One of the most important films in the history of cinema owes its form to a production disaster — but the connection between them is more complicated than people usually think.
1976. Andrei Tarkovsky takes on an adaptation of Arkady and Boris Strugatsky’s novel “Roadside Picnic.” The screenplay is written in close collaboration — the Strugatskys rework the text again and again, adjusting to the director’s vision, which has not yet fully settled into shape itself. Location shooting was initially planned in Tajikistan: a mountain landscape, alien, fractured, carrying the feeling of post-apocalyptic emptiness. But the plan literally collapsed — an earthquake in early 1977 made the location unusable. Production moved to Tallinn, Estonia, where suitable industrial ruins and abandoned power stations were found. Even at this stage, the film began to veer away from its original conception: the texture of Central Asia and that of Estonia are two different languages, two different skies, a different humidity in the air and a different color of concrete.
The early screenplay versions really were unlike what viewers saw in 1979. They had more action, more plot twists, more of the science-fiction narrative in the spirit of the novel. But it would be a crude mistake to imagine that Tarkovsky had intended to make a genre sci-fi action picture, then accidentally “discovered” meditative cinema. From the very beginning, the director gravitated toward slowness, toward a metaphysical register — it is enough to remember “Mirror,” released a year earlier. It was simply that the direction in which the film was moving had not yet been carried to the radical point it would reach after the catastrophe.
The cinematographer of the first version was Georgy Rerberg — a brilliant professional who had shot “Mirror” for Tarkovsky. Their partnership seemed tested and dependable. Filming in Tallinn went on in 1977, and enough footage had accumulated for its loss to become a real blow.
The film stock was not Soviet. They used Kodak 5247 — imported negative stock supplied by producer Sergio Gambarov. In Soviet film production, this was a rare and valuable resource: Western film stock offered different color rendition, different latitude, different possibilities for the cinematographer. Once it was developed, it became clear that a substantial part of the footage was unusable — the image had acquired a characteristic dark-green cast, killing both color and detail.
What exactly happened remains a question with no single answer. The theories include a technical error by the lab during development — the wrong chemical process for that type of stock — a defect in the Kodak batch itself, substituted film stock, even sabotage. The participants’ accounts differ in their details: some spoke of the footage being completely ruined, others of only partial unusability. What gets me every time is how differently people who stood beside one another on set describe the same event — as if each had seen their own defect on their own film stock.
The common legend that “all the film was lost” is a simplification. By some accounts, part of the footage survived and even existed for a while as rough working fragments. But for a director striving for absolute visual control, partially damaged material could be equivalent to material lost altogether — it was impossible to splice it into a coherent film.
The technical accident coincided — and “coincided” sounds far too mild — with the escalating conflict between Tarkovsky and Rerberg. Their relationship had deteriorated throughout the shoot. The director was moving ever further away from their original agreements on the visual approach, demanding something else, changing himself as he went. Rerberg, a man with his own sharp eye and professional pride, was not the kind of cinematographer who silently carried out instructions.
The conflict was creative and personal at once — in cinema, such things cannot be separated, because a film crew lives together for months in a confined space, under the pressure of deadlines and budgets. After the defective film stock was discovered, the situation became unbearable. Rerberg left the project. Who “fired” whom, or who “left of his own accord,” is another point where the memoirs diverge. But the outcome is unequivocal: the cinematographer who had defined the visual language of “Mirror” never worked with Tarkovsky again.
At the same time, the Strugatskys continued revising the screenplay. In essence, by the collapse of the first shooting cycle, the text had already moved far away from “Roadside Picnic” — the Zone was turning from a science-fiction premise into a metaphysical space, the Stalker from an adventurous guide into a holy fool, and plot mechanics were yielding to an inner journey. The Strugatskys, to give them their due, accepted these rewrites, although every new version led them further from their own prose. Different figures are cited for exactly how many screenplay drafts were written, but we are talking about many iterations, each one a full-scale reworking rather than a cosmetic edit.
In April 1978, Tarkovsky suffered a heart attack. For a man who had just lost an enormous amount of work, was in conflict with his closest collaborator, and was trying to reassemble the film from scratch, it was a predictable physical response to overload. Production stopped.
When filming resumed in June 1978, it was, in essence, a different film. The new cinematographer was Alexander Knyazhinsky — a professional of an entirely different type, less of a “star” than Rerberg, but ready to work in a state of complete submission to the director’s vision. And it was in this partnership that the visual logic of “Stalker” as we know it was born.
Knyazhinsky and Tarkovsky built a system in which the image functions as a tuning fork for inner states. The world beyond the Zone is monochrome, faded, sepia-toned, like an old newspaper photograph. The Zone is muted color: green, water, moss, rust. The transition between them is not abrupt but gradual, the way the eye adjusts to darkness. The contrast could have been a poster-like device — “gray reality versus colorful miracle” — but Tarkovsky made it subtler: the color of the Zone is not bright, not festive. It is more sickly, damp, like the color of mold on walls.
The long takes of “Stalker” are not a stylistic whim, nor are they the result of a shortage of editing material. Knyazhinsky’s camera moves slowly, with almost physical effort, like a person wading through water. And there is so much water in the film that it becomes almost the main character: puddles, streams, flooded rooms, rain, condensation on metal. Water in “Stalker” is the medium in which time is dissolved. The famous tracking shot over objects lying on the bottom — a syringe, coins, an icon, a pistol — is not symbolism spelled out in capital letters, but the archaeology of an instant, frozen and submerged.
Tallinn’s industrial ruins proved to be the perfect texture. Abandoned power stations, concrete structures, rusted-through pipes — all of it looked not like the set of a science-fiction film, but like a reality from which function had departed while matter remained. The Zone in “Stalker” does not look alien — it looks like our world with the meaning removed. I think it is precisely this ordinariness that makes the film so unsettling: nothing supernatural happens on screen, yet every frame is saturated with the sense that the laws here are different.
The film’s rhythm is its most radical element. The final version of “Stalker” has around 142 shots in the entire film, over a running time of more than two and a half hours. By comparison, the average Hollywood film of the same period contained hundreds of cuts. Every shot lasts so long that the viewer stops waiting for the next one and begins to live inside the current one — studying the texture of a wall, listening to droplets, following the movement of fog. Time ceases to be a frame and becomes the medium in which the image exists.
What remained of the first shooting cycle? Some fragments, according to accounts, were kept by editor Lyudmila Feiginova, who worked on the film. In the mid-nineties, a fire broke out in which Feiginova died, and with her went the surviving footage from the first version. Which means that we will most likely never see that other “Stalker” — the film Tarkovsky began shooting with Rerberg. We cannot compare them. We can only reconstruct it from descriptions, memories, screenplay drafts.
This loss also makes it impossible to definitively verify what, exactly, was on the damaged film stock. We do not know for certain how visually different the first version was from the second, how different its rhythm, editing, performances might have been. We know that the screenplay had been substantially reworked by the time production restarted. We know that the cinematographer changed. We know that the director suffered a heart attack and spent several months living with the sense that his major work had been destroyed.
The romantic version of this story goes like this: fate destroyed a mediocre film and forced a genius to make a masterpiece. It is a beautiful construction, but wrong in several essential respects. First, nothing suggests that the first version was mediocre — Tarkovsky would have been making auteur cinema with Rerberg too, not a genre action picture. Second, “fate” means ruined film stock, human conflict, the bureaucratic pressure of the Soviet film system, and a heart attack, not some mystical sign. Third, the price of this “gift of fate” was monstrous: budget overruns which, under Mosfilm conditions, could mean the project being shut down; the physical destruction of the director’s health; a break with the cinematographer with whom he had made his previous film; years of work gone nowhere.
The film changed not in a sudden lurch, but layer by layer. The Strugatskys’ screenplay went through countless iterations before the film-stock accident. The conflict with Rerberg was escalating in parallel. The move from Tajikistan to Estonia had already altered the visual environment. The heart attack forced Tarkovsky to rethink not only the film but his own relationship to time and death — and this inevitably found its way into the material. Knyazhinsky’s arrival made it technically possible to realize the degree of visual asceticism the director had reached as a result of all these accumulated layers. None of these factors was decisive by itself — they worked together, like a mechanism with several gears, each turning the next.
“Stalker” premiered in 1979. The film runs 163 minutes. It is almost devoid of special effects. Its three main characters walk through an abandoned space and talk. The camera moves so slowly that it seems — it is not sure itself whether it should go any further. And this film, assembled from the wreckage of a failed production, from a screenplay rewritten to pieces, from illness and conflict, became one of those works after which cinema could no longer pretend it did not know that time existed as material to work with. Tarkovsky died in 1986, in exile in Paris, of lung cancer — he was fifty-four, and “Stalker” remained the middle of his journey, the point where everything came together and from which everything diverged.