cs.HCApr 4, 2026
SaveFlueBricks: A Construction Kit of Flute-like Instruments for Acoustic Reasoning
Organizations: National Taiwan University, Taipei, Taiwan · Independent Researcher, Taipei, Taiwan
Abstract
We present FlueBricks, a construction kit for acoustic reasoning via building and customizing flute-like instruments. By assembling generator, resonator, and connector modules that embody various aeroacoustic properties, users gain deeper understanding of how blowhole, tube length, and tone-hole placement alter onset, pitch, and timbre through hands-on experimentation. This forms a designer-player loop of configuring and playing to form, test, and refine acoustic behaviors-acoustic reasoning-shifting acoustic instruments from static artifacts to dynamic systems. To understand how users engage with this system, we conducted an exploratory study with 12 participants ranging from novices to professional musicians. During their explorations, we observed participants fluently switching between designer and player roles, scaffolding designs from familiar instruments, forming and refining their acoustic understanding of length, tone holes, and generator geometry, reinterpreting modules beyond their intended functions, and using their creations for performative acts such as pedagogical showing and musical expression. These collectively demonstrated FlueBricks's potential as a pedagogical tool for embodied acoustic reasoning.
Figures & tables
Figure 1. FlueBricks and the designer–player loop. FlueBricks is a construction kit of 3D-printed building blocks for building and customizing flute-like instruments. This figure illustrates the designer–player loop : users iteratively switch between designing instruments and playing them to form, test, and refine acoustic behaviors. (A–C) Users browse tangible building blocks that embody aeroacoustic properties, select a mouthpiece, and breathe into it to hear the sound it generates. (D) As designers, they connect modules to assemble a resonator body. (E) As players, they immediately test the sound by playing. (F–H) Based on what they hear, users reconfigure tangible modules to test new hypotheses (F&G) and play again to evaluate the sound effect (H). Through these iterations—hypothesizing, configuring, playing, and refining—users engage in acoustic reasoning . FlueBricks and the designer–player loop. FlueBricks is a construction kit of 3D-printed building blocks for building and customizing flute-like instruments. This figure illustrates the designer–player loop: users iteratively switch between designing instruments and playing them to form, test, and refine acoustic behaviors. (A–C) Users browse tangible building blocks that embody aeroacoustic properties, select a mouthpiece, and breathe into it to hear the sound it generates. (D) As designers, they connect modules to assemble a resonator body. (E) As players, they immediately test the sound by playing. (F–H) Based on what they hear, users reconfigure tangible modules to test new hypotheses (F and G) and play again to evaluate the sound effect (H). Through these iterations—hypothesizing, configuring, playing, and refining—users engage in acoustic reasoning.
Figure 2. Overview of the FlueBricks modular component system. A complete flute constructed from FlueBricks modules, comprising three component families— Generator , Resonator , and Connector —each exposing parameter-level control for acoustic reasoning. The system offers two tiers: a standard set of pre-assembled variants for getting started, and an advanced set for fine-grained acoustic manipulation. A zoomed-in sketch illustrates the Generator’s internal structure, highlighting its modularity and granularity. Overview of the FlueBricks modular component system. A complete flute constructed from FlueBricks modules, comprising three component families—Generator, Resonator, and Connector—each exposing parameter-level control for acoustic reasoning. The system offers two tiers: a standard set of pre-assembled variants for getting started, and an advanced set for fine-grained acoustic manipulation. A zoomed-in sketch illustrates the Generator's internal structure, highlighting its modularity and granularity.
Figure 3. Three coupled phases of flute-like sound production. (a) Air-jet formation : blowing into the mouthpiece creates a focused jet of air. (b) Jet-edge interaction : the air jet encounters the labium, a sharp edge, and oscillates between the inside and outside of the pipe, producing an edge tone. (c) Resonance amplification : The jet-labium interaction generates pressure oscillations that excite the resonator, which selectively reinforces certain frequencies through resonance to sustain the sound. These acoustic mechanisms inform FlueBricks’s generator design: following the granularity principle, we decompose the generator into 8 acoustically meaningful submodules that map to these geometries (flue tunnel, air chamber, windway, window, splitting edge, sound chamber), enabling parameter-level control over tone production. Three coupled phases of flute-like sound production. (a) Air-jet formation: blowing into the mouthpiece creates a focused jet of air. (b) Jet-edge interaction: the air jet encounters the labium, a sharp edge, and oscillates between the inside and outside of the pipe, producing an edge tone. (c) Resonance amplification: The jet-labium interaction generates pressure oscillations that excite the resonator, which selectively reinforces certain frequencies through resonance to sustain the sound. These acoustic mechanisms inform FlueBricks's generator design: following the granularity principle, we decompose the generator into 8 acoustically meaningful submodules that map to these geometries (flue tunnel, air chamber, windway, window, splitting edge, sound chamber), enabling parameter-level control over tone production.
Figure 4. How changes in pipe geometry affect pitch. This figure compares three types of resonator configurations. Left: an open pipe creates sound by supporting vibrations along its full length, with both ends open. Middle: a closed pipe supports a different vibration pattern, resulting in a longer effective length and a lower pitch compared to an open pipe of the same size. Right: a tone hole acts like a shortcut: if large enough, it defines a new endpoint for vibration, effectively shortening the pipe and raising the pitch. The cyan region shows the effective length—the part of the pipe that actually shapes the sound. How changes in pipe geometry affect pitch. This figure compares three types of resonator configurations. Left: an open pipe creates sound by supporting vibrations along its full length, with both ends open. Middle: a closed pipe supports a different vibration pattern, resulting in a longer effective length and a lower pitch compared to an open pipe of the same size. Right: a tone hole acts like a shortcut: if large enough, it defines a new endpoint for vibration, effectively shortening the pipe and raising the pitch. The cyan region shows the effective length—the part of the pipe that actually shapes the sound.
Figure 5. Generator fine-grained control configuration. FlueBricks enables airflow and tone production control through parameter-level manipulation. (a–f) The generator is assembled from eight modular parts, each controlling specific acoustic parameters. (b) Varying air deflector angles (90°, 78°, 35°) reshapes airflow, shifting tone from focused high frequencies to overtone-rich textures. (c) Changing the air facade (sharp, flat, arc) affects timbre brightness and ease of onset. (e–f) A shorter splitting edge raises the overblowing threshold, improving pitch stability in lower registers. Generator fine-grained control configuration. FlueBricks enables airflow and tone production control through parameter-level manipulation. (a–f) The generator is assembled from eight modular parts, each controlling specific acoustic parameters. (b) Varying air deflector angles (90 degrees, 78 degrees, 35 degrees) reshapes airflow, shifting tone from focused high frequencies to overtone-rich textures. (c) Changing the air facade (sharp, flat, arc) affects timbre brightness and ease of onset. (e–f) A shorter splitting edge raises the overblowing threshold, improving pitch stability in lower registers.
Figure 6. Resonator length control configuration. FlueBricks enables pitch control through modular length manipulation. (a) Basic nodes of different lengths provide discrete pitch steps. (b-1 to b-4) Stacking nodes extends the air column, progressively lowering pitch and deepening timbre. (c) Attaching an adapt cap shifts to closed-pipe mode for further pitch lowering. (d) Telescoping nodes enable continuous pitch bending for glissando effects. Resonator length control configuration. FlueBricks enables pitch control through modular length manipulation. (a) Basic nodes of different lengths provide discrete pitch steps. (b-1 to b-4) Stacking nodes extends the air column, progressively lowering pitch and deepening timbre. (c) Attaching an adapt cap shifts to closed-pipe mode for further pitch lowering. (d) Telescoping nodes enable continuous pitch bending for glissando effects.
Figure 7. Tone-hole manipulation configuration. FlueBricks enables pitch and timbre control through hole size, position, and continuous adjustment. (a) A fully-opened branch node creates a high pitch by shortening the effective resonator length. (b-1 to b-3) Adapt caps adjust pitch by varying tone-hole size across caps and changing direction through rotation. (c-1) A tuning node with a vertical slot enables broader tone shaping. (c-2) A sliding tweak ring supports real-time pitch modulation for glissando effects. Tone-hole manipulation configuration. FlueBricks enables pitch and timbre control through hole size, position, and continuous adjustment. (a) A fully-opened branch node creates a high pitch by shortening the effective resonator length. (b-1 to b-3) Adapt caps adjust pitch by varying tone-hole size across caps and changing direction through rotation. (c-1) A tuning node with a vertical slot enables broader tone shaping. (c-2) A sliding tweak ring supports real-time pitch modulation for glissando effects.
Figure 8. Connector configurations and instrument hybridization. (a) An all-in-one (L) generator module is attached to create a distinct timbre. (b) Tapered connectors are added to provide ergonomic angling and restrict airflow for better breath control. (c) A bifurcated air hub integrates a standard recorder with a single-note flute for drone accompaniment. (d) The setup is upgraded to a multi-note flute , enabling simultaneous melody lines and harmonic complexity. Connector configurations and instrument hybridization. (a) An all-in-one (L) generator module is attached to create a distinct timbre. (b) Tapered connectors are added to provide ergonomic angling and restrict airflow for better breath control. (c) A bifurcated air hub integrates a standard recorder with a single-note flute for drone accompaniment. (d) The setup is upgraded to a multi-note flute, enabling simultaneous melody lines and harmonic complexity.
Figure 9. Overview of the user study procedure. The study consisted of six phases, alternating between interviews, open-ended exploration, and goal-directed design. The 90-minute hands-on exploration period enabled participants to engage in iterative hypothesize–configure–evaluate–conclude cycles, which we analyzed as evidence of acoustic reasoning (Section 5 ). While time blocks were planned, the schedule was flexibly adjusted based on each participant’s engagement level and curiosity. Overview of the user study procedure. The study consisted of six phases, alternating between interviews, open-ended exploration, and goal-directed design. The 90-minute hands-on exploration period enabled participants to engage in iterative hypothesize–configure–evaluate–conclude cycles, which we analyzed as evidence of acoustic reasoning. While time blocks were planned, the schedule was flexibly adjusted based on each participant's engagement level and curiosity.
Figure 10. Module-level acoustic use across participants. Each row represents a participant and each column a module. Colors indicate the highest level of sound-related use: Non-Acoustic (no acoustic intent), Exploratory (acoustic testing), or Intentional (sound-related function). Module-level acoustic use across participants. Each row represents a participant and each column a module. Colors indicate the highest level of sound-related use: Non-Acoustic (no acoustic intent), Exploratory (acoustic testing), or Intentional (sound-related function).
Figure 11. Participant-level perceptual and emotional responses. Each column represents a participant, grouped by musical or woodwind background. Rows indicate Pitch Variation , Timbre Variation , Onset Sensitivity , and Sense of Control ; the bottom row shows an affective summary (Positive, Neutral, Negative). Color intensity encodes frequency ranges (0–2, 3–4, 5+ times). Participant-level perceptual and emotional responses. Each column represents a participant, grouped by musical or woodwind background. Rows indicate Pitch Variation, Timbre Variation, Onset Sensitivity, and Sense of Control; the bottom row shows an affective summary (Positive, Neutral, Negative). Color intensity encodes frequency ranges (0–2, 3–4, 5+ times).
Figure 12. The designer–player loop in action. P12’s (Pro Woodwind) session illustrates the designer–player loop : rapid alternation between configuring and playing, with each sonic outcome informing the next adjustment. The designer–player loop in action. P12's (Pro Woodwind) session illustrates the designer–player loop: rapid alternation between configuring and playing, with each sonic outcome informing the next adjustment.
Figure 13. Exploration timelines of all 12 participants. Each row shows one participant’s session across three phases (dashed lines); point markers denote reasoning events— instrument scaffolding (blue stars), rule formation (green circles), and reinterpretation (purple diamonds); horizontal bars denote duration-based performative acts ; right-side charts summarize per-behavior frequency. Exploration timelines of all 12 participants. Each row shows one participant's session across three phases (dashed lines); point markers denote reasoning events—instrument scaffolding (blue stars), rule formation (green circles), and reinterpretation (purple diamonds); horizontal bars denote duration-based performative acts; right-side charts summarize per-behavior frequency.
Figure 14. User reinterpretations of FlueBricks components. Despite their original functions—(a) tuning node , (b) tweak ring , (c) air hub , and (d) bridge connector —participants repurposed them creatively, revealing unexpected uses and modular flexibility. User reinterpretations of FlueBricks components. Despite their original functions—(a) tuning node, (b) tweak ring, (c) air hub, and (d) bridge connector—participants repurposed them creatively, revealing unexpected uses and modular flexibility.
Figure 15. Pedagogical showing as performative acts. Top row: P5 (Pro Woodwind) used systematic pitch recording, melody demonstration, and a classroom game proposal to build a structured teaching strategy. Bottom row: P11 (Pro Musician) explored timbre manipulation and proposed open-ended discovery activities. Pedagogical showing as performative acts. Top row: P5 (Pro Woodwind) used systematic pitch recording, melody demonstration, and a classroom game proposal to build a structured teaching strategy. Bottom row: P11 (Pro Musician) explored timbre manipulation and proposed open-ended discovery activities.
Figure 16. Musical expression as performative acts. Left: P12 (Pro Woodwind) performing the melody of My Heart Will Go On with a custom multi-note setup. Right: P12 improvising freely across eight distinct pitches. Pitch contours were extracted using SwiftF0 ( Nieradzik, 2025 ) . Musical expression as performative acts. Left: P12 (Pro Woodwind) performing the melody of My Heart Will Go On with a custom multi-note setup. Right: P12 improvising freely across eight distinct pitches. Pitch contours were extracted using SwiftF0.
Figure 17. Completed instruments from the scenario-based study. All participants except P3, regardless of musical background, successfully built playable flutes using FlueBricks. P3 instead created a prank toy. Completed instruments from the scenario-based study. All participants except P3, regardless of musical background, successfully built playable flutes using FlueBricks. P3 instead created a prank toy.
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 18. Pitch contrast and multiphonic effects through basic and branch node combinations. (a) Raising the tone hole vertically (with fixed tube length) narrows the pitch gap between covered and uncovered states. (b) Extending the horizontal branch progressively lowers the covered tone (E4 → Eb4 → Db4 → Bb3) and eventually produces multiphonic effects when the upper note jumps an octave (G4 → F4 → F5), demonstrating how spatial layout enables musical expressivity. Pitch contrast and multiphonic effects through basic and branch node combinations. (a) Raising the tone hole vertically (with fixed tube length) narrows the pitch gap between covered and uncovered states. (b) Extending the horizontal branch progressively lowers the covered tone (E4 to E-flat 4 to D-flat 4 to B-flat 3) and eventually produces multiphonic effects when the upper note jumps an octave (G4 to F4 to F5), demonstrating how spatial layout enables musical expressivity.
| Participant | Age | Gender | Profession | Expertise Level | Creativity (Self) | Design Exp. | LEGO Play Style |
| P1 | 29 | Male | Engineer (IC design) | Novice | High | No | Exploratory |
| P2 | 36 | Female | Planner (marketing) | Amateur | Low | No | None |
| P3 | 30 | Male | Designer (educational games) | Novice | Low | No | Exploratory |
| P4 | 32 | Female | Teacher (early childhood) | Amateur | High | Yes | Exploratory |
| P5 | 51 | Female | Professor (innovation / design, flute) | Pro Woodwind | High | No | Goal-driven |
| P6 | 27 | Male | Student (music, dizi/suona) | Pro Woodwind | High | Yes | Exploratory |
Table A.1. Overview of participant demographics and creative backgrounds, including age, gender, profession, musical expertise, self-reported creativity, instrument design or modification experience, and LEGO play style, collected during the pre-interview. Instrument design/modification experience includes both casual and professional experiences. LEGO play style was used as a proxy for creative behavior, categorized as None , Goal-driven (structured builds), or Exploratory (open-ended play). Overview of participant demographics and creative backgrounds for 12 participants (P1–P12), including age (20–51), gender, profession, musical expertise (Novice, Amateur, Pro Woodwind, or Pro Musician, three per group), self-reported creativity (Low, Medium, High), instrument design or modification experience (Yes or No), and LEGO play style (None, Goal-driven, or Exploratory), collected during the pre-interview. Instrument design or modification experience includes both casual and professional experiences. LEGO play style was used as a proxy for creative behavior. Participants span engineers, teachers, musicians, artists, and students.
More in this topic
Audio Reasoning
Audio reasoning studies how AI systems infer events, relationships, intent, and context from sound, beyond recognizing words or matching labels.
See all 64 papers