[HN Gopher] My favourite fonts to use with LaTeX (2022)
___________________________________________________________________
My favourite fonts to use with LaTeX (2022)
Author : todsacerdoti
Score : 179 points
Date : 2025-05-17 02:53 UTC (4 days ago)
(HTM) web link (www.lfe.pt)
(TXT) w3m dump (www.lfe.pt)
| Xophmeister wrote:
| Palatino with Microtype is my go to for all my LaTeX documents.
| It looks so good.
| ahartmetz wrote:
| My standard font package is "mathpazo", which is Palatino with
| maths support. I obviously like Palatino - and if it isn't
| available, Garamond is similar.
|
| If ever have to do much LaTeX again though, I'll check out the
| alternatives because the mess of partially compatible modules
| and the troubles with figure placement are still bad in LaTeX.
| dhosek wrote:
| It's hard to see Palatino and Garamond as similar, but
| perhaps that's just my typographic training at play. Palatino
| is much closer to its calligraphic origins than Garamond and
| has a darker color on the page (the seldom seen Palatino Book
| weight is a great improvement over the Palatino Medium that's
| the default Palatino weight for extended text).
|
| Note also that Palatino was originally designed for Linotype
| hot metal typesetting and has incorporated in its design the
| limitations of that system (which, in some ways is actually a
| bonus for naive digital setting where ligatures may be
| limited or non-existent). The most obvious case of this is
| the lack of character kerns--that is, characters cannot
| extend beyond their typeset width. This makes the italics
| look cramped since, e.g., _d_ , _l_ and _f_ cannot reach over
| the following letter with their ascenders.
| ahartmetz wrote:
| They are both derived from medieval scripts and they have
| fairly bold serifs, right? Would you agree that they are
| more similar to each other than to Times New Roman or
| Computer Modern (which I both dislike for their thin and
| pointy serifs)?
| drob518 wrote:
| Palatino has been my favorite font since it first appeared on
| the Macintosh.
| Syzygies wrote:
| As the only oldstyle font in the "original 35" Postscript
| desktop publishing fonts (think "original seven" Mercury
| astronauts), Palatino saw overuse. Among some designers it's
| second only to Comic Sans for its amateur use infamy.
|
| I'm among the guilty. Palatino does appear spread out
| compared to alternatives, for better or worse.
|
| The article does note that Palatino was originally designed
| for display text. I had long heard that Hermann Zapf was
| horrified at its adoption as a body font, which I couldn't
| confirm. The best I could do was to find a quote,
|
| "One day I got up the nerve to ask 'Mr Zapf, what do you do?'
| He replied, 'I correct the errors of my youth.'"
| spookie wrote:
| Microtype is such an insanely good package, I love it
| sneak wrote:
| Palatino and Baskerville are underrated. I don't know why Times
| is the default in so many things.
| spongeb00b wrote:
| Back in school we were told essays had to be submitted (as
| paper printouts, this was early 2000s of course) in Times.
| The rebel that I am, I submitted them set in Baskerville as I
| loved how much better it looked.
| groos wrote:
| OP should try Scholax as well. I'm very partial to it.
| mhd wrote:
| I would recommend most fonts that Michael Sharpe worked on. He
| did some nice refinements on already-decent fonts, often to bring
| copies closer to the original[1]. Heck, I'd recommend them
| outside of LaTeX, too.
|
| Favorites out of those are XCharter, ScholaX, Etbb and Erewhon.
|
| Also have a look at Algol Revived, which is a remake of a font
| made for French Algol 60 books by famed type designer, Adrian
| Frutiger.
|
| [1]: https://ctan.org/author/sharpe
| fodkodrasz wrote:
| Your link is the most authentic LaTeX experience for beginners:
| fonts with no preview/screenshot on their homepage... LaTeX
| users already know it, why make screenshots :D
|
| The blog post does this some justice though, the author does
| live in the current century, and does have nice examples.
|
| /me is former LaTeX user from uni, not overly fond of it, for
| usability issues not unrelated to its legacy.
| nanna wrote:
| CTAN is the central repository for all TeX packages, it's the
| natural link to give. If you want to see examples it's as
| easy as clicking on the relevant package name (eg XCharter)
| and then click 'Package documentation' which will request the
| relevant pdf. TeX outputs into PDF so it's the natural medium
| for demonstrating and documenting things. It's got nothing to
| do with not 'living in the current century'.
| uniqueuid wrote:
| The word 'natural' is doing titanic work in your sentences
| here ... :)
| fidotron wrote:
| Great to see Bembo being properly appreciated.
|
| It's kind of ironic that a system that ships with Computer Modern
| doesn't end up creating more Bodoni/Didone fans.
| lanstin wrote:
| I hadn't noticed Bembo before but it was quite good looking.
| dhosek wrote:
| I'm a fan of the closely related Poliphilus which is a
| somewhat more rumpled interpretation of the same sources.
|
| Really, the whole Stanley Morison-era catalog of Monotype
| designs is great stuff. I think the only collection of type
| designs that rivals it is perhaps the Sumner Stone-era Adobe
| originals.
| dhosek wrote:
| I think a big part of that is that Bodoni-style fonts don't do
| so well at low resolutions (and 300 dpi counts as low
| resolution). It's somewhat akin to the problem of Optima which
| loses much of its subtlety when laserprinted even at high
| resolutions. Bodoni and Didot both call for fine hairlines
| which lose their grace at screen resolutions and don't really
| do that great at laserprinter resolutions either (at 300dpi,
| with "writes white" laser technology where the whole page is
| charged and the laser removes the charge from the white parts
| of the page, the hairlines of CM would often vanish
| necessitating the "writewhite" hack for the printer modes for
| those printers).
| delta_p_delta_x wrote:
| Indeed. To add some detail--Didone typefaces (th have very
| high contrast between thick and thin strokes, which at low
| resolutions seriously affects readability as the thin strokes
| either disappear altogether, or are badly hinted and anti-
| aliased.
| jovas wrote:
| I like some variant of STIX (I never know the exact difference)
| and Gillius Sans
| amoshebb wrote:
| stickstootext
| mynameisash wrote:
| I've heard so much praise for _The Visual Display of Quantitative
| Information_ , but why, I'll never know. I bought it after
| hearing someone rave about it, and I'll be damned if I didn't
| hate every page of the book. It felt like a rebuff to _The Design
| of Everyday Things_.
|
| The former is currently sitting in my car, and I'll be trying to
| offload it to someone who actually wants it.
| atrettel wrote:
| I'm curious. Why do you feel like Tufte's book is a "rebuff" of
| Norman's book? I've read both and find them complementary in
| many senses. The one criticism that I have found by Norman of
| Tufte's work is that Tufte preferred high "data density" but
| Norman argues that this is not always appropriate.
| mynameisash wrote:
| Norman talks a good deal about cognitive load and that good
| design is intuitive. Ideally, you shouldn't notice good
| design because it's near invisible.
|
| OTOH, I recall Tufte going on and on about cutting the "data-
| ink ratio" to the point of making graphs that we generally
| understand at a glance suddenly very unintuitive. I can dig
| into the book again if necessary, but I recall he essentially
| argued that box-and-whisker plots became just a few dots.
| There's meaning conveyed by the boxes and the whiskers, and
| changing that convention - even if it uses more ink than
| absolutely necessary - adds significant cognitive load.
| iisan7 wrote:
| VDQI explains why chartjunk is bad, and if it reduces the
| amount of chartjunk in the world, that is good. Many maintain a
| soft spot for VDQI probably because it's their first a-ha
| moment in terms of appreciating graphic communication.
| Certainly there are other directions one could go but most
| people are not designers. Although, I don't see the same
| contradiction between DOET and VDQI, I'm curious what you mean
| about that.
| flenserboy wrote:
| Quite fond of kpfonts. As much as I use XeTeX because of
| OpenType, I find myself going back to PDFLaTeX so I can get the
| benefits of Microtype. I tried some of my docs through LuaTeX,
| but the results, while fine, were still inferior to the two
| options above.
| jpfr wrote:
| Microtype support for LuaTeX doesn't look so bad. At least on
| paper...
|
| https://tex.stackexchange.com/questions/654089/microtypograp...
| kragen wrote:
| The Bembo variant I used for Dercuano, Derctuo, and Dernocua
| (without LaTeX) is Edward Tufte's ET Book, linked in the article,
| using the old-style numerals variant. Unfortunately its Unicode
| coverage is very limited. Fortunately, URW Palladio L (URW's
| freely-licensed version of Palatino) has fine Unicode coverage,
| so I used that and Palatino as fallbacks. Unfortunately ET Book's
| metrics are not very comparable to URW Palladio L's, leading to
| letter size mismatches when letters mix; its x-height is
| especially different.
|
| So in, for example, where
| https://dercuano.github.io/notes/finite-function-circuits.ht...
| says "S [?] S", the "S" is noticeably shorter than the other
| full-height characters. It looks a little bit better in the half-
| assed PDF rendering I produced with my hurriedly-written HTML-to-
| PDF renderer:
| http://canonical.org/~kragen/dercuano.20191230.pdf#page=1572
|
| The other big problem you can see on that PDF page is that I
| chose Latin Modern Typewriter Condensed (lmtlc) for fixed-width
| text so that I could get 80 columns onto the narrow cellphone
| screens I was targeting with the PDF, but lmtlc completely omits,
| for example, Greek, so the examples using Greek are totally
| screwed up.
|
| The formula display in that note is definitely worse than LaTeX
| would do, but I flatter myself to think that my half-assed Python
| script still produced better-looking math output than I usually
| see from Microsoft Word.
| behnamoh wrote:
| I might pay the karma tax on this one but I've come to really
| appreciate "Times New Roman" (or TeX Gyre Termes + STIXTwoMath).
|
| Like the OP, I used to care a lot about fonts. Heck, at some
| point my Windows boot time got slowed down because of the sheer
| number of fonts it had to load!
|
| I used to think the default Latex font gives off a "serious" and
| "scientific" vibe. And I thought to myself: why would anyone ever
| use TNR when more "soulful" fonts exist?
|
| Now that I'm older (33), I resort back to TNR or TeX Gyre Termes
| but with one change: I add "FakeBold" to text to make it look
| like old papers and books:
| https://x.com/OrganicGPT/status/1920202649481236745/photo/1. I
| just want my text to convey my thoughts, and I don't want any
| fancy "serifness" get in the way (so no to Bembo and Palatinno).
| ayongpm wrote:
| I swear I was looking for exactly what FakeBold did years ago.
| Thanks for sharing that!
| drwu wrote:
| Unfortunately, the rendering of fakebold (especially if the
| factor is not 2) depends on PDF readers and printers. Some PDF
| programs handle it well, some fix the factor at 2, some do not
| implement it at all.
| GavinMcG wrote:
| How much is FakeBold doing there, though? Easy to say it's
| "just" TNR, but if the features of TNR that make it
| characterless have been supplanted with something soulful, then
| haven't you just found a soulful typeface that you like?
| sneak wrote:
| the default latex font is called Computer Modern, and I
| specifically use a webfont version of it on my personal
| site/blog because of the "serious" and "scientific" vibe you
| describe.
| neilv wrote:
| For LaTeX documents like technical papers...
|
| Something like Palatino (or even Computer Modern Roman) for body
| text.
|
| But for headings, humble Helvetica looks good, and a bit less
| "academic". (I really dislike the default CMR at large point
| sizes.)
|
| For monospace bits, again I dislike the unusual-looking TeX
| default, so something serifed or otherwise clearly unambiguous
| (for "1" and "l", "0" and "O"), and thick enough to be legible
| (some Courier are too thin). Inline, at a slightly smaller point
| size than body text, to look proportional, and maybe a little
| smaller in code blocks.
|
| For a book, I was thinking something slightly flashier for
| headings, at least on chapters, maybe Linux Biolinum.
| yawaramin wrote:
| I was a heavy user of `pxfonts` back in the day. Good times!
| blt wrote:
| I always liked the look of ACM journals/conferences more than
| other venues. Their template uses Libertine, so it is my choice
| too.
| guizzmo wrote:
| Genuine question : why are all those fonts serif ones? I
| personnaly find them much harder to read than sans serif ones
| (including when printed on paper, not just on screens).
| MarcusE1W wrote:
| I think the general idea is that Serif-fonts are better
| readable in book style documents and Sans-Serif fonts are
| better for short documents or maybe computer displays.
| Synaesthesia wrote:
| I thought the whole point of serif fonts is that they are
| easier to read. But the author addresses your question:
|
| > This survey focuses on serif fonts as these are the usual
| choice for longer documents such as articles or books (although
| sans-serifs have become more popular for longer text in recent
| years). However, in keeping with the reasoning above, I have
| also selected accompanying sans-serif fonts for each of the
| seven roman choices below (all of which have maths support of
| some form or another).
| adrian_b wrote:
| It is likely that you find them harder to read mostly because
| you are less habituated to them.
|
| For many of the older generations, who have spent their
| childhood reading thousands of books printed using serif fonts,
| at a time when there was no easy access to computer terminals,
| serif fonts are easier to read.
|
| In general, the readability of a typeface is determined more by
| other features than by whether it is has or it does not have
| serifs.
|
| The sans serif typefaces have appeared first after the
| Napoleonic wars, as simplified typefaces, suitable for low-
| quality printing used for titles or advertisements that should
| be readable from a distance.
|
| Having simplified letter forms, sans serif typefaces remain
| preferable for low-resolution displays of for very small text
| or for text that must be read from a great distance.
|
| However, until WWII, the simplification of the letter forms has
| been pushed too far, resulting in many letters that are too
| similar, so they can no longer be distinguished. There are many
| such sans serif typefaces that have been modified very little
| from their pre-WWII ancestors, like Helvetica and Arial, which
| are too simplified so that they should be avoided in any
| computing applications due to the great probability of
| misreading anything that is not plain English.
|
| After WWII, and especially after 1990, there has been a
| reversal in the evolution of the sans serif typefaces, away
| from excessive simplification and towards making them more
| similar to serif typefaces, except for the serifs.
|
| A well-known example of such a sans-serif typeface is FF Meta
| (Erik Spiekermann, 1991), but it had a lot of imitators. Such
| non-simplified sans-serifs typefaces have e.g. traditional
| Caroline shapes for the lower-case "a", "g" and "l" and also
| true italic variants (i.e. not just oblique variants).
|
| Besides the removal of serifs, a traditional simplification in
| the sans serif typefaces is the removal of the contrast between
| thin lines and thick lines, making the thickness of the lines
| uniform. At least for me, any long text printed with a font
| with uniform line thickness looks boring, so I strongly prefer
| the sans serif fonts that go even further in their resemblance
| with serif fonts, by having thin lines and thick lines, for
| example Optima nova and Palatino Sans.
|
| While any sans serif typeface by definition does not have
| serifs, when Hermann Zapf has designed Optima (which was
| released in 1958), he has found an alternative to serifs, which
| achieves a similar optical effect. Starting from a line that
| has the form of a long rectangle, instead of attaching serifs
| to the ends, one can make the 2 lateral long edges of the
| rectangle concave, instead of flat. After that, one has 2
| alternatives for how to terminate the ends of the line. The
| first is to also make concave the 2 short terminal edges. This
| results in sharp corners for the line and it is the solution
| chosen by Zapf in Optima. The second method of line termination
| is to keep the terminal edges flat or even slightly convex and
| to round the corners where they meet the concave lateral edges.
| This is the method chosen by Akira Kobayashi in Palatino Sans
| (2006), under the influence of the similar line terminations
| used in Cooper Black and in the rounded sans-serif typefaces
| that are popular for public signage in Japan.
|
| This alternative to serifs, with concave lateral edges, is in
| my opinion superior both to serifs and to classic sans serifs,
| but unfortunately it is effective only on very high-resolution
| displays or on paper (even a cheap laser printer has better
| resolution than the most expensive monitors), because on low
| resolution monitors any slightly concave edges will become
| straight.
|
| In any case, for the best readability, I never use fonts with
| ambiguous characters, like Helvetica/Arial and most other sans
| serifs. Instead of that, serif fonts are better, but even
| better are good modern sans serifs which have been designed
| carefully, to have distinctive characters. With good monitors,
| fonts with contrast between thin lines and thick lines, or even
| with concave lateral line edges, are preferable.
|
| I read this HN thread rendered in the Palatino Sans mentioned
| in TFA (with the italic of Palatino nova configured as its
| italic form; the italic of Palatino Sans Informal is also a
| good choice, but I prefer a stronger contrast between the
| regular and the italic variants of a font; that is why I have
| also configured the italic of Bauer Bodoni as the italic for
| Optima nova).
|
| While in TFA Palatino Sans was dismissed with regret, for
| having to be purchased, I have bought Palatino Sans, together
| with a few other high-quality typefaces and I use them on
| Linux, instead of the available free fonts. I consider the
| money that I have spent on good typefaces as some of the best
| purchasing decisions that I have made.
|
| However, for CLI/TUI applications and program editing, I use
| the free JetBrains Mono. Unlike for regular text, for
| programming there are many high-quality free fonts, but I
| prefer JetBrains Mono because it supports an extended character
| set, with many Unicode mathematical symbols that are missing in
| other programming fonts.
| lukeyoo wrote:
| it's merely about LaTex font: I use mostly the html-side font for
| text to spread the weight with markdown renderer. It is also made
| lighter by using KaTex(which comes with some compromise due to
| the lack of extension). I found this more coherent along with
| other types of articles. I'm very fine with non-serif. But this
| preference varies by person.
|
| Examples
|
| [1] https://lukeyoo.fyi/test/data/render/latex-statements-1.md
|
| [2] https://lukeyoo.fyi/recap/2025/5/statistical-inference-1
|
| [3] https://lukeyoo.fyi/test/data/render/latex-integrals-1.md
|
| [4] https://lukeyoo.fyi/test/data/render/latex-integrals-in-
| comp...
| libbrfish wrote:
| Does anyone have any suggested fonts for presentations in the
| beamer package?
| me_jumper wrote:
| I always use Atkinson Hyperlegible. It's great for the bad
| presentation situations one often is in (bad presenters, or
| simply the head of the person in front of you cutting off the
| lower part of the letters).
| jhanschoo wrote:
| I'm personally a fan of using EB Garamond together with Garamond-
| Math. Ironically I'm actually excited about Granjon's types in
| it. The Greek is very nice, though unconventional for someone not
| used to Greek texts. (The Greek letters will look unconventional
| when used as Greek letters for mathematical typesetting.)
| JustFinishedBSG wrote:
| Kind of expensive but there isn't actually a lot of choice for
| fonts with matching math fonts so for my PhD thesis I used Minion
| 3 + MinionMath.
|
| For mono fonts there are a lot of nice choices but I used
| PragmataPro for no other reasons that I own it and it provided a
| nice readable contrast.
|
| Otherwise for the free options, Palatino+mathpazo or StixTwoText
| + StixTwoMath are quite good options. Honestly anything but
| ComputerCM is a good option; it's imho not a very good font
| nowadays; it's way too thin. It was designed with the assumption
| that it will be printed on old, fairly bad, printers with
| significant ink overspill.
| PJ_Maybe wrote:
| MLModern [0] is a thicker version of Donald Knuth's Computer
| Modern and the Latin Modern project, in case you are
| interested.
|
| [0] https://ctan.org/pkg/mlmodern
| SubjectToChange wrote:
| +1 for Minion Pro, Minion Math, and PragmataPro. Those fonts
| have been my preferred defaults for years now. While they are
| expensive, it is worth it to write in a beautiful font and to
| compensate the artistic work that went into making them.
| trueismywork wrote:
| Does MinionMath cover all ligature? My reason for mostly
| sticking with Latin Modern Math is because I don't have to
| worry about random characters being unprintable.
| Aardwolf wrote:
| When I was at university I'd sometimes print PDFs as 4 pages per
| side to save paper, and some LaTeX-generated PDFs had some
| horrible font that became unreadably pixelated (while on screen
| it looked fine iirc). Not all PDFs were like this but LaTeX ones
| more often were culprits.
|
| HTML on the other hand printed perfect tiny vector glyphs, so it
| wasn't the printer's fault.
|
| I don't know enough now to find back what bad font this was back
| then, but I hope the situation is better today for LaTeX PDFs...
| jltsiren wrote:
| Printing is black magic.
|
| One common issue is that you don't have all fonts embedded in
| the PDF file. There may not be anything wrong on screen, as the
| PDF reader can probably find the font if you have it installed.
| But the software you use for printing may not know enough about
| the capabilities of your printer, and it may choose to hope for
| the best. If your printer does not have the font and the
| printing software didn't bother sending it with the document,
| the printer will use another font as a substitute.
___________________________________________________________________
(page generated 2025-05-21 23:02 UTC)