Feedback A microphone picks up subtle vibrations in the space. The signal is amplified through the circuit, and a sound emerges from the loudspeaker: “...hiss........” The microphone receives it again and amplifies it, while the loudspeaker plays it back once more, and the sound gradually turns into “shiiii~~~.” Receiving, amplifying, and playing back—the cycle repeats, and the initially faint noise is gradually amplified into a resonance that fills the entire space: “eeee———.” This is how acoustic feedback operates: as if the system were listening to itself and changing itself through the loop. It does not require any deliberate external input; background noise in the air or within the circuit itself can be amplified through the feedback process and eventually develop into a massive sound. These sounds vary according to the structure, materials, and acoustic characteristics of the space, allowing the space itself to participate in the evolution of the sound. This series of works began with experiments in acoustic feedback, using operational mechanisms to establish relationships among objects, technology, environment, and audience. Rather than beginning with the direct production of content, the works leave their final outcomes undetermined. Sound emerges from a system that operates autonomously and changes continuously, while the space shifts from being merely a site in which the work is placed to becoming part of the work itself.

Feeding (2020) consists of several acoustic feedback units, each containing a suspended microphone, a single loudspeaker driver, and a sensor that detects sound intensity. When the sound becomes too quiet, the microphone descends toward the loudspeaker, strengthening the feedback; when the sound becomes too loud, the microphone rises and moves away from the loudspeaker, reducing it. The sound intensity thus feeds back into and regulates its own feedback state, keeping the system in a continuous process of seeking equilibrium.

Cross (2021) forms an audio loop from four sets of microphones and loudspeakers. The sound received by the first microphone is transmitted to the next loudspeaker and then passes sequentially through all four units. The sound received by the final microphone is eventually sent back to the first loudspeaker. During each transmission, the sound undergoes subtle changes caused by the different frequency responses of the loudspeakers, spatial reflections, and interventions by the audience. These changes continuously re-enter the system through the loop, causing the sound to deform over time and generate sonic results that cannot be precisely predicted.

Cyberfly (2021) consists of a small motor, a single propeller blade, a counterweight, and a microphone suspended above a loudspeaker. The imbalance caused by the rotation makes the device twist continuously, while the microphone attached to its end constantly changes its distance and angle relative to the loudspeaker. The variations produced by these unstable interactions interfere with the formation of acoustic feedback, causing feedback that might otherwise continuously intensify to emerge and recede intermittently, producing irregular low-frequency sounds.

Tree Noise (2022) attaches a contact microphone to the trunk of a tree, receiving vibrations and sounds transmitted through the tree itself and from the surrounding environment. These sounds are played back through several mushroom-shaped devices distributed around the tree. The sounds are then picked up again by the microphone on the trunk from different distances and re-enter the system, forming a loop jointly constituted by the tree, the environment, and the devices.

Wrong Sound The works described above all belong to a series of experiments that I previously referred to as “Wrong Sound.” When using audio systems, many phenomena need to be avoided in order to maintain signal quality, equipment stability, and hardware safety, and feedback is one of them. When phenomena that are normally regarded as technical problems become the primary sonic material of a work, the “error” itself also becomes an object of listening. &#x611b extends this line of thought by using another phenomenon that is normally meant to be avoided: the turn-on/turn-off transient. When equipment is not switched on or off in the recommended sequence, or when power is directly connected or disconnected, the sudden change in current may be transmitted through an amplifier to the loudspeaker, producing a short and forceful pop and, under certain conditions, potentially damaging the speaker. In &#x611b (2020) , I convert continuously generated text from an online forum into binary signals using UTF-8 encoding, then use 0s and 1s to directly control the connection and disconnection of power to the loudspeaker, producing a sequence of electrical transient sounds.

Here, the loudspeaker no longer functions according to its intended purpose of reproducing an audio signal; instead, it produces sound through the switching of electrical power itself. Does the so-called “error” exist in the sound itself, or does it arise from the technical system’s assumptions about what constitutes normal operation? When phenomena that would normally be excluded are removed from established conventions of use and become the primary sound source of a work, are we listening to an error, a manifestation of the physical operation of a technical device itself, or another mode of listening constructed through technical rules?

From Luigi Russolo’s incorporation of industrial and urban noise into musical vocabulary, to Steve Reich’s use of a swinging microphone to allow acoustic feedback to unfold on its own, and Alvin Lucier’s explorations of the relationships between sound, space, and resonance...... noise, feedback, and spatial acoustics all have art-historical lineages that long precede my own practice. As I developed these works, Adam Basanta became one of my more direct contemporary references. By moving microphones, changing the distance between microphones and loudspeakers, or working with the acoustic properties of a space, he uses acoustic feedback as a mechanism for revealing the relationships between devices, space, and listening. Looking back at my own works, I had tried to establish differences through the formal design of the installations, more complex operational logics, and connections between the works and personal experience. However, if these works are understood solely as explorations of noise and spatial acoustics, such differences do not seem sufficient to constitute a substantial theoretical distinction.

Uncertainty This gradually led me away from works centered on loudspeaker systems and toward considering the questions shared across these experiments from a broader perspective. One thread that began to emerge was how uncertainty from different sources participates in the operation of a system. In the works described above, uncertainty was in fact already present in different forms: Feeding breaks mechanical repetition through irregular pause durations, while interactions among multiple units lead to unstable and unpredictable states; Cross is disturbed by the frequency responses of different loudspeakers, spatial reflections, and interventions by the audience; the unbalanced mechanical structure in Cyberfly produces irregular twisting movements; Tree Noise incorporates the complexity of the external environment into its loop; and in &#x611b , continuously changing online text prevents the binary sequence entering the system from being predetermined.

Although all of these works developed from system noise, looking back, &#x611b actually marked an important shift in my practice. The Feedback series mainly experimented with how different uncertain factors act upon the process of sonic change, whereas &#x611b extended this question outward toward the choice of the source that drives the system. Depending on the rules through which they operate, the uncertain factors selected in a work can exert different degrees of control over the outcome and may act on different levels of the work. These relationships, in turn, determine how the source manifests itself within a particular work. From this point onward, I gradually shifted my attention toward how different system configurations determine which aspects are controlled by the artist and which are jointly formed by the environment, materials, and biological processes.

In Soul (2022) , Soul-vinyl (2023) , the audience’s entrance into the exhibition space triggers an electrostatic generator that produces static electricity. Electromagnetic sensing devices receive electromagnetic signals in the space and transmit them to a performance system that activates different instruments. I define the chords and rhythmic rules available to the system, while the development of the music is influenced by continuously fluctuating electromagnetic signals.

Microbial Ensemble (2026) uses electrical signals measured from an MFC (microbial fuel cell) to drive a physical drum kit. The triggering thresholds and instruments are predetermined by me, while the timing of each strike is influenced by microbial activity that cannot be precisely predicted.

In these works, I determine the basic structure of the system, the sensing methods, the conditions under which signals enter the system, and which sonic parameters they can ultimately affect. Once these rules have been established, when the signals appear, how they fluctuate, and what results the system produces at a given moment are left to develop through the system’s own operation. Artificially defined random mechanisms, such as random functions in software, can have their ranges specified in advance and therefore impose a certain degree of constraint on the outcome. Variations produced by mechanical structures, space, the environment, or biological activity, however, only emerge through actual operation and contact with the material environment. External uncertainties may also become entangled with random functions within the system. The controllable and uncontrollable therefore coexist at different levels of the work and influence the outcome to different degrees. Introducing uncertainty does not mean that the work has escaped control, nor does it imply the withdrawal of the author. What changes is where control is situated and the extent to which different factors can participate in the development of the work. I have shifted from directly arranging sonic outcomes toward defining the conditions under which those outcomes can emerge.

Conclusion In the acoustic feedback series, sound is more than an outcome produced by the system. It also reveals how space, devices, and rules interact with one another, becoming a way of sensing the site and its changes. In later works, when factors such as microbial activity, electromagnetic signals, online information, and audience behavior encounter the mechanisms, sensing processes, transformations, and methods of sound production established by the work, sound emerges as an event produced through these encounters. Because these systems operate in real time, each act of listening captures only a partial moment of their ongoing processes.

Beyond the Rules At this stage, my works still operate primarily within a fixed set of rules. Although uncertain factors affect the sound and the state of the system, they still enter the work only as inputs. I am interested in whether the source could develop more layered behaviors and relationships within the work. For example, if the results produced by a work could feed back into and modify its own mode of operation, could long-term iteration lead to more complex forms of self-regulation? In addition, several of the works discussed above consist of multiple units that influence one another through shared spaces, signals, and environments. This has drawn my attention to another question that has yet to be fully explored: if each system follows a simple set of rules while also becoming an input or operational condition for other systems, could their coupling give rise to collective behaviors that no individual system possesses on its own? These two questions, which I am still exploring, may become directions for the future development of my practice.

Magic Magic is often composed of materials, spells, emotions, and effects. The preceding discussion has attempted to clarify how I configure systems to distribute control, introduce uncertainty, and bring different materials and phenomena into relation. To me, these processes feel much like the study of magic: placing an imagined idea (emotion) into real media (materials) and technical mechanisms (spells), waiting for them to produce an unpredictable sound (effect), and observing how the resulting outcomes respond in turn to the initial motivation.

Feedback 麥克風接收空間中細微的震動，經由電路擴大，聲音從喇叭傳出「...嘶........」。麥克風重複接收、放大，喇叭再播出，聲音慢慢變成「西~~~」。接收、放大、播出，反覆循環，原本微小的雜訊逐漸被放大成整個空間的共鳴「咿———」。聲音回受機制如此運作，像是自己聆聽自己，並且在循環中改變自己。它不需要刻意加入其他外部輸入，任何空氣中或電路本身的底噪都可能在回授的過程中被放大，最終衍生出巨大的聲響。這些聲響會隨著空間結構、材質與聲學特性產生差異，使空間也參與著聲音的演變。此系列實踐從聲音回受的實驗出發，嘗試透過運作機制建立物件、技術、環境與觀眾之間的關係。作品不以直接生產內容出發，最終的結果也不會被預先決定，聲音是從一套自主運作並持續變化的系統中生成，而空間也從放置作品的場域變成作品構成的一部分。

Feeding (2020) 由數個聲音回授單元組成，每個單元包含一組垂釣的麥克風、一顆單體喇叭以及偵測聲音強度的感測器。當聲音過小時，麥克風會下降靠近喇叭使回授增強，當聲音過大時，麥克風則上升遠離喇叭使回授減弱。聲音強度回過頭來調節自身的回授狀態，使系統持續處於尋找平衡的狀態。

Cross (2021) 由四組麥克風與喇叭構成一個聲音迴路。第一組麥克風接收的聲音會傳送至下一組喇叭，依序經過四個單元，最後一組麥克風收到的聲音會再傳回到最初的喇叭。聲音在每一次傳遞中受到不同喇叭的頻率響應、空間反射與觀眾介入的影響而產生細微變化，這些變化隨著循環不斷重新進入系統，使聲音持續變形並產生無法精確預期的聲響。

Cyberfly (2021) 由小型馬達、單片螺旋槳、配重與麥克風構成並懸吊於喇叭上方。旋轉造成的不平衡使裝置持續扭動，尾端連接的麥克風不斷改變與喇叭之間的距離和角度。這些不穩定連動產生的變化干擾聲音回授的產生，使原本可能不斷增強的回授時而形成、時而消退，產生不固定出現的低頻聲響。

Tree Noise (2022) 將接觸式麥克風固定於樹幹，接收樹木內部與周遭環境傳導至樹幹上的震動與聲響，透過分布於樹木周圍的數個蕈狀裝置播放。這些聲音又會從不同距離被樹幹上的麥克風接收重新進入系統，使樹木、環境與裝置共同形成循環。

Wrong Sound 上述作品皆屬於我過往稱為「錯誤的聲音」的一系列實驗作品。在使用音響系統時，為了維持訊號品質、設備穩定與器材安全，存在許多需要被避免的現象，回授便是其中之一。當這些原本被視為技術問題的現象成為作品主要的聲響，「錯誤」本身也成為了聆聽的對象。 &#x611b 延續這一思考，使用了另一種通常需要被避免的現象：瞬間訊號（turn-on/turn-off transient）。當設備未依照建議的順序開啟或關閉或電源被直接接通與切斷時，瞬間的電流變化可能經由擴大器傳送至喇叭，形成短促而強烈的爆裂聲（pop），在特定情況下也可能對喇叭造成損害。在 &#x611b (2020) 中我將網路論壇中持續產生的文字透過 UTF-8 編碼轉換為二進位訊號，並以 0 與 1 直接控制喇叭的通電與斷電，產生一連串電源突波聲響。

喇叭在這裡不再依照原本的用途播放一段聲音訊號，而是透過電源切換本身發聲。所謂「錯誤」究竟存在於聲音本身，還是來自技術系統對正常運作方式的預設？當原本需要被排除的現象脫離既定的使用規範，成為作品主要的聲音來源時，我們所聆聽的究竟是一個錯誤、一套技術裝置自身物理運作的顯露，還是另一種由技術規則所建構的聆聽方式？

從 Luigi Russolo 將工業與城市噪音納入音樂語彙，到 Steve Reich 讓聲音回授隨著麥克風擺動自行展開，以及 Alvin Lucier 對聲音、空間與共鳴關係的探索......。噪音、回授與空間聲學都有著比我的實踐更為悠久的藝術史脈絡。在我發展這些作品的過程中，Adam Basanta 則是更為直接的當代參照之一。他透過移動麥克風、改變麥克風與喇叭之間的距離，或利用空間的聲學特性，使聲音回授成為揭露裝置、空間與聆聽關係的機制。回頭檢視自己的作品，我過去曾試圖透過裝置造型、更複雜的運作邏輯以及作品與個人經驗之間的連結來建立差異。然而，如果仍只將這些作品理解為對噪音與空間聲學的探索，這些差異似乎並不足以構成實質的理論區隔。

Uncertainty 這使我逐漸離開以喇叭系統為主要機制的作品，並開始以更廣泛的角度思考其中共同存在的問題。其中一個浮現的線索是不同來源的不確定性如何參與系統的運作，在上述作品中，不確定性其實已經以不同形式存在： Feeding 透過不固定的暫停時間打破機械性的重複、多個單元的互相影響導致難以預期的不穩定狀態； Cross 受到不同規格喇叭的頻率響應、空間反射與觀眾介入的擾動； Cyberfly 中不平衡的機械結構產生不規則的扭動； Tree Noise 將複雜的外在環境納入循環； &#x611b 中即時變動的網路文字使進入系統的二進位序列無法被預先編排。

儘管都是以系統噪音作為發展基礎，回頭來看，&#x611b 其實構成了我的實踐中一次重要的轉向。Feedback系列中的不確定因素主要作用於聲音變化的過程，而 &#x611b 則更往外涉及了驅動系統的來源（source）的選擇。作品所選擇的不確定因素在不同的規則下對結果會有不同程度的控制，也可能作用於作品的不同層面，而這種關係也會使得source在特定作品中有特定的顯現。從這點延伸，我漸漸將思考的重點放在作品如何透過不同的系統配置決定哪些部分由藝術家控制，哪些部分由環境、物質與生物共同形成。

Soul (2022) , Soul-vinyl (2023) 中，觀眾進入展場後觸發起電裝置產生靜電，電磁感測裝置接收空間中的電磁訊號，並將訊號傳送至演奏系統觸發不同的樂器。我設計系統可以使用的和弦與節奏規則，而音樂的發展則受到持續波動的電磁訊號影響。

Microbial Ensemble (2026) 利用MFC（微生物電池）中測量到的電訊號驅動一組實體爵士鼓。觸發的閾值與樂器由我預先設定，而敲擊發生的時間受到微生物無法被精確預測的活動影響。

在這些作品中我決定了系統的基本結構、感測方式、訊號進入系統的條件和它最終能夠影響哪些聲音參數，在這些規則建立之後訊號何時出現、如何波動以及系統在特定時刻形成什麼結果則交由系統自行發展。人為設定的隨機機制，例如程式中的隨機函數，其範圍可以事先被界定並對結果有著一定程度的規範。而機械結構、空間、環境或生物活動所產生的變化則要在實際運作的過程與物質環境接觸才會發生，外部不確定性有時也會與系統內的隨機函數有所牽連，作品中的可控與不可控是同時存在於系統的不同層面並以不同程度共同影響結果。因此引入不確定性並不能代表作品真正擺脫控制，更不意味著作者的退出，它改變的是控制被配置在什麼位置，以及不同因素能在多大程度上參與作品的發展。與其說我放棄了對結果的控制，不如說控制從對聲音結果的直接安排，轉移到了對其發生條件的設定。

Conclusion 在聲音回授系列創作中，聲音不只是系統產生的結果，它同時暴露了空間、裝置與規則彼此如何互動，成為了感受場域及其變化的線索。在後續作品中，當微生物活動、電磁訊號、網路資訊、觀眾的行為等不同因素遇上作品所建立的機制、感測、轉換及發聲方式，聲音也是這些因素彼此相遇時所發生的事件。由於這些系統都是即時運作，每一次聆聽都是對其運作過程的一次局部截取。

Beyond the Rules 現階段的作品仍主要在一套固定規則中運作，不確定因素雖然影響聲音與系統狀態，卻還只是作為一項輸入進入作品，我很好奇有沒有可能讓source在作品中建立更多層次的行為與關係，例如：如果作品產生的結果能進一步反過來修改自身的運作方式，經過長時間的迭代是否會形成更複雜的自我調節？ 另外，上述有幾件作品存在多個單元透過共享的空間、訊號與環境互相影響，這也讓我注意到另一個尚未充分展開的問題：如果每個系統都遵循一套簡單的規則，並進一步成為其他系統的輸入或運作條件，彼此耦合之後是否可能形成任何單一系統都不具有的整體行為？這兩個目前仍在思考的問題可能成為後續實踐發展的方向。

Magic 魔法通常由材料、咒語、情感與效果所構成。前面的討論試圖釐清我是如何透過系統配置控制、引入不確定性，並使不同的物質與現象產生關係。而這些過程對我而言就很像是在研究魔法：將一個想像（情感）放進現實媒介（材料）與技術機制（咒語）之中，等待它產生某種無法預期的聲響（效果），觀察它們如何回過頭來回應最初的動機。