---

When Perception Becomes a Unit
​The Many Times of Walking through Taipei Main Station

當感知成為單位
​在台北車站行走的那些次數

A Scale Beyond Scale
​
​This research, Metrics of Reality: Beyond Human, extends from the fundamental spirit of surveying: re-understanding space through scale, position, distance, and form. While ancient civilizations used the human body as the origin of measurement, and late eighteenth-century France established the metric system through natural phenomena, the question of whether human perception itself can become a unit of measurement has remained insufficiently discussed. The body possesses complex sensory systems: seeing, listening, touching, walking, and breathing. If traditional measurement relies on quantifiable lengths and data, can perception also function as a stable (or unstable) reference for “measuring reality”? Can such a reference be named, standardized, or translated into a method of artistic practice? This research takes the work “Anthropometries” by the French artist Yves Klein as a historical point of departure, understanding how the body once intervened in art through its own scale, producing traces and serving as a medium to testify to reality. It also draws on the French philosopher Alain Milon’s discussions in Virtual Reality: Are We There or Not? concerning “virtuality,” “machine-mediated reality,” and the modes of existence of body and image. Once reality is reproduced, stored, and encoded by machines, does human perception still possess the potential to function as a unit of measurement? Thus, Metrics of Reality is divided into two phases: “Beyond the Human” (this project) and “Within the Human.” The present phase focuses on perceptual measurements generated by nature, space, and bodily walking “outside” the machine, and compares these with data provided by machines (Google Street View), in order to rethink how space is represented, measured, and seen.

Measurement: From Physical Scale to Perceptual Scale

Measurement is not only a tool of architecture and engineering, but also a mode of seeing. It produces different forms based on mass, material, scale, and application, and marks positions in space through one-, two-, and three-dimensional systems. This research, however, focuses on a measurement “beyond scale”: through walking, pausing, looking, filming, and breathing, how does the body leave unquantifiable traces in reality with each movement through space? During 2025, I wandered and recorded Taipei Main Station dozens of times: 9/20: In the morning, Taipei Main Station felt somewhat empty, but as I was a traveler, I did not record the moment and walked through quickly.

• 9/19: At dawn, Taipei Main Station had no people. It was empty and silent, with only the scrolling LED reflections on the floor. The underground passage still felt like a maze; the ceiling became lower. The main hall was empty. It was not until 2 a.m. that I realized the station also closes, yet it was still alive, like a breathing organism. The scrolling lights continued to reflect onto the metal shutters. The hall was no longer a box, but a creature breathing with light.
• 9/18: At 3 p.m., the hall was filled with soft scattered light. Off-peak, the station was as quiet as a warm afternoon. At 11 p.m., the exterior was wrapped in deep blue light. It was very quiet.
• 4/30: In the afternoon there was no sunlight; it was overcast. The station turned a bleak pale white.
• 4/26: At 6 a.m., backlit, the station appeared dark. People had not yet woken up. There were no footsteps in the hall, only the sound of luggage wheels. Reflections of the scrolling lights on the floor gradually blurred with the rising sun, while the clean, new floor coated the LED colors with another layer.
• 3/12: In the afternoon, blue sky and white clouds were visible through the station. Sunlight scattered over the roof; the crossing shadows divided the hall into fragmented shapes.
• 2/27: In the afternoon, I walked from the metro to the underground passage and then to the main hall. I did not go outside. I recorded all the traces of my walking. There were many people and it was crowded, but looking back at the footage, people appeared only as blurred objects, while Taipei Main Station looked like a puzzle of sliced-up objects.
• 2/19: At noon, sunlight could not be felt. Standing in the center, I recorded the surroundings, continuously rotating myself as the central point, measuring the distance between myself and the hall.

Upon replaying, people in the footage gradually became blurred, while space became clearer and more continuous. The body’s steps turned into a kind of “stride unit,” camera angles into “angle units,” and breathing and pauses into invisible temporal scales. These traces can all be understood as “units of perception.”

Two Modes of Measurement: Human and Machine

These repeated walks and recordings at Taipei Main Station formed two parallel trajectories of measurement: one from the body, the other from the machine. Both rely on perception, but perception is positioned in drastically different ways. Through this “dual-track measurement,” the produced space contains both present perception and past spatial representations. Both forms of measurement involve human perception, but one operates through rational recording and post-processing, while the other records through sensory randomness.

Human Measurement (by human)

The human body measures in an unstable manner, easily affected by the environment. The rhythm of walking, the length of breath, the height of the gaze, and the slight trembling of the hand-held camera all form “perceptual units” of space. With each recording, bodily condition, emotion, and environment differ. Thus, people in the image become increasingly blurred, while space becomes more persistent and clearer within that blur. The body is not recording space itself, but rather the “distance between the body and space.”

Machine Measurement (by machine)

By contrast, Google Street View reconstructs the same city endlessly through fixed specifications: focal length, scale, coordinates, shooting paths, and stitching modes. Faces erased for privacy, paths stitched by algorithms, and parallax produced by fixed camera height together compose a “machine version of history.” It is a continuously automated recording, yet this objectivity still contains gaps, omissions, and biases. It is a form of “computed reality.” In bodily images, people gradually blur; in machine images, people are erased. Both point to the same fact: neither human perception nor machine perception is a complete reality. What remains are different versions of reality, one from bodily movement, the other from algorithmic logic.

Bodily Record and Machine Memory: Point Clouds as a Third Reality

Between bodily walking and machine computation exists another form of spatial representation: the point cloud. This project uses space composed of countless coordinate points as a third path for observing “reality.” A point cloud is neither photograph nor video, but spatial data composed of “points” and the “distances between points.” Differences in scanning device precision, scanning angles, and data density lead to entirely different spatial outcomes, and these deviations themselves record the state of reality at the moment of scanning. The qualities of point clouds allow them to carry both machine logic and traces of the operator’s body. While machines generate data through laser or optical principles, the scanning path, position, shakiness, and occlusion still come from the human. Thus, it exists between body and machine, neither fully objective nor fully subjective, presenting an “incomplete reality” composed of information volume, error, and density. Bodily perception becomes a fluid and unfixed unit of measurement. It is not metric; it has no millimeters or scale bars, yet it generates distinctive rhythmic images, stride lengths, and differences in viewpoint. Conversely, machine measurement, though standardized and comparable, produces gaps due to privacy processing, image stitching, and coordinate algorithms. These voids are not flaws but reveal the logical structure of “computed reality.” The differences between human and machine spatial records instead construct multiple versions of reality. In this context, the point cloud becomes a third way of seeing: through the relationships between coordinate points, variations in data density, and missing scanning paths, it presents porosity, discontinuity, and incompleteness. Rather than restoring reality, it points through its incompleteness toward a form of truth closer to perception

Inside and Outside Taipei Main Station

​Taipei Main Station as a site makes the interweaving of these three realities more apparent. In bodily recordings, people gradually blur; in machine images, people are directly erased; in point clouds, both people and objects remain only as differences in density. These layered blurs, speeds, and absences ultimately constitute the central question of this project: how is reality dismantled, recomposed, and re-represented through different modes of measurement? It is not metric, but it can generate differences in image, rhythm, stride, and perspective. Machine measurement, though objective, still produces gaps through privacy, stitching, and algorithms. What it shows is a “computed reality.” The spatial differences recorded by humans and machines thus reveal “multiple versions of reality itself.” Point clouds provide a third mode of measurement: presenting an “incomplete reality” of space through information volume and error. Taipei Main Station as a site layers speed and blur within reality. In bodily recordings, people gradually blur; in machine recordings, people disappear altogether. This phase, “Beyond the Human,” completes a comparative framework between machine vision, human vision, and spatial representation. The next phase will further investigate how perception redefines scale and reality when “humans are within.” The next stage of the research will be developed collaboratively with Nicholas Jia-Jun Toh, Mikyo Sho, and Chi-Lun Feng, advancing human-centered modes of measurement and perception from bodily behavior, sensory systems, sound, stride, and breathing, and will ultimately develop into the Taipei Main Station · No One series.

尺度之外的尺度

本研究《衡量現實的單位：在人之外》延伸自測量學（surveying）的基本精神：以尺度、位置、距離與形狀重新理解空間。然而，古文明以人體作為測量起點，18 世紀末法國透過自然現象制定公制，但「人類感知」是否也能成為一種測量單位，始終是未被充分討論的問題。 身體具有觀看、傾聽、觸摸、行走、呼吸等複雜的感知系統。若傳統測量仰賴可量化的長度與數據，那麼感知是否也能作為一種用來「衡量現實」的穩定（或不穩定）基準？此一基準是否能被命名、被規範，或被轉譯為創作方法？ 本研究以法國藝術家 Yves Klein 的「人體測量學」（Anthropométries）為歷史起點，理解身體曾以自身的尺度介入藝術、生產痕跡，並作為證明真實的媒介。同時亦參考法國哲學家 Alain Milon 於《虛擬真實：我們的身體在或不在？》中對「虛擬性」、「機器介入的現實」、「身體與影像存在方式」的討論，提出：當現實由機器再現、儲存與編碼之後，人類感知是否仍具有作為測量單位的可能？ 因此，「衡量現實的單位」分為兩個階段：「在人之外」（本計劃）與「在人之內」。本次成果著重於「人在機器之外」的自然、空間與身體行走所產生的感知測量，並與機器（Google 街景）所提供的資料進行比較、對照，重新思考空間被再現、被測量與被觀看的方式。

測量：從物理尺度到感知尺度

測量不僅是建築與工程的工具，也是一種觀看方式。它依據質量、材質、尺度與應用產生不同形式，也依據一維、二維、三維在空間中標記位置。然而，本研究關注的是一種「尺度之外」的測量：透過行走、停留、觀看、錄影、呼吸，身體在空間中的每一次移動，如何在現實中留下不可度量的痕跡？
2025年間，我在台北車站進行數十次遊走與紀錄：
9/20：早上的台北車站有點空，但我是個旅人，所以我沒有記錄當下，而是迅速走過
9/19：凌晨，台北車站沒有人，很空、很安靜，剩下地板的跑馬燈反射，地下道依然像一個迷宮，天花板變得更低了，大廳空無一人，直到兩點，我才知道台北車站也會關門，但台北車站還是活著，如同拉下的鐵門一般，跑馬燈依舊反射在鐵捲門上，大廳不再是一個方盒子，而是一個跟著燈光呼吸的生物。
9/18：下午三點，台北車站的大廳散落很柔和的燈光，非尖峰時刻的台北車站跟午後陽煦一樣安靜。半夜十一點，外頭的台北車站披著深藍色的燈光，它很安靜。
4/30：下午台北車站沒有陽光，是陰天，它變成慘白的顏色。
​4/26：早上六點，台北車站因為背光而呈現暗黑色，人們還沒起床，大廳沒有人行走的聲音，只有行李箱車輪經過的聲音，地面上的跑馬燈反射因東昇的陽光逐漸變得模糊，而乾淨嶄新的地板將跑馬燈的顏色重新上了一層顏色。
3/12：下午，台北車站看得到藍天與白雲，陽光散射在屋頂，屋頂交叉的陰影把大廳切割成散狀。
2/27：下午，我從捷運走到地下道，再走到大廳，我今天沒有出去，我把所有走過的痕跡錄了下來，人很多、很擠，但回頭看這些片段，發現人只是一個個模糊的物體，而台北車站是一片片切割下來的物件拼圖。
2/19：中午，台北車站感受不到陽光，我站在中間，將四周的場景記錄下來，我一直旋轉，把自己當作中心點，測量我與台北車站大廳的距離。
回放之後，鏡頭中的人逐漸變模糊，空間卻反而變得清晰而持續，身體的步伐成為一種「跨距」，鏡頭角度成為一種「角度單位」，呼吸與停留成為不可見的時間尺度，這些痕跡皆可視為「感知的單位」。

兩種測量方式：人與機器

上述在台北車站的多次行走與紀錄，形成了兩條並行的測量軌跡：一條來自身體，一條來自機器。兩者都依賴感知，但感知被放置在截然不同的位置。透過這種「雙軌測量」所產生的空間包含現有的感知以及過去的空間再現，兩種測量都包含了人的感知介入，但一種以理性的紀錄與再處理，另一種以感官隨機的方式紀錄。

人類測量（by human）

人的身體以一種不穩定、易受環境影響的方式進行測量。 走路的節奏、呼吸的長度、視線的高度、手持鏡頭的微晃，都形成了空間的「感知單位」。每一次拍攝，其身體狀態、情緒與環境條件皆不同，因此影像中的人逐漸模糊，而空間卻在模糊中顯得更持續、更清晰。 身體不是在紀錄空間本身，而是在紀錄「身體與空間之間的距離」。

機器測量（by machine）

相較之下，Google Street View 以固定的規格——焦距、比例尺、座標、拍攝路徑、拼接模式——無限次地重建同一個城市。 因隱私而被抹除的臉孔、因演算法而被縫合的路徑、因鏡頭高度固定而產生的視差，都組成了「機器版本的歷史」。 它是持續自動化的紀錄，但這樣的客觀仍帶有空缺、遺漏與偏差，是一種「經過運算的現實」。 在身體的影像中，人變得模糊；在機器的影像中，人則被抹除。兩者共同指向一件事：人類的感知，和機器的感知，都不是完整的真實。它們所保留下來的是不同版本的現實，一種來自身體的運動感知，另一種來自演算法的邏輯運算。

身體的記錄、機器的記憶：點雲作為第三種現實

在身體的行走與機器的運算之間，存在著另一種介於兩者之間的空間再現方式——點雲（Pointcloud）。本計畫以無數座標點構成的空間作為觀察「現實」的第三條路徑。點雲既不是照片，也不是影片，而是一種由「點」與「點之間的距離」所組成的空間資料。不同掃描設備的精度、掃描角度的選擇、資料密度的差異，都會導致截然不同的空間結果，而這些偏差本身便記錄了掃描當下的現實狀態。 點雲的特性，使其既具備機器運算的邏輯，也保留了一部分操作者的身體痕跡：機器透過雷射或光學原理生成資料，然而掃描的路徑、位置、是否晃動、是否遮蔽，卻依然來自人。於是它介於身體與機器之間，不完全客觀，也不完全感性；它呈現的是一種由資訊量、誤差與密度所構成的「不完整真實」。 身體的感知成為一種流動而無法固定的測量單位。它不是公制，它沒有毫米，也沒有比例尺，但能生成特有的影像節奏、移動跨距與視角差異。相對地，機器的測量雖然標準化、可比對、可重建，卻因隱私處理、影像拼接、座標演算法而產生空缺。這些空洞不是缺陷，而是揭示了「經過運算的現實」的邏輯特徵。人與機器各自留下的空間紀錄差異，反而構成了現實的多重版本。 在此脈絡下，以點雲當作第三種觀看方式：以座標點之間的關係、資料密度的變動，以及掃描路徑的缺失，呈現空間的多孔、不連續與未完成。它並非要還原現實，而是以其不完整性指向另一種更貼近感知的真實。

在台北車站的內與外

台北車站作為場域，使這三種現實交錯得更加明顯。身體的記錄中，人逐漸模糊；機器的影像中，人被直接消失；而在點雲中，人與物都僅以密度的差異留下痕跡。這些層層堆疊的模糊、速度與缺席，最終共同構成了本計畫所關注的核心問題：現實如何在不同測量方式下被拆解、重組與再現？ 它不是公制，但它能形成影像、節奏、跨距與視角的差異。機器的測量雖然客觀，但仍會因隱私、拼接、演算法而產生空缺。它呈現的是「經過運算的現實」。人與機器所記錄的空間差異，反而呈現「現實本身的多重版本」。點雲提供了第三種測量方式：以資訊量與誤差呈現空間的「不完整真實」。 台北車站作為場域，使現實的「快」與「模糊」層層疊加。在身體記錄中，人逐漸模糊；在機器記錄中，人直接被消失。本階段「在人之外」完成了機器視角、人類視角與空間再現三者之間的比較框架；下一階段將深入探討「人在其中」時，感知如何重新定義尺度與真實。下一階段研究將與杜家駿、蔣美喬、馮紀倫共同發展「以人為出發」的測量與感知方式，從身體行為、感知系統、聲音、步伐、呼吸等層次延伸新的測量語言，發展為 《台北車站‧沒有人》 系列創作。

---