Marvin
Visions
EXPANDS
& BECOMES
A VARIABLE
FONT WITH
2 OPTICAL
SIZES & A
RANGE OF
WEIGHTS

During the research phase Kabel, Serif Gothic and Avant Garde quickly proved to be the typographical staples of the genre (together with a collection of more unusual ones). To a degree these typefaces, along with the often lavish illustrations, have given science fiction its look in the collective mind. To reach a new audience, SF needs a new image. The well established one is unfortunately also associated to a lot of clichés that put off some otherwise interested people. The quest for a typeface that could tread the line between a modern look and a retro feel had begun.

Ideally, you’d want a classic with a twist. Something that’s a bit like Kabel, sorta like Serif Gothic, and with the impact of Avant Garde. Marvin doesn’t really tick all these boxes but it somehow captured my imagination as soon as I spotted it. I felt a hint of science fiction about it, only realising later that it had been used for that purpose. But more importantly, I could sense something mysterious, and maybe even comical, hiding inside it. Unfortunately there was no proper digital version available. Everything out there was poor and somehow managed to suck the life out of it. That’s why I decided to look for what made it magical to my eyes.

Marvin gets its distinctive voice not only from its Art Nouveau vibe but also from its almost geometrically perfect construction. Its roundness and familiarity with Bauhaus typefaces shows its roots in geometric sans serifs at the same time. Typefaces inspired by the Art Nouveau lettering are usually built with a brush stroke — a handmade feel that could transport you to Paris. The original Marvin doesn’t care much for brush strokes. It’s built like a tank with fat stems, little to no optical correction and a C that’s simply half of the letter O. Marvin doesn’t have the time to dilly‑dally.

I only managed to find a handful of examples in use but the vast majority uses the Lubalin-esque tight-not-touching style. Maybe to support that trend, Marvin had generally narrow, plump, letters – a structural decision that can help to reduce the white space between the shapes. By going away from Roman proportions and with the limitations of a seemingly modular typeface constantly re-using bits of itself to build itself, Marvin had a few letters that felt awkward. Whether it’s because of the rigidity of the construction or not, letters like P, R, C, K, Y were much narrower than the rest, producing an uneven rhythm within words. After closer inspection, the curve on the A and V (which looks like it’s simply been flipped) leans in, flattening the shape as it curves. Creating more white space above the curve felt inconsitent with the intent of use with tight spacing.

The first things that Marvin shows you – and they’re in the name – are its A, N and V. My instinct was that the magic was hiding in there and it was crucial to get these letters right. As soon as my first printed specimens came out, I started to see issues. The words had no balance and some letters, like the tower of Pisa, were about to topple.

I had a different image of Marvin in my head. For me, it had strong and sturdy shapes, sure on their feet. The narrowness and the flatness of the curves reduced the counters when my idea of Marvin was more open and maybe more ‘Roman’. I particularly wanted larger counter to increase its readability and friendliness so it could be used in different contexts. As a result, Marvin Visions is based on the initial impression that Marvin made on me rather than a true reproduction of its lettershapes.

With over 42 different versions, the A was one of the most time consuming glyph. The task was to find a curve that could fit A, V and N whilst maintaining the impression of visual harmony. They have a wider, more pronounced curve that feels smooth throughout, arching over the stem. It creates space inside the letter and closes the gap you’d have next to shapes like T. Obviously the curve couldn’t be transplanted to the V and N without being adjusted, paying particular attention to the joints and shoulders.

When I started to work on the O, I was surprised by its simplicity. This was a round, unapologetically geometrical shape. Wide, with a generous counter, to the point that any idea of an overshoot had been forsaken. A perfect circle inside another, the pinnacle of geometric sans-serifs.

However, there are very good reasons why circular shapes need special treatment in terms of design and I couldn’t dismiss them completely. I tried to keep the outside circle as perfect as possible, in line with the original intent, and focused on the counter to compensate the optical illusions. The resulting shape should still look very much like a perfect circle with a more mathematical approach to the solutions offered by the likes of Futura and Johnston.

There’s only so long you can look at the word COG without going mad. The rounded shapes were redesigned together with the idea that the O could overshoot minimally but the G and C should rest on the baseline and meet the cap height like other letters to keep visual harmony and consistency. Because of the white space either side of the O, the overshoot is not really noticeable but helps to lift it to the level of the other letters. When it came to the C, whichever way you cut it, the half-circle had a tendency to be troublesome: it looked squashed by two towering forces when used in the middle of a word, breaking the rhythm. Following my idealised Marvin, I extended it a fraction in order to give it more room to breathe next to letters with a straight stem.

If you were to draw an R, would you just tack a leg onto a P? Maybe. They share some of the same DNA after all, but it doesn’t mean they have to be clones. If you don’t consider what you want the R to look like first, you increase the chances of the leg sticking out like a sore thumb.

Marvin suffered from that because that’s exactly how it had been built. The resulting shape was definitely too narrow to accomodate the leg and looked strange at the start or end of a word. To remedy this, I sketched the proportions of the R first and built the P based on these. I then went back to the R to finish the job, taking care of adjusting the curve to account for the thickness of the joint.

It only took a few weeks for me to get a full alphabet, but I could see and feel that something wasn’t quite right. I needed to cement my vision of Marvin to get it looking the way I imagined it. Examples in use that could guide me were hard to find, particularly because I didn’t know where to look. I decided to let the project sit for a while.

On holiday in Ukraine I found this box for a model boat. I didn’t care for the inside as much as I did for the outside. This was the first time I saw Marvin in the wild and it was in Cyrillic. I still haven’t resolved how or why or even who but the fact is: someone made a Cyrillic version of Marvin and it was used to sell model boats. Note the inverted V to make the ‘El’, continuing what had been done for the A, V and N.

I spotted Marvin a second time on that trip, in giant letters on the side of a lorry. Ritter Trans, an Austrian transport company uses Marvin as its logotype. Again, I have no idea how they maintained this in the digital age but here it is. When I came back, I was ready to tackle the project again. Based on the box and research on the alphabet, I drew a set of Cyrillic glyphs. It came out spikier than latin Marvin but also somehow cooler and prompted me to start looking the latin alphabet again.

Thanks to Toshi Omagari’s expert advice, the general look and feel of the diacritics has improved dramatically. One of them is particularly worth highlighting: the ogonek. If a glyph has a brush stroke in its DNA it’s this one and it doesn’t really jam with steamrolling Marvin. Luckily, research showed they can be more graphic in display faces, sometimes reduced to a square or rectangle. It was an opportunity to bring out the curve from the A again, reducing the ogonek to its simplest movement.

Finally, The german double S (know as eszett) was an interesting one to draw. Until very recently there wasn’t an official capital version of the uppercase ß in a single glyph. It’s not a new letter and there are extensive articles on the history of it. The debate is still happening it seems, but the fact is, ẞ is now part of the German letterset, so Marvin Visions supports it.

The specimens I could find only had a handful of punctuation marks but I had no luck finding them in use once again. The comma, in particular, was a problem. It looked great on its own, but as part of a sentence it looked too much like a full stop. It was so round it couldn’t fulfil its role. I particularly wanted a flick in the tail, feeling that a slightly diagonal stroke was important to the semantic meaning of it. The one that made the grade was derived from the original, retaining its roundness with an exaggerated tail. The same comma worked perfectly as a quote mark. For the comma accent, it would have been easy to keep using that new design but it was reaching too far down. Instead I went back to the initial comma and there it looked right at home.

The ‘at sign’ – or as they say in Romanian, the monkey tail – was an interesting challenge. Usually the lowercase ‘a’ would be your starting point, but Marvin being uppercase only I had to come up with something different. The final design embraces the circular nature of the glyph and Marvin’s affinity for perfect circles. Looking forward to seeing that on a business card!

The figures were a different kettle of fish and had to be redrawn entirely. The 2 and 3 had a great look to them but didn’t feel balanced. The 1 looked like a capital I, the 5 needed a serious face-lift, the 4 had some thickness problems. The rest was workable but needed correcting. It became an optical game to fit in the thickness necessary for the strokes but keeping the width consistent and the counters large enough.

Marvin is not naturally suited for this kind of job, like an elephant in a china shop, it tramples happily any subtelty and detail. However, Marvin is a geometric sans at its core, and despite its Art Deco influence, it looks surprisingly normal if you don’t use letters like V/A/N (without looking neutral, or ever having that pretension).

As soon as the proportions of troublesome glyphs were fixed, I started on spacing. I didn’t want to make the same mistake I had made before and start spacing too late. Because the main use was still display, I wanted to keep the tight-not-touching spacing I had for the bold weight but the spacing still had to be a little looser to accommodate the far larger counters.

When the first proofs were out the first reactions were ‘This is nice but, what if it was even lighter? Like, super thin?’. These were comments by designers — type users — and not type designers which made it all the more relevant. If a designer sees how they could use a specific weight of a font, sees something that they like in a regular weight but can visualise a thinner one working very well for something, it’s worth listening. I followed up very quickly with an extrapolated, auto generated, thin weight. Although it was all over the place, it was a great way to see how things might look slimmed down to the extreme with low effort. It got the thumbs up.

The problems I had encountered with the regular weight were only exacerbated. The same letters looked completely squished against towering others like O which pushed aside anything and everything out of sheer circularity. The spacing had to be much wider to avoid words looking like a series of barcodes on the page. The number 3 needed a completely different construction and the whole alphabet simply started to look lanky.

The challenge had just gotten harder. Thin sans were very trendy a few years back and a lot of families have extremely thin weights. But when it came to capturing the same 60’s Marvin feel, with a touch of the Art Deco, thin examples were much scarcer. I did find a few examples that I could look at but most of them dealt with proportions in extreme ways. Either everything was extended or they stuck more to the Art Deco style and ran with letters with massive differences in width. In the meantime I had also acquired the full Face Photosetting catalogue and could draw inspiration from other faces designed by Michael Chave and his team.

This time, the O had to forego a little of its perfection and accept to become more oval, if only for the sake of its siblings. The K and the R have their legs stick out much further, when they’re traditionally very narrow in Art Deco lettering. The C is properly extended to compensate its half circle nature. The U is wider, as are a lot of the letters, to compensate the loss of fat. If you lose weight, you gotta build muscle to replace.

Because of the tight-not-touching spacing, the side bearings work differently that they normally would. There is less compensation to apply for open counters because you set your space based on whether the sides of the letters are touching. As a result, whenever the new display weight were tracked to be used smaller, the spacing didn’t look elegant enough at small sizes. As the space between the letters increases, the need to optically adjust the spacing also increases. For instance, in the bold weight, you get 15 units of space on either side of the H but also on either side of the F, which would traditionally be a lot smaller to the right. But as the distance between the letters increases, the space on the right side of the F might actually get smaller because of its very open nature.

Feeling like I needed to learn more about what designing for small sizes entailed before I jumped back, I got hold of the very useful Size-specific adjustments to type designs. The book takes a very exhaustive approach to the different things that can affect how we perceive letters. Essentially the very things that made the display weights of Marvin interesting were also detrimental to their use in text size. The letters needed to be squatter, shorter and wider, and the weight progression needed to be less extreme. The thin weights needed to be bolder and the bold weights needed to be less bold. The three display masters were redrawn to create three small masters. They’re a little bit shorter, wider, and even larger counters. The diacritics have been repositioned and redrawn to match the new looser spacing adapted to small sizes.

For the same reasons, I drew an alternate round top A that would distribute space around itself a bit better. The slopping A created too much space at the top and while it still works, the round top A sometimes look more comfortable. There’s also an alternate V and slopping A for both optical sizes. These had already appeared in the original version thanks to a simple flip, for instance in the logo for the Teatre Victoria in Barcelona.