Sonic Facade – Comprehensive Study Notes
- Author: Alina Granville
- Degree: Bachelor of Science in Architecture, MIT, June 2012
- Thesis submitted: May 25, 2012
- Supervisor: John Fernandez (Associate Professor of Architecture and Building Technology and Engineering Systems)
- Committee: Terry Knight, Rizal Muslimin (Shape-Grammar critics)
- Title: “Sonic Facade, Creating a Sounding Architecture”
- Copyright © 2012. MIT may reproduce and distribute the thesis in any medium.
- Accompanying media: Embedded/CD audio tracks 1–14 corresponding to precedents, tube tests, and example buildings.
Motivation
- Architecture is dominated by the visual; acoustic potential is often ignored.
- Everyday built elements already generate sound (e.g., footsteps on floors, wind at windows).
- Materials either amplify/reflect or dampen/absorb sound.
- Form can channel sound, creating spatial auditory experience.
- Author’s background as a woodwind player → curiosity about wind-driven sound in pipes.
- Inspiration from sound artists who scale traditional instruments for nature:
- Wind replaces a musician’s breath, water can replace bellows, rain acts as mallets, etc.
- Historic tension: Murray Schafer notes concert halls evolved to separate sonic experience from everyday life; author aims to reintegrate them.
- Critique by Bernhard Leitner: modern architecture lacks relationship among “sound, space, body.”
- Thesis challenge: Can a façade itself become a wind-powered musical instrument that also informs occupants about exterior conditions (wind, rain, passer-by interaction)?
Objectives & Scope
- Design a wind-powered sonic façade that integrates into existing walls.
- Develop a seven-rule shape grammar enabling users to decide placement/implementation on varied buildings.
- Conduct extensive studies:
- Wind behavior around buildings & tubes.
- Tube acoustics, scaling, and typology.
- Spatial grammars for aggregation.
- Limit thesis to three demonstrative building applications; grammar intended to be universal.
- Produce hypothetical composite audio for each example based on recorded tube samples.
Precedents
- Instrument classification (Hornbostel–Sachs): focus on aerophones (vibrating columns of air).
- Simple pan flute → complex brass & woodwinds show variety of tube geometries.
- Large wind-driven sound projects:
- Singing, Ringing Tree (Tonkin Liu) – mixed tube types; Track 1.
- Aeolus (Luke Jerram) – Aeolian harp strings + resonating tubes; Track 2.
- Songlines (Joe Snell) – 10 m cylinder with Aeolian harps + LEDs; Track 3.
- Chimecco (Mark Nikkon) – bridge-hung pipes act as chimes; Track 4.
- Sonambient (Harry Bertoia) – rod clusters producing metallic drones; Track 5.
- Organ of Corti (Liminal) – sonic crystal tube array filtering/amplifying ambient sounds; Track 6.
Instrument & Tube Typologies Examined
- Straight circular tubes: open–open, open–closed (+ block), transverse-slot variants.
- Bent tubes: mild bend, 180^{\circ} U-bend, U-bend with closed end, split-length U-bend.
- Split & partition tubes: dual branches, interior cross walls (little acoustic change).
- Transverse-opening tubes (slots, flaps) – most variants showed negligible change except long narrow slot.
- Aeolian-harp hybrids (strings across square/circular sections) – ineffective in wind tests.
- Bottle/organ-pipe and tapered sections – required direct wind; unsuitable for façade context.
- Harry Bertoia style rods – sound, but hard to excite naturally.
Wind Studies
- Wind seeks shortest path around obstacles → acceleration at edges.
- High-velocity zones near corners & openings ideal for tube placement.
- Visualization via Wind Tunnel Pro HD (iPad app): pressure, speed, vectors over drawn building profiles.
Possible Tube Placements (Section & Plan)
- Parallel, slanted, reverse-slanted, perpendicular; options:
- Exterior surface mount.
- Wall-penetrating (inside–outside sonic link).
- Ground-supported positions rejected (airflow blocked).
- Only placements ensuring unobstructed mouth exposure to wind are viable for sound.
Sound Studies
Basic Acoustics
- Sound waves = longitudinal; speed depends on conditions.
- Use constant c\;=\;340\,\text{m/s} (approx.).
- More precise: c\;=\;331.3 + 0.6\,t (with t in °C).
- Two primary tube categories:
- Open–open (both ends pressure nodes).
- Open–closed (pressure node + antinode).
- Tube with two closed ends unusable – no radiation path.
Fundamental & Harmonics
- For any tube: f = \dfrac{c}{\lambda}.
- Open–open:
- Fundamental wavelength \lambda = 2L.
- Length correction: L = \dfrac{n\,\lambda}{2} - 0.8d (end correction; d = diameter).
- Open–closed:
- Fundamental \lambda = 4L.
- Length correction: L = \dfrac{n\,\lambda}{4} - 0.4d.
- Harmonic content:
- Open–open supports all integer n.
- Open–closed supports only odd n.
Human Hearing & Design Limits
- Audible band: 20 Hz – 20 kHz; piano range 27.5 Hz – 4186 Hz chosen as practical subset.
- Selected façade tube diameter: 6 in (≈150 mm) – compromise between ease of excitation and structural scale.
- Extreme lengths derived:
- Open–open at 27.5\,\text{Hz}\Rightarrow L \approx 19.88'.
- Open–closed at same frequency \Rightarrow L \approx 9.94'.
- High piano C (1318.5\,\text{Hz}) corresponds to impractically short lengths (<0.5').
- Minimum usable length set to 2' to keep wind speed requirements reasonable.
Empirical Testing
- Air sources & measured speeds:
- Desk fan ≈ 9 mph.
- A/C vent ≈ 7.5 mph.
- Hair-dryer: low 5 mph, high 12.5 mph.
- Human blowing: ~23 mph but unrealistic for façade.
- Early piccolo-scale tubes (1/4"–1") failed → scaling up to 2"+ diameters succeeded.
- Hundreds of variants tested; only 3 selected for final grammar (see below).
- Type 1 (T = 1): Circular, open–open – medium-pitch hum.
- Type 2 (T = 2): Circular, open–closed – one octave lower.
- Type 3 (T = 3): Circular, open–closed + long transverse slot – higher pitch with whistle overtone (fundamental & 1st harmonic recorded).
- Recordings cleaned in Audacity using noise-cancel & amplification.
Spatial Relation & Basic Grammars
Pillar Abstraction
- Tube represented by square pillar in plan; two pillars placed 15° rotated form base relation.
- Pillar has 16 symmetry operations → 16 label conditions.
Basic Grammar Outcomes
- One-rule grammars (16 variants): tight/loose spirals, fans, lines, identity (no growth).
- Two-rule grammars (256 possible): patterns A–B–A–B; identity rules (labels 2,5,10,13) halt growth (unsuitable for façades).
- Three-rule grammars (4096 possible): patterns A–B–C; increased complexity, less predictable aggregations; all examined sets used non-turning labels (1,6,9,14).
Seven-Rule Sonic Facade Grammar
- Rule 1 – Locate Façade
- User marks start & end points on existing building; progression counter-clockwise in plan.
- Rule 2 – Tube Height
- Bottom ≥ 3" off grade.
- Length limits 2' \le L \le 19.88'.
- Relative to building height: equal, above, or below roofline.
- Rule 3 – Calculate Quantities
- For façade segment distance D: N = \dfrac{D}{6"} tubes.
- Count corners C.
- Rule 4 – Place First Pillar
- Offset ≥ 3" from wall.
- Initial labels: tube number Z = 1, tube type T = 1.
- Rule 5 – Aggregate Pillars
- Apply straight rule while Z \le N.
- If Z=N & C>0 apply one turn rule (pillar width sized to maintain 3" clearance).
- Tube-type labeling sequence:
- If L > 9.94' → pattern 1-2-1-2-…
- Else L \le 9.94' → 1-2-3-1-2-3-…
- Rule 6 – Color Coding
- Apply one of four possible color orientations on each pillar; drives connection-detail parametrics.
- Rule 7 – Substitute Tubes & Details
- Pillar replaced by physical tube according to T.
- Connection bracket depth s = 2.5\"; offsets: 0.1L or 0.2L for aesthetic staggering.
- Grammar supports non-determinism & user choice; accommodates openings; not yet written for curved surfaces.
Example Applications
Example 1 – “Complicated with Awning”
- Two walls: pattern (1-9-14) on short side, (1-6-9) on long side; one corner turn.
- All tubes 8' long (equal height), W = 3", G = 3".
- Visual result: undulating spiral forming awning + niches.
- Acoustic output: only 3 base sounds (equal lengths).
Example 2 – “Simple with Openings”
- Patterns use only label rule 6; differentiation via starting label positions.
- Tube lengths: 2', 3.25', 9'; varying ground clearances 3"–7.25".
- No corner tubes; façade wraps around windows/doors.
- Produces 9 distinct pitches; overall higher register.
Example 3 – MIT Green Building (Bldg 54)
- Site wind rose (Project Full Breeze) → prevailing NW wind.
- Façade located on NW wall, framing entry near Bldg 56.
- Tube length maximized 19.88'; two tube types (open–open, slot).
- Tubes 15' above grade; low, droning composite sound.
Creating Final Sounds (Composites)
- Method in Audacity:
- Layer multiple recordings of each tube type.
- Use Change Speed, Change Pitch, Amplify, Reverse per layer.
- Pitch shifts match relative tube lengths; number of layers matches count of each type.
- Tracks 12–14 correspond to Example 1–3 composites.
Discussion & Future Work
- Applicability: best for public, educational, or museum façades; domestic/office users may need optional tube caps to mute sound.
- Strengths of Grammar: adaptable, user-driven aesthetics, limited tube palette keeps sonic range intelligible.
- Limitations:
- No curved-wall logic.
- Fixed tube lengths within a run; finer musical control deferred.
- Opening creation rule could be formalized.
- Does not compute exact frequencies; author argues environment-dependent variability makes strict tuning less relevant.
- Next Steps:
- Physical mock-ups to validate acoustics.
- Extend grammar to structural elements, interior-exterior sonic dialogue, furniture scale.
Numerical & Equation Highlights
- Speed of sound (ideal): c = 340\,\text{m/s} ; temperature-corrected: c = 331.3 + 0.6t.
- Fundamental wavelengths:
- Open–open: \lambda = 2L.
- Open–closed: \lambda = 4L.
- Tube frequency: f = \dfrac{c}{\lambda}.
- Practical length bounds for 6" tubes: 2' \le L \le 19.88' (≈ 0.61 m – 6.06 m).
- Wind test air speeds: 5–12.5 mph (mechanical); 23 mph (human blow) – unrealistic.
References & Credits (Condensed)
- Major sources: Araujo & Dykes (Project Full Breeze), Fletcher & Rossing (Physics of Musical Instruments), Schafer, Kahn, Hopkin, etc.
- Image credits: Tonkin Liu, Luke Jerram, Joe Snell, Mark Nikkon, Harry Bertoia studio, Liminal.
- Sound credits: YouTube clips of Singing Ringing Tree, Aeolus, Aeolian Harp, Chimecco, Sonambient, Organ of Corti.