an outline font so that it lines up with a rasterized grid. At low screen resolutions, hinting is critical for producing clear, legible text. It can be accompanied by antialiasing and (on liquid crystal displays) subpixel rendering for further clarity.

For the purpose of on-screen text display, font hinting designates which primary pixels are interpolated to more clearly render a font.

The open-source FreeType 2 font rendering engine uses an auto-hinter when such hinting data are not present or their use is restricted by a software patent.

Postscript fonts: A scalable font technology from Adobe that renders fonts for both the printer and the screen. PostScript fonts come in Type 1 and Type 3 formats. Type 1 fonts use a simple, efficient command language and are widely used, but Type 3 are not. Type 3 fonts can use the entire PostScript language to create complex designs, and Type 3 fonts can also be bitmaps.

source:, wikipedia,

share|improve this answer
A postscript font itself looks like a bitmap – Nope, postscript is based on vector data. Anyway, even if it weren’t, you have to obtain the bitmap data somehow, and that’s where hinting comes into play. Moreover, could it be that you are confusing hinting and anti-aliasing in your first paragraph? – Wrzlprmft Jun 16 at 5:49
Printing postscript fonts on an inkject and what will come out is only that hinting data. If you print postscript with only the screen font or a corrupted version, it will look like a bitmap. – go-junta Jun 16 at 15:35
Either you have a totally different definition for hinting or bitmap than me (and than quoted in your post) or you are misunderstanding something. Either way, I fail to follow what you want to say. Also, what do you understand to be a screen font or hinting data? – Wrzlprmft Jun 16 at 16:00
Here, all explained: – go-junta Jun 16 at 16:14

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.