.jpg)
There is a number that every camera company in Japan chased through the 1950s, and it was not a sales figure. It was an f-number, and the point of it was to be smaller than everyone else’s. F/1.5 had been the state of the art since Ludwig Bertele drew the Sonnar for Zeiss in 1932. By 1953 a firm of forty people in Nakano, Tokyo, had got to f/1.1. By 1961 Canon had got to f/0.95 and was telling the world the lens was four times faster than the human eye. In between, four companies — Teikoku Kōgaku, which became Zunow; Fuji Photo Film; Konishiroku, which became Konica; and Nippon Kōgaku, which became Nikon — put out four standard lenses faster than f/1.2 in the space of thirty months, and Canon, which had started the decade copying Zeiss, ended it with the lens that closed the argument.
This is the story of that race as the Japanese sources tell it, because the English ones mostly do not. It begins in the navy.
A lens for dusk, 1943
In 1930, in the trough of the Depression, a man named Suzuki Sakuta (鈴木作太) founded the Teikoku Kōgaku Kenkyūjo — the Imperial Optical Research Institute — in Tokyo. His technical director was Hamano Michisaburō (浜野道三郎), and Hamano had come from Nippon Kōgaku, which is the first of many times in this story that a man walks out of Nikon’s door carrying what he learned there. Teikoku ground lenses; it did not make cameras. In 1943, with the war going badly, the Imperial Navy asked it to research a 50mm lens at f/1.1 for aerial reconnaissance at twilight — a lens to see by when there was almost nothing to see by. Japanese Wikipedia, citing the standard histories, has the request; MAP Camera, telling the same story from the other end, says Suzuki carried the experience of an experimental f/1.2 into the postwar work.
The war ended before the lens did. A prototype of twelve elements was working by 1950. Everything the company had made and written down during the war burned, MAP Camera says, and the postwar lens was started again from what the men remembered. A line that circulates in the collectors’ literature has Hamano making fifty thousand drawings over ten years; I could not find it on any page I read, and I give it here only as a thing people say. The decade is right.
Nippon Kōgaku spent those same years in a different kind of trouble. Its optical designers had built the navy’s rangefinders and periscopes; in August 1945 the company had to make something civilians would buy. Nikon’s own history of its lenses, the Nikkor Thousand and One Nights essays written by its designers, tells it through one man: Murakami Saburō (村上三郎), who between 1945 and 1949 designed six Leica-mount lenses almost alone, redesigning around whatever glass the works still had — “ガラスの在庫がなくなりそうです,” the glass is about to run out — until in 1948 the glass plant steadied and the designs could be finished. “ようやくこれまでの努力が報われる時がきた”: finally the time had come when the effort would pay. Those six lenses, the essay says, kept the company alive while the Nikon I and the Nikon S found their feet.
Two of them matter here. The Nikkor 5cm f/2 of 1946 and the Nikkor 5cm f/1.5 of 1949 were Sonnars — the f/1.5 a close study of Bertele’s, seven elements in three groups. Nippon Kōgaku was not ashamed of that, and had no reason to be; the Sonnar was the best fast lens in the world and every Japanese maker started from it. What Nikon did next was the thing that mattered: in late 1950 it replaced the f/1.5 with its own 5cm f/1.4, still a Sonnar type but its own computation, and that is the lens David Douglas Duncan’s Korean War pictures made famous. The New York Times told American photographers in December 1950 that Japanese lenses had outperformed German glass in Duncan’s hands, and from that morning every fast lens made in Tokyo had a market waiting for it in New York.
Canon starts by copying, and one man refuses
Canon’s position in 1950 is the one the company does not put on its history page. Its museum says it: the Serenar 50mm f/1.5 of 1952 “is developed based on Zonner 50mm f/1.5 manufactured by Carl Zeiss,” because “production of a large aperture standard lens was considered as a status symbol of a high class camera manufacturer at the time.” The Japanese Wikipedia biography of Itō Hiroshi (伊藤宏) says how Canon had been going about it: it had been hiring engineers away from Nippon Kōgaku and copying German products. Itō, who joined in the spring of 1948 straight from university — he had wanted to design aircraft, and the end of the war had ended that — decided that a latecomer copying the same lenses would never catch up. He got hold of the German patent publications, Bertele’s among them, and studied the theory instead.
The result, in June 1951, was Canon’s 50mm f/1.8: six elements in four groups, a Gauss type, the design that Zeiss’s Planar and Leitz’s Summicron were also built on and that everyone knew flared badly wide open. Itō’s contribution was a way of analysing the aberrations that let him tame the coma — the smeared, comet-shaped highlights that had made fast Gauss lenses unusable at full aperture — to the standard of the Sonnar 5cm f/2. He patented it in Germany as well as Japan. Frank Mechelhoff, the German collector, calls it “a fine, quite fast lens fully comparable with the best quality of the world, LEICA and ZEISS,” and Canon built more than a hundred thousand. Itō was twenty-six. In 1980 the Japanese state gave him the Purple Ribbon for it.
This is the hinge of the whole decade, and it is worth stating plainly. The Sonnar is a lens for a world without coatings: it has few air-to-glass surfaces, so it holds contrast, and it pays for that with an asymmetry that cannot be pushed much past f/1.5 without the corners falling apart. The Gauss type is symmetrical, correctable, and hungry for surfaces; once anti-reflection coatings existed — and they did, by 1950 — and once the new high-index glass arrived, it could be driven to apertures the Sonnar could never reach. Every lens that follows in this story is a fight between those two families, and the companies chose sides.
Zunow fires first, 1953
On 20 October 1953 Teikoku Kōgaku announced the Zunow 5cm f/1.1: the twelve-element prototype refined to nine elements in five groups, sold in Leica thread and in the Nikon and Contax bayonets. It was the fastest lens ever offered for a 35mm camera, from a company nobody outside Japan had heard of, and it was a Sonnar — the design pushed, by main force, a full stop past where Bertele had left it. It was launched under the slogan “the brightest lens in the world” — MAP Camera has the copy — and the rear element bulged about a centimetre out of the back of the barrel to earn it, which is why the first version is called the ping-pong ball, and why, MAP says, it was disliked as hard to use. MAP Camera’s staff, shooting an early example, found rainbow ghosts that came and went “as capriciously as a cat,” corners that stayed dark long after stopping down, and a wide-open image like a soft-focus filter. They also found it beautiful, which is the only reason anyone pays for one now.
The lens did not sell, and the price says why: ¥95,000, MAP Camera records, when a salaried man earned about ¥10,000 a month — the better part of a year’s pay for a lens, nearly twice what Canon would ask for the f/0.95 eight years later. The photo-china column that tells the company’s story puts it in one line: it seems hardly to have sold at all, which is why it is so rare and so expensive. Teikoku incorporated in 1954 and put out a 5cm f/1.3 beside the f/1.1, and by then the other three had answered.
The new glass and the panic, 1954
Fuji Photo Film, a film company with an optical department, had the answer that changed the rules. Its Fujinon 5cm f/1.2 was in the June 1954 Asahi Camera as a new-product memo and on sale that November, and it was the first Japanese lens built on the new rare-earth glass — lanthanum glass, high in refractive index and low in dispersion, which let a designer bend light harder with less aberration. The designer was Doi Ryōichi (土居良一), who wrote his own account in Shashin Kōgyō in 1955: a lineage from his Solinon 50mm f/1.5 patent through an f/1.3 study to a nine-element trial and an eight-element production lens, six of whose elements — a reader of the Nikon blog that reproduces the memo counts them — were of four kinds of the new glass. Doi described the finished formula as readable either way, an improved Sonnar or a modified Gauss, and the Japanese literature has argued about which ever since.
What the new glass did to Zunow is the drama of the middle of the decade. The Fujinon was a tenth of a stop slower and, by the accounts, plainly better, and it was made by a company with a factory. Teikoku’s response was to redesign its flagship from the ground up, and the two men who did it tell you what kind of company it had become: Kunitomo Kenji (国友健司), from Nippon Kōgaku, and 藤陵厳達, from Yashū Kōgaku. The 1955 Type II Zunow is eight elements in five groups, three of them in the new glass, with a flat rear and the ping-pong ball gone — a second Nikon man’s work, twenty-five years after Hamano’s — and MAP Camera says it was well received, which at that price did not help.
Konishiroku, the oldest photographic company in Japan, showed its entry as a development at three Tokyo department stores in October 1954 and sold it from December 1955: the Hexanon 60mm f/1.2. Nobody has explained the focal length. Sixty millimetres needed its own finder and had no rangefinder frame anywhere; it may have been the only way the designer could reach f/1.2 with the glass and the back-focus he had, and the designer’s name is not recorded in any source I could find, Japanese or English. Two collectors’ sites put the run at about two hundred or at most three hundred. It is the phantom of the four, and it went on sale two months before the lens that ended the first round.
Nikon answers, deliberately not with a Sonnar, 1956
The Nikkor-N 5cm f/1.1 appeared in February 1956, and Nikon’s own essay on it is unusually candid about where it stood: “F1.2より明るい超大口径レンズは、昭和28(1953)年発売の「ズノー 5cm F1.1」に次いで二番目の登場でした” — among lenses faster than f/1.2 it was the second to appear, after the Zunow of 1953. Not the world’s first, not the world’s best, the essay says; second. What it was, was a refusal. Murakami Saburō, now head of the design department’s third mathematics section, did not push the Sonnar as Zunow had. He built a nine-element, six-group Gauss type and hung its speed on three convex elements of the new lanthanum glass, and he did the computation, the essay says, with an abacus and a table of logarithms — “そろばんと対数表です” — over two years of redesign and prototype. The patent was filed in 1957 and granted in America in 1958.
The lens weighed so much that the first version, which used the Nikon S’s internal bayonet and let the body’s own helical carry it, went back to the factory for its mounts to be adjusted; from 1959 an external-mount version carried its own helical. Eight hundred and thirty-five of the first, fifteen hundred of the second, by Rotoloni’s count of the factory records. Nikon’s essayist finds coma and vignetting wide open and a spiral to the blur that he calls a flaw, and excellent centre sharpness by f/2.8 — the Gauss’s virtues and vices, exactly as Itō had found them in 1951 at f/1.8.
And it is the ancestor of something. The 5.8cm f/1.4 that Murakami designed for the Nikon F in 1960 is the next lens in the line, and the essay on that one explains why every reflex camera killed the Sonnar: an asymmetrical design cannot leave room behind it for a swinging mirror, and the Sonnar, which cures its aberrations by “fighting poison with poison,” falls apart if you try to move it. The Noct-Nikkor 58mm f/1.2 of 1977 descends from the 5.8cm, and so, at one remove, from the abacus.
Canon, twice
Canon’s 50mm f/1.2 of September 1956 is the lens that gets misdescribed. It is often written — Mechelhoff writes it — that it was the fastest lens for 35mm film in the world when it appeared. It was not: the Zunow had been on sale for three years and the Nikkor for seven months. It was the fastest Canon, and the fastest lens of the Planar type, seven elements in five groups by Itō, at ¥60,000 — more than a camera. It replaced the Sonnar-copy f/1.5 at the head of Canon’s range, and it is the least loved of the lenses in this story now: soft to f/2, glowing, flaring, and, in BastianK’s phrase, “by modern standards simply not a well performing lens.” That is not the point of it. The point was that Canon had stopped copying Zeiss and could now be measured against Nikon on its own designs, and that a company which had started the decade poaching Nippon Kōgaku’s engineers had, in Itō, produced one of its own.
The 85s ran the same race a length behind. Nippon Kōgaku’s 8.5cm f/1.5 was begun by Murakami and finished by Wakimoto Zenji (脇本善司) through eight redrawings and seven trial lenses between 1950 and 1953, a Sonnar with a sixteen-bladed iris that Domon Ken and Kimura Ihei both carried — Kimura so devotedly it was called his trusted blade. Canon’s 85mm f/1.5, Itō’s, was on sale in June 1952, in a chrome barrel weighing 730 grams. The wide-angles had their own race, and Fuji started that one too: the Fujinon 3.5cm f/2, sold beside the 5cm f/1.2 in 1954, was the fastest 35mm lens in the world when Zeiss’s best was f/2.8 and Leitz’s f/3.5, and koiwide, which has the only full account of it, says it was the spark. Nikon’s 3.5cm f/1.8 came two years later and took the title, Zunow’s 3.5cm f/1.7 the same year, Canon’s 35mm f/1.8 in 1957 and f/1.5 in 1958 — each a tenth of a stop past the last, from lenses small enough to lose in a coat pocket. A Japanese collector, shooting the Zunow, the Fujinon, the Canon f/1.2 and the two f/1.4s side by side on a Leica M9 half a century later, concluded that the best balanced of them all was the Canon 50mm f/1.4 — the lens that was never the fastest at anything.
Zunow dies and Canon wins, 1958 to 1961
Zunow, renamed in December 1956, had by then decided that a lens company could not survive on lenses. Suzuki set up a subsidiary, Shinkō Seiki, to build the camera that its designer Arao Kiyoshi (荒尾清), who had worked on Minolta’s development, wanted: a single-lens reflex with a pentaprism, an instant-return mirror, a hinged back, a single non-rotating shutter dial in even steps, and the first fully automatic diaphragm in the world, sold with a 58mm f/1.2 — the brightest standard lens then on sale for a reflex — and a 100mm f/2. The Zunow Pentaflex went on sale in April 1958, a year before the Nikon F, with every feature the F would have. A reader of the Nikon blog that reproduces the Shashin Kōgyō mechanism article of July 1958 remembered it from his schooldays in Nakano: a rich boy brought one to class, and its form was overwhelming. He remembered the red flags going up at the factory, too.
Nearly every part was made outside and only assembled in-house; the parts did not fit the tolerances the drawings assumed, the yield was bad, the cameras broke, and Japanese Wikipedia puts the total under five hundred with about two hundred sold. Shinkō Seiki was wound up at the end of 1959 and its tooling destroyed. On 1 January 1961, when its main lens customer, Arco Shashin Kōgyō, went under and the camera maker Neoca ran into the wall behind it, Zunow Kōgaku folded and Yashica bought what was left. The company that had fired the first shot of the race, eight years earlier, did not see the last one.
That came in August 1961. Canon’s 50mm f/0.95, designed by Mukai Jirō (向井二郎) and announced at Photokina the year before, was sold as a standard lens for the new Canon 7, on a three-lug bayonet outside the screw thread because no thread could carry it. Canon’s museum says it plainly: “the largest aperture in the world for photographic lenses available in the market at the time,” and “four times faster than the human eye.” It cost ¥57,000 when a civil servant’s monthly pay was about ¥12,000 — five months’ salary, and you needed the Canon 7 to mount it. Wide open it is soft, glows, vignettes by four stops and flares at everything, and either you love it for that or you don’t. It is the Dream Lens because that is what it was called, and because the race, from a navy request in 1943 to a lens faster than the eye in 1961, had become a dream by then. Leitz took five more years to answer it, with the Noctilux f/1.2 of 1966, and never went faster than f/0.95 at all.
What it was for
Ask what the race produced and the honest answer is: a great many lenses that are worse at f/1.1 than at f/2.8, sold to people who wanted to own the number. The Zunow hardly sold; the Hexanon is a rumour; the Nikkor strained the camera it was mounted on; the Canon f/1.2 is soft. The lens the whole decade was actually judged by, in the end, was Itō’s f/1.8 of 1951, which was not fast and was made by the hundred thousand.
But that is what races are for. The abacus computation that Murakami did for the f/1.1 became the Gauss lenses of the Nikon F; the coma analysis Itō did for the f/1.8 became every Canon lens after it; the rare-earth glass Doi risked on the Fujinon became the glass everyone used. Zunow’s engineers, ex-Nikon men and all, went to Yashica. Fifty years later a collector in Japan shot the five of them side by side on a digital Leica and found them, wide open, quite capable for portraits. One tenth of a stop at a time, and it took eighteen years.