Hi Noah. First thing I notice is that all of your vertical strokes are too heavy. The -o- is too thin, as is the -c- and the bowl of the -d-. The -e- fills in on the left side. Perhaps instead of thinning the vert strokes you could thicken the thinnest areas of all of the letters. The exit curve on the -c- seems too thin. The bowl on the -b- seems too small, this could be because there isn't a solid base. Curves on the -m- seem too angular in comparison to the curve on the -n-. The bowl on the -p- seems to be squashed at the top right. The -c- also seems to be leaning backwards. What is your resource? From where is your inspiration coming?

Tiffany, attached is a minor revision and the original version in the same pdf. The quotation is in the revised version. The other revised sections are marked.
------
minorRevision.pdf
------

I drew the -a- free, it was the first character I made. I then made the -o- with the stress (?) at -15 degrees. The construction of the remaining characters was loosely influenced by bembo, various garamonds, and the -a- and -o-. Mostly I just made each glyph in relation to each other glyph, though I guess they aren't entirely harmonious yet.

Thank you for the prompt feedback! Just how thick should the thin areas become? Do you find what I"ve done sufficient? I'd like to retain a degree of contrast.
Yours,
Noah

Progress is slow, for those who are interested...is anyone interested, or am I just cluttering up the board with self-indulgent posting? ... Here is a sample of the font so far. I'm starting to feel out how I want the UC to be, and I've got about half of it done. What do you think of the new -n-, -m-, -s- ?
Thanks,
Noah
sample here

Hi Grod

As Tiffany noticed the 'o' and 'c' look too thin. But also the 'd' and the 'b' look like the curves are just a bit too thin. Try making them a little bit thicker than the stem width.
To me it also looks like the 'o' is tilting too much to the left. It could just be me but the 'd' having the same tilt as the 'o' also looks a bit like it wants to take off to the left.
I think you should try to stick to one type of serif on that 's'. I like the top one.
The right legs of 'n' 'h' and 'm' looks thinner than their left ones. And the hairline in those letters also looks very thin.
The serif on the 'q' is weird. It should be longer to the left than it is to the right.

I have probably repeated some of Tiffany's words here but I still think the contrast in certain letters needs work.

I look forward to see more of this font.

magnus

In the spirit of honest critiques, this looks to me like Bembo if Bembo were a lush! I'm afraid I simply don't understand the rationale for the modifications you've made to Bembo; them seem altogether arbitrary. As the others have pointed out, the colour is just plain wunky as a result. The relationships between thicks and thins across letters are not consistent. For example, the bowls of the p, d, b, and q are the similar, but not in the relevant way, viz stroke axis. More importantly, they don't match up well with those of a, m, n, h, and r, especially where they meet the stems.. Keep in mind that Bembo is of calligraphic heritage. You have kept the proportions from this face, while basically abandoning the calligraphic strokes and axes. I think you should probably either keep the calligraphic style, or change the proportions altogether if this design is to make sense.

magnus_gaarde, thanks for the additional feedback. I took your suggestions to heart and attempted to implement them. Overall I believe the font is improved, however I'm certain there are many more modifications I need to make and continued feedback is not only welcome but highly desired. Please note that the uppercase is complete but imperfect as of yet. I'd like to read any suggestions, criticism, praise, or objurgations from anyone willing to take the time to examine the samples and the energy to reply.
Yours,
Noah
Here is another sample

edit nepenthe, heh, figures. I post a reply to magnus_gaarde only to discover that, while I'd been composing my reply, you'd commented. First, thank you for the honesty of your crit. Now that you point these inconsistancies out to me many of them are painfully obvious. I think my main rationale is that I don't want to make Bembo, but I do want to be influenced by Bembo. Perhaps the randomness is due to fear of influence. Regardless, I can see I have a lot more work than I'd initially thought. Thanks for your observation, while somewhat frustrating, I find I mostly agree. Now to figure out what to do about it.

nepenthe, tell me how you feel about the font now. It is not meant to be Bembo, nor a revival, nor an homage, nor an interpretation. I drew on the proportions of Bembo, it is true, but I also drew on a variety of other sources and my own ideas. If the result is a drunken Bembo, then this is a failure. Here is another sample.
EDIT: 2 hours later
So I just wandered over to Kinko's and paid their absurd $.20/min to use a computer and nearly 1$/page to print a sample of frimbo on heavy weight, blindingly bright, laser printer paper. The Xerox printer had a maximum setting of 1200dpi, which I employed. The result, a crisp, sharp, very black (for a laser printer) printout. Looking at it now, I find that it is far from "wunky" and has an even, readable quality to it, at least at decent resolutions. While I'll continue to tweak, I think it rather extreme to label the frimbo with the epithet, "wunky."

I also tried printing your new sample from 1200dpi from my printer. I will try to post a detailed critique when I get the chance.

I eagerly await your opinion. In the meantime I've begun work on the italic. Still very rough, but at least all the characters are sketched out now. Italics sample
-- -- --
EDIT:
Here is a sample of the italics as they now stand. The overall spacing remains imperfect but it's getting there. The individual characters harmonize a bit better now, as well.

hmmm, lost interest?

Seriously, I'd really appreciate some feedback on the italic, both as an italic and as a companion to the roman. I'm still a complete novice when it comes to type but since no one has said "you suck, give it up" I'm encouraged to keep trying.
-Noah

I really don't have too much time, which is why I haven't commented, but since you're on the edge ;) ...

The Italic:

Stem widths aren't consistent. For one example, check the n against the r. The r looks noticeably thinner than the n. Also, the n is thicker on the left than it is on the right. If anything, it should probably be the other way around, but only by a couple of units.

You have a swash serif on the A in a place that doesn't make sense.

Some of the uppercase letters have both adnate and abrupt serifs. This doesn't make sense either.

The y looks bent out of shape.

The weights of the numerals seem fairly inconsistent to me. Problems here are similar to the LC.

It's good that you made the ascender serifs more horizontal than the exit serifs.

The leg on the R and the K should be more similar than they are now.

As this italic seems mostly influenced by Palatino and the roman is influenced more by Bembo, I'm not so sure they work well together. But perhaps if you even out the colour of the glyphs it will be OK. In terms of overall stems, italic and upright seem to match fairly well, but the italic is too loose.

I have to get back to work now, but good luck with it!

Interesting, thank you. I'm surprised at the Palatino resemblance, I've never looked at Palatino Italic, I'm pretty sure I've never used it for anything. Concerning the A, I tried it with swashes on both sides in sort of chancery style but it felt to calligraphic (well, duh) and I wanted the uppercase to be closer to a slanted roman than to a script -- Hmm, just took a look at Palatino Linotype Italic, I like the z. -- I'll work on regularizing some of the stem widths and cleaning up that y. I'm not certain what I'll do with the A. I don't feel constrained by history or convention and will happily mix styles and influences but if the result is crap I'm not going to stand back and call it art. I think I'll treat this face like a coherence set, as long as it is internally consistent it is true :-) (you are a philosophy major, right? hmm, but in Canada, you guys tend to be more interested in Being than epistemology)... thanks for the feedback, you've given me a lot to reconsider. Did you ever reach a conclusion re: wunky (isn't the word "wonky") color in the roman?

Interesting, thank you. I'm surprised at the Palatino resemblance, I've never looked at Palatino Italic, I'm pretty sure I've never used it for anything. Concerning the A, I tried it with swashes on both sides in sort of chancery style but it felt to calligraphic (well, duh) and I wanted the uppercase to be closer to a slanted roman than to a script...Hmm, just took a look at Palatino Linotype Italic, I like the z. I'll work on regularizing some of the stem widths and cleaning up that y. I'm not certain what I'll do with the A. I don't feel constrained by history or convention and will happily mix styles and influences but if the result is crap I'm not going to stand back and call it art. I think I'll treat this face like a coherence set, as long as it is internally consistent it is true :-) (you are a philosophy major, right? hmm, but in Canada, you guys tend to be more interested in Being than epistemology)... thanks for the feedback, you've given me a lot to reconsider. Did you ever reach a conclusion re: wunky (isn't the word "wonky") color in the roman?

This comment removed.

Sorry, that spacing is pretty rough.

hhp

Any examples of specific pairs that throw you?

this comment removed

Well, maybe InDesign could do better too - although from what I understand it depends on how good your base spacing is (and I haven't had the time to look). But one place that IKern is doing worse is in the word "such" (at the end of the third line before the end in the Roman) where it's leaving the "su" very tight but loosening the "uc" and "ch". It's also unduly loosening the "ub" in "dubious" on the last line. So maybe there's something about your "u" which is throwing it off? Dunno.

This isn't to say IKern is useless, just not nearly as good as a rigorous human spacer can manage. BTW, that overlay is cool-looking, but hard to "read".

hhp

on the u left sidebaring:25, right: 24; "su" pair -6 kern, uc pair -4 kern, ch pair no kerning. I really suck at spacing, even following Tracy's recommendations, that's one of the reasons I wanted to try iKern.

hrant, I've followed Tracy's instructions to the letter (sorry) and generated another sample. What is your opinion?

I'm not sure how much you're changed what*, but I think it does look better now.
Here are some things I noticed:

"a" right: too loose
"d" right: too loose
"s" right: too tight
"t" left: too tight
"u" right: too loose

* Especiall since there's a pretty severe limit to how far Tracy's method will take you; to me it's more of a -good- guideline.

The Indy setting looks a bit better overall I think, but:
1) It seems to be under-compensating for some of the 5 problems above;
2) Sometimes it's messing up, like in that "result".

hhp

and now?

Hey, that's much better! Good going. Just keep on tweaking though.
Like the left of the "a" is a hair tight... There's always something! :-/

hhp

I had a quick look at some of the roman pdf samples (using Acrobat reader 7.0 on Linux), but I have not printed any. Some of the uneven stroke widths in smaller sizes look like hinting artifacts to me. They do not appear to such an extent in the largest sizes. Zooming in alters the effect. I get a slightly different effect using gv on Linux, which uses the Ghostscript font rasteriser.

I have seen this effect in some badly hinted Type1 fonts. For example, omitting or misplacing a vertical stem hint can grossly thicken the stroke on screen, but it is OK on a high resolution print.

I wonder if you have some real (but fairly small) irregularity in the stroke widths, or some other problem, that is confusing the automatic hint generation.

Glyn Adgie
I think you're right. I don't know how to hint manually so I just use the autohinting routine. The results on one of my other fonts (kisbefe, a sans) are really bad. In this font I've noticed the r appears darker than it is at some magnifications. I've redrawn most of the characters, nothing major, just refinements, but the hinting remains less than perfect.
Hrant
If I've spaced the characters a la Tracy and the font is designed for between 6 and 14pts, on average what is the tightest kern I should make? I've looked at what Indy does and I've looked at the kerning in Minion Pro, and it seems that pairs are rarely kerned by -12 units or so. The problem I have is that what looks good at 128pts in the metrics window is too tight at 8pts on paper. How do you deal with the optical, um, illusion(?) of spacing?

Here is the revised version and kerning sample

> on average what is the tightest kern I should make?

Wow, that sounds tough... What does "average tightest" mean? :-/ Do you mean the max negative value you should use? I can't think how anybody could come up with a number... But -12 seems like way too little. I personally use values over 50 often, and I know I'm not alone.

Anyway, don't worry about the numbers too much, worry about the color. Once you arrive at a spacing that can't be improved without kerning, you're very close to the color you want, and you kern to bring errant pairs in line. This is sort of simplified, since the eventual avilability of kerning should be taken into account in the base spacing (and even the black letterforms), but in the real world it's 95% there.

> How do you deal with the optical, um, illusion(?) of spacing?

I think the only way must be by applying experience
to (via acquiring experience from) carefully crafted
tests you carry out.

But looking at your latest sample, I would say that your overall* spacing is slightly too loose for the vertical proportions and color of the face. And once you make the spacing slightly tighter, I think you can shoot for around 11 point setting - just great for text (although I personally prefer darker color). You won't really be able to go as low as 6 though; maybe 8.

* I say overall because it's still a bit too irregular - like look at "adjusted. But I think it's almost there.

hhp

Here is another kerning sample Please note, I decided to jump on the class kerning bandwagon and I'm still figuring things out. I thought I could throw A, smallcapA, and xheightA in the same class but that messes things up. Just an example. Anyway, Hrant, I've tightened the overall spacing and adjusted a bunch more, is it dark enough for you? Or is it too dark now? The last version took 17 complete lines with the word "memory" on the 18th line. This version requires 17 lines with room to spare. Optical kerning still manages to be slightly more compact. What do you think? And any hints on the class kerning? For example, what do I do with ligatures, what do I do with punctuation? Also, I'm very interested in your overall opinion of Frimbo as a face. Don't pull your punches, either.
Yours,
Noah

I think the overall spacing is a good tightness now, and the color is fine. However, it seems that maybe there are more spacing anomalies now. This is natural, since the more you tighten a face the harder it is to space... Maybe I should have told you that in advance. :-/

As for what I think of the design itself: I would make more of it more idiosyncratic; the distribution of funk seems somehow imbalanced. But this isn't designed for me, so don't fret.

hhp

Yet another attempt at getting the spacing right.

The spacing is looking pretty good now!

Some things that pop out:
- The right of the "e" is slightly tight.
- The left of the "a" is slighlty tight.
- I might tighten the right of the "r" a hair.
- Kern the round-round pairs looser; this is expected.
- The "fi" should be wider.

hhp

BTW, I keep forgetting to mention: your blank space could be a hair narrower.

hhp

OK, I've adjusted the metrics/kerning again (out of curiosity, when I've 'finished' with this -- assuming such a thing possible -- what do I do with it? Could I sell it? Whom would I approach? Or should I just keep it so in a few years when I finally managed to make a really good font I can look back and laugh?) and made another sample which you'll find Here. I've also set (very quickly [about 10 min] so no need to point out the lack of typographical refinement) a longer body of text at 11pts. It can be found here. As usual, I need all the help I can get. :-)

What have you changed? The "ea" in "heard" (end of first line) is still way too tight.

hhp

OK, I replaced the files linked to in my previous post. Changes should be evident now. BTW, ae isn't kerned, so I guess when I tightened the metrics I over did it a bit. What puzzles me is that I only increased the sidebarings on a few glyphs and added less than a hair of space between round/round pairs but the paragraph now appears to take the same or slightly more space than the one with no kerning at all.

- The spacing seems decent now. Just tighten the blank space.
- Maybe add some "r"-round negative kerns.
- The "ea" (what you meant) shouldn't need a kern anyway.
- You increased sidebearings on the "a" and "e", and considering how frequent those two letters are it's expected that they'd add a lot of set width to text. But you must have meant as compared to the non-kerned setting of the previous version, right?

Just to be clear though: this is all based on some short minutes looking at your limited sample texts. There might be -and most probably are- things we haven't caught yet; and I'd be shocked if the kerning is all -or even almost- finished*. But my New Year's resolution prevents me from taking this any further for free. :-)

* Whatever the hell that could mean in type design...

hhp

It seems I made an unfortunate mistake a few weeks ago when I posted a sample of Frimbo and claimed that the kerning was the result of Igino Marini's excellent kerning software iKern. It was not. This error arose not from malice but carelessness on my part. To perform the kerning Mr Marini requested the font be in TrueType format. Since I work with Type1 outlines I decided to export the metrics and kerning from the file Mr Marini returned to me and import them into the original postscript formatted file. Somewhere along the line I made a mistake. I'm still trying to determine exactly what I failed to do, regardless the error is mine entirely and iKern did not produce the results that I claimed it had. I am deeply sorry for this mistake. I know how hard Mr Marini has worked on this project and the results, the true results, of his efforts are breathtaking, one need only examine the fonts he offers free of charge on his own site to see power of his program. It is my sincere hope that my accidental misrepresentation of iKern has not tainted anybody's perception of this incredible software. The fault is entirely mine, the rough spacing mistakenly attributed to iKern came from my own incompetence. I only hope that Mr Marini and, indeed, the members of this community can forgive my error.
Sincerely,
Noah Feldman

I like how your caps turned out! They are really nice.

Overall, I still think this font is bit rumpled, if not quite "wunky" as I misspelled earlier. But perhaps that is what effects its unique character! For persons familar with different typefaces, it is interesting to see the different influences, e.g. a, f and r are so Bembo, while s is so Palatino. (I am actually surprised that you say the italic has nothing to do with Palatino, since there is such a strong resemblance.) And although these elements, among others, are familiar, the overall effect is so unusual that I really don't know how to judge it! Anyway, congratulations on sticking with it and trusting your own taste. I look forward to see more of your work!