| Field | Value |
|---|---|
| Project | Refayee Masjid full-masjid audio + acoustic design (RIF-2026-001) |
| Scope | Whole masjid: main hall, extension, mezzanine, riwaqs ×3, women's musalla, mihrab/adhan, courtyard, minaret, basement tie-in. Basement Nawawi PA already installed (Next Audiocom W6W) — bridged, not rebuilt. |
| Confidence convention | ⚑EST estimate pending survey/architect · INDICATIVE pre-gate figure · SOLID traces to DXF title-block or confirmed physics · GATE must clear before freeze |
| Basis of all α / material data | textbook / placeholder — confirm on-site finishes + product datasheets before BOQ |
The masjid. A large domed Friday mosque in Kafr Sousa, Damascus: four levels (basement / ground / mezzanine / roof), ~2,150 worship capacity, a 1,142 m² main prayer hall (الحرم) under a tall bulbous pointed (onion) Ottoman dome Ø≈21.5 m, plus a 580 m² extension, a 500 m² mezzanine (السدة), three riwaqs, a women's musalla, a 500 m² courtyard, and a minaret. The basement Nawawi hall already has a working PA (local-vendor rush install); this study covers everything else and ties the basement in.
The concave onion dome is the keystone of the whole design. An earlier 2-D ray-trace suggested a short 47 ms / −6 dB return; independent engineering review refined the dome geometry and established the correct result. The corrected trace shows the dome throws a distinct ~150–200 ms LATE ECHO off the crown back down onto the floor — a discrete slap, worse for intelligibility than any Haas-region arrival, at roughly −10 to −16 dB below direct in the 2-D model (treat −13 dB as a floor — the real 3-D caustic concentrates a point focus and will be louder). This is the single most important finding in the project and it changes the fix.
The earlier framing ("you can't absorb the dome, so it's a pure directivity problem") was set aside on independent engineering review as a false dichotomy. The correct fix has four coordinated moves:
Build the hall on DSP beam-steerable columns — the only class that gives a slim, heritage-discreet, electronically re-aimable vertical null on the dome plus an independent mezzanine beam. EAW (raised as a candidate during review) is the right philosophy (its Adaptive tech is genuine beam-shaping) but the wrong product for the qibla wall — too large/heavy for a heritage dome hall, and EAW has no slim steerable column. Spec EAW as point-source fills (riwaqs, mezzanine soffit, courtyard), where it genuinely excels.
Areas trace to the architect title-block schedules, double-confirmed across two DXF exports (zero delta). Scale resolved: DXF is centimetres (×0.01 m); the 400-unit module = the 4.00 m column grid; hall ≈ 34 m consistent with 1,142 m².
| Zone (AR / EN) | Area m² | Level | Confidence |
|---|---|---|---|
| الحرم — main prayer hall | 1,142 | ground | area SOLID · H ⚑EST (~10 m clear/cornice) |
| توسعة الحرم — extension | 580 | ground | area SOLID · H ⚑EST |
| الصحن — courtyard (open) | 500 | ground | outdoor; SPL/coverage only |
| الرواق القبلي / الشرقي / الغربي | 232 / 165 / 160 | ground | side/rear fills |
| مصلّى النساء — women's musalla | 28 → ~68 | ground | small fill; 68 m² envelope = cheap insurance |
| المحراب وغرفة الأذان — mihrab + adhan room | 17.2 | ground | highest feedback/mic sensitivity |
| السدة — mezzanine | 500 | mezzanine | independent delay/level |
| قاعة الإمام النووي (basement) | 90 / 49.3 | basement | PA installed (Next W6W) — sched vs deployed to reconcile |
| قاعة الإمام البخاري (basement) | 122 | basement | future zone |
| قبو التوسعة (basement) | 485 | basement | future zone |
| مئذنة — minaret | 11.2 | — | external adhan horns |
Combined hall + extension = 1,722 m². Worship capacity ≈ 2,150 (~0.8 m²/person).
Cross-checked and refined through independent engineering review. The crown returns cluster at 150–210 ms (imam @ mihrab 154–206 ms, median ~195 ms; under-dome ~160 ms), deep in the “distinct late echo — worst case for STI” region. The earlier 47 ms estimate sits far to the left, shown for comparison. Click to enlarge.
The refinement: an early trace suggested a short 47 ms / −6 dB return computed on a simplified drum geometry. Independent engineering review re-examined the dome section, modelled the drum open, and established the true crown path — a materially different and more critical result.
The validated finding: with the drum modelled open, the crown reflects the imam’s rising energy straight back down onto the floor as a distinct late echo — imam @ mihrab 154–206 ms (median ~195 ms); under-dome ~160 ms.
Level: ≈ −10 to −16 dB below direct in the 2-D model (long crown path). 2-D under-models a point focus; the real 3-D caustic concentrates it louder. Treat −13 dB as a floor, not the answer — the 3-D gate (§9) sizes the true focal level.
Why it is worse than the old number: a 47 ms arrival sits near the Haas/C50 boundary — borderline useful. A ~195 ms arrival is unambiguously late — a discrete slap the ear resolves as an echo, degrading intelligibility and (in a reverberant stone hall) smearing consonants.
A provisional 3-D axisymmetric ray-trace confirms the crown echo (imam ~150 ms, under-dome ~185 ms) plus an on-axis caustic on estimated heights. Re-run on surveyed apex/springing to FREEZE the focal level — the 2-D / estimate under-models the caustic. Click to enlarge.
To bury a −13 dB echo to −22 dB by whole-shell absorption alone needs α ≥ ~0.87 across the entire dome interior — physically impossible on a heritage onion shell whose developed surface is ≥ ~726 m² (hemisphere lower bound; the onion profile is materially larger, so this only makes pure absorption harder). That arithmetic is correct (energy form, 10·log). But the engineering conclusion is explicit: this is a reductio that rejects whole-shell absorption — NOT proof that absorption is useless. The right answer is the four-part hybrid (§1, §3). Framing it as "directivity-only" was set aside on independent engineering review.
Room modes sit at ~25–45 Hz, below the speech band. Schroeder frequency is low; bass-trapping is not the priority. Invest absorption in 500 Hz–4 kHz (the consonant band) and put the directivity budget into keeping speech energy off the dome. Every panel spec below is optimised for 500 Hz–4 kHz.
03 · ACOUSTIC TREATMENT PACKAGEα = octave-band, textbook — placeholder pending product datasheets.
| Code | Product | α [125/250/500/1k/2k/4k] | NRC | Role |
|---|---|---|---|---|
| BB1 | 50 mm mineral wool (60–70 kg/m³), fabric-wrapped, 25 mm air-gap | 0.20/0.55/0.85/0.95/0.95/0.90 | ~0.85 | Broadband wall panels (rear/side mid-band workhorse) |
| BB2 | 100 mm mineral wool (48–70 kg/m³), fabric-wrapped, rigid backing | 0.45/0.85/1.00/1.00/0.98/0.95 | ~0.95 | Drum-ring segments — broadband interception + durability (in-band 500–4k ≈ BB1) |
| AP | Acoustic plaster / lime-compatible spray (BASWA / Pyrok class) | ~0.30/0.55/0.70/0.75/0.72/0.70 | ~0.70 | Drum band + rear targeted reversible absorption; matches historic lime finish |
| DIFF | Muqarnas / coffering / geometric relief (Damascene plaster/GRG) | 0.05/0.08/0.12/0.15/0.15/0.12 | ~0.12 | Crown scatter — breaks the ~150–200 ms echo 10–15 dB; locally fabricable, import-immune |
| DRP | Heavy cotton/wool velour ~0.5 kg/m², 50–100 % fullness, spaced 100 mm off glass | 0.10/0.35/0.55/0.72/0.70/0.65 | ~0.60 | Adjustable drapes at ground-level glazing |
| CPT | Wool loop carpet on 8–10 mm bonded-felt underlay | ~0.15/0.35/0.60/0.72/0.72/0.75 | ~0.55 | Floor — dominant absorber; underlay's real uplift is 125–250 Hz |
| CLD | 50 mm mineral-wool ceiling cloud/raft, fabric-wrapped | 0.30/0.65/0.90/0.98/0.95/0.90 | ~0.90 | Enclosed mezzanine soffit only |
| CLD-W | Hydrophobic mineral wool behind perf-metal/micro-perf, or slotted-timber broadband | ~0.25/0.55/0.80/0.90/0.88/0.85 | ~0.80 | Semi-open riwaq + courtyard soffit — weather/humidity-rated |
| Sub-zone | Treatment | Location | ~Area | Rationale |
|---|---|---|---|---|
| Crown / lower dome + squinch | DIFF (muqarnas/coffering relief) | Springing-to-mid-dome band; apex left decorative | ~150 m² | Scatters the ~150–200 ms crown echo 10–15 dB. Shell absorption rejected (α≥0.87 impossible). Locally fabricable. |
| Drum ring | BB2 segments + AP between arched windows | Cylindrical drum band, ~16 m up, 360° between glazing | ~85 m² | Intercepts dome-return energy at the neck before it reaches the shell — highest-leverage absorption in the building. Windows stay clear. |
| Rear + upper side walls | BB1 fabric-wrapped, upper band above sightline | Rear + both side walls, ~4 m band above ~3 m | ~200 m² | Kills rear-wall slap + long lateral tail; mid-to-rear placement preserves front support and first-row STI. |
| Front / qibla wall | Provisional light — MODELED, not assumed | Front third | ⚑ TBD | Do not auto-assume "keep it live." Diffusion if it combs/feeds back; zero if it supports recitation. Gate: qibla-wall model (§9). |
| Floor | CPT (wool carpet + felt underlay) | Full hall | 1,142 m² | Dominant absorber; underlay lifts 125–250 Hz (existing; underlay = disruptive upgrade). |
| Large glazing | DRP heavy drapes | Ground-level windows/entrance glazing only (not drum) | ~120 m² | Tames hard glass at ear height; adjustable Ramadan-vs-daily lever. |
Extension توسعة الحرم (580 m²): BB1 rear + outer side upper band ~120 m². Mezzanine السدة (500 m²): engine already predicts RT60 ≈ 0.58–0.68 s occupied — do not over-dry; CLD clouds ~60 m² only. Riwaqs ×3 الأروقة: CLD-W (hydrophobic) rafts ~180 m² total (≈75/55/50) — fabric-wrapped wool would sag/mold in a humid Damascus arcade. Women's musalla: BB1 ~30 m². Mihrab + adhan المحراب: niche untreated (natural horn); BB1 ~20 m² behind the mic to raise gain-before-feedback. Courtyard الصحن: CLD-W soffit strip ~40 m². Basement: no action this package.
Occupied hall, Fitzroy + dome basis. Indicative pre-gate figures; final STI must come from a modeled impulse response, and the STI cells are contingent on the PA directivity design, not an absorption output.
| Scenario | 250 Hz | 500 Hz | 1 kHz | 2 kHz | Speech-avg RT60 | STI (indic.) | Verdict |
|---|---|---|---|---|---|---|---|
| Carpet only (no panels) | ~5.5 | ~4.8 | ~4.5 | ~4.0 | ~4.5–6.0 s | <0.40 | FAILS — echo + long tail |
| S6 moderate (this package) | ~1.7 | ~1.45 | ~1.35 | ~1.25 | ≈1.38 s | ≈0.67 | ✅ DESIGN INTENT |
| S4 full-treat (add shell/side abs., front-wall) | ~1.0 | ~0.9 | ~0.85 | ~0.80 | ≈0.88 s | ≈0.75+ | ❌ over-dry + heritage-intrusive |
Choose S6. It lands squarely in the 1.2–1.5 s target with intelligible speech and recitation that keeps its tarteel bloom. S4 over-dries (a mosque at <1.0 s sounds clinical), demands absorption on the front wall and shell (violating rules 1 and 3), and buys an STI gain (0.67→0.75) that is inaudible for amplified speech. The residual ~150–200 ms echo at S6 is handled by directivity + crown diffusion, not by grinding RT to 0.88 s.
| Item | Zone | Qty | Purpose |
|---|---|---|---|
| DIFF muqarnas/coffering relief | Hall crown (lower dome + squinch) | ~150 m² | Scatter the ~150–200 ms echo; locally fabricable, NOT absorption |
| BB2 100 mm drum-ring segments | Hall drum band | ~85 m² | Intercept dome-return at the neck |
| AP acoustic plaster | Hall drum band + rear | ~90 m² | Reversible targeted absorption, lime-compatible |
| BB1 50 mm wall panels | Hall rear + side upper band | ~200 m² | Kill rear slap + lateral tail |
| BB1 wall panels | Extension rear/side | ~120 m² | Absorb extension tail |
| CLD ceiling clouds | Mezzanine soffit + rear | ~60 m² | Local reflection control (light) |
| CLD-W weather-rated rafts | Riwaqs ×3 + courtyard soffit edge | ~220 m² | Overhead flutter control, humidity-safe |
| BB1 panels | Women's musalla / Adhan room (behind mic) | ~30 / ~20 m² | Small-room intelligibility / raise gain-before-feedback |
| DRP heavy drapes, adjustable | Ground-level glazing (hall+ext+entries) | ~120 m² | Tame glass; variable-acoustic lever |
| CPT wool carpet + felt underlay | Hall+ext+mezz+women floors | ~2,250 m² | Dominant absorber; disruptive underlay upgrade |
| Qibla/front wall + mihrab niche | Hall + mihrab | modeled | Do not auto-assume; gate on qibla model |
You cannot absorb the dome shell. So the loudspeaker must do what treatment cannot: launch a wavefront sharply bounded in the vertical plane — full level at seated ear height (~1.2 m), falling off a cliff above head height so almost nothing climbs into the drum windows or the shell. Horizontal width is the easy axis; the vertical null above the congregation is the whole game. Two device properties deliver it: (1) a tall radiating aperture (a column metres long) for high vertical DI at 250 Hz–2 kHz; (2) DSP beam-steering so the beam is aimed and shaped in software and re-aimed after measurement without re-rigging on heritage stone.
The model's ~20 dB dome-off-axis suppression (omni −16 dB → steered −36 dB) is an idealized figure. It must be proven with the chosen array's real GLL/polar data, octave band by octave band. A column that holds Q 8–16 at 500 Hz–1 kHz but sprays 2–4 kHz will still light the dome. No PO until the GLL confirms it (§9).
Essentially invented for reverberant stone worship spaces with domes and balconies.
| Product line | What it is | Fit for the dome hall |
|---|---|---|
| Renkus-Heinz IC Live X (ICL-FR / ICL-F) & Iconyx Gen5 (IC7 / IC16 / IC24-RN) | Self-powered, network-controlled steerable arrays; onboard DSP; multiple independent beams from one array (BeamWare + RHAON II). Iconyx Gen5 = high-resolution install line. | Best-in-class for this problem. One front array throws a lower beam across the floor and a separate upper beam to the mezzanine while placing a null on the dome between them. Primary hall recommendation. |
| Fohhn Focus Modular / Linea Focus | German; real-time electronic steering (Fohhn-Net), re-aimable live. Superb build, very high output. | Co-primary / premium alternative. Steering and finish top-tier; slightly higher cost, thinner regional spares. |
| Kling & Freitag VIDA | German DSP electronically-steered line array. | Genuine competitor on steering — carry it in the head-to-head. (Corrects an earlier draft that misspelled the brand and wrongly said K&F has no steering.) |
| Pan Acoustics Pan Beam | German DSP-steered columns, lower cost. | Strong value engineered-tier option for the extension and secondary zones; a notch below RH/Fohhn on output/beam sophistication. |
Why it wins: the dome null and the mezzanine each want a separately aimed beam. Only electronic multi-beam gives you that from a single visually-slim column on a stone pier — heritage restraint and acoustics align. Value tier for fills (not the hall spine): RCF (VSA is self-powered steerable; C-series passive), Community ENTASYS, Apart COLS — useful as delayed fills, not the primary hall source under the dome.
EAW is a serious, historically pattern-control-obsessed brand (Jericho horns, Radius). The friend's instinct isn't wrong in spirit. EAW's genuinely relevant tech — Adaptive (ANYA / ANNA / OTTO): large-format arrays whose coverage is defined in software and whose DSP shapes the wavefront across dozens of transducers. This is real beam-shaping, conceptually in the same family as what a dome room needs. Credit where due: this is EAW's legitimate seat at the directivity table.
Why Adaptive is nonetheless the wrong fit here:
Where EAW genuinely fits and I'd happily spec it: point-source fills — EAW QX/KF (superb horn pattern control, install-tough) and MK/RSX make excellent riwaq, mezzanine-soffit, and courtyard fills. A large modern courtyard main or arena-scale overflow → Anna would be a top-3 pick. The pattern-control DNA lives inside the fill recommendation: Fulcrum Acoustic (coaxial fills) was founded by EAW's former chief engineer.
Bottom line: EAW's Adaptive is real beam-shaping and EAW is a directivity-first brand — but for the dome hall spine, EAW does not compete with the beam-steering column specialists (Renkus-Heinz, Fohhn, K&F VIDA) on the one axis that decides this project. Use EAW for point-source fills, not for the qibla-wall mains.
Hall ≈ 34 m deep (SOLID from area) × ~33.6 m wide (⚑ derived, near-square inference, not given). Steering lets a front array serve two throw-zones (near-floor beam + far/mezzanine beam), reducing box count.
| Throw-zone | Covers (m from qibla) | Array @ | Delay | Loudspeaker role | Eng qty | Prem qty |
|---|---|---|---|---|---|---|
| Z1 front (qibla) | 0–11 | 0 m | 0 ms | Steerable columns flanking mihrab + across width; lower beam floor, upper beam grazes Z2 & feeds mezzanine | 4× RH IC Live X | 6× RH Iconyx Gen5 / Fohhn Focus |
| Z2 mid | 11–22 | ~11 m | 47 ms | Upper beam of Z1 + a delayed steerable/high-Q ring on the ~11 m line | 4× high-Q column (RCF VSA/C) | 4× RH IC Live X |
| Z3 rear | 22–33 | ~22 m | 79 ms | Delayed steerable/high-Q columns on rear grid | 4× high-Q column | 4× RH IC Live X / Fohhn Linea |
| Z4 back-wall lip | 33–33.8 | ~33 m | 111 ms | Low-level delayed point-source to erase back-wall slap (a ~0.8 m slap zone, not an 11 m coverage zone) | 2× coaxial point-source (Fulcrum) | 2× premium coax |
Subwoofers — speech-first → none required. Optional 2× compact sub for recitation warmth only. If kept, band-limit ~45–50 Hz HPF (above the 25–45 Hz modes) + ~90–100 Hz LPF handing off to the mains — not a 90 Hz HPF (which strips exactly the band a sub exists for). Cleaner still: drop it.
| Zone | Area / notes | Recommended class | Directivity justification |
|---|---|---|---|
| Extension توسعة | 580 m² | Delayed steerable/high-Q columns, time-aligned. Eng: RCF VSA / Pan Beam · Prem: RH IC Live X / Fohhn / L-Acoustics Syva | Continues off-dome vertical control into the contiguous volume |
| Mezzanine السدة | 500 m², independent delay | Upper beam of Z1 + under-soffit coaxial point-source + front-rail fills | Low, tightly-controlled soffit sources keep the balcony from over-driving the mains into the dome |
| Riwaqs ×3 | 232/165/160 m² | Coaxial point-source, off-axis from piers, locally delayed. Eng: Fulcrum / EAW QX-MK · Prem: d&b / Fulcrum | Short throw behind colonnades; controlled conical pattern. EAW fits well here |
| Women's musalla | 28→68 m² | 2× compact point-source or 1 small column | Even low-level fill; no feedback path to imam mic |
| Mihrab + adhan room | 17.2 m² — highest feedback risk | 1–2× small isolated confidence monitor, not in the main PA loop | Reciter hears without opening a feedback loop |
| Courtyard الصحن | 500 m² outdoor | Weatherproof (IP54+) point-source/column. Eng: Community/RCF/EAW weatherized · Prem: L-Acoustics/d&b | Fixed downward pattern acceptable outdoors (no dome); durable/serviceable |
| Minaret مئذنة | adhan broadcast | 2–4× high-output weatherproof horns (RCF HD-series) | Max intelligibility-per-watt outward, not back at the structure |
| Parameter | Value | Source |
|---|---|---|
| Speed of sound c | 343 m/s (20 °C) | standard; ⚑temp-sensitive (§5.6) |
| Design SPL interior / courtyard | 78 dBA (75–80) / 80 dBA | brief §6 |
| Headroom | +12 dB → 90 dB peak interior | brief §6 |
| Uniformity | ±3 dB pref / ±6 dB accept | brief §6 |
| STI | > 0.60 primary / > 0.50 secondary | brief §6 |
| Haas bias | +15 ms per downstream zone | prior peer review |
| Hall volume V | ≈ 11,441 m³ (1142 × ~10 m) | ⚑EST height |
| Occupied RT60 (S6) | ≈ 1.38 s (band 1.2–1.6) | brief §3, indicative |
Point source (compact box/cluster/ceiling tap): loss = 20·log₁₀(d) — 6 dB/doubling. Line / column / steered array (near field): loss = 10·log₁₀(d) — ~3 dB/doubling. A single front cluster @ 33.8 m point-source = 117.6 dB@1m ✓; a delayed column @ 11 m line-source = 97.4 dB@1m ✓. This is why a single cluster is unusable and the zoned delayed spine is mandatory.
| Zone | Covers | Far throw | SPL@1m (line, lower) | SPL@1m (point — SIZE TO THIS) |
|---|---|---|---|---|
| Z1 front | 0–11 | 11 m | 100.4 dB | 110.8 dB |
| Z2 mid | 11–22 | 11 m | 100.4 dB | 110.8 dB |
| Z3 rear | 22–33 | 11 m | 100.4 dB | 110.8 dB |
| Z4 back lip | 33–33.8 | ~2–3 m | ~93 dB | ~98 dB |
Formula: SPL@1m = 78 + 12 + k·log₁₀(d), k = 10 (line) / 20 (point). For 75/80 dBA: −3 / +2 dB uniformly. Amp/driver read (corrected — no double-count): at 92 dB/W/m, continuous ≈ 0.4 W, peak ≈ 7 W (line) / ≈ 76 W (point ceiling). The earlier draft counted the 12 dB headroom twice. Plan ~250–400 W amp channel per hall zone-array (⚑assumes 92 dB sensitivity — confirm on datasheet).
Arrays at 0 / 11 / 22 / 33 m from qibla. Delay = (distance from imam ÷ c) + 15 ms Haas.
| Zone | Array @ | Propagation | +15 ms | DSP delay |
|---|---|---|---|---|
| 1 | 0 m | 0.0 ms | ref | 0 ms |
| 2 | 11 m | 32.1 ms | +15 | 47 ms |
| 3 | 22 m | 64.1 ms | +15 | 79 ms |
| 4 | 33 m | 96.2 ms | +15 | 111 ms |
Verified at the ear (the real Haas test): a worshipper at 36 m hears the imam's direct wavefront at 105 ms; the zone-4 array (fires at 111 ms, +8.7 ms to the listener) arrives at ~120 ms = +15 ms after direct → inside the Haas window, localises to the imam, no perceived echo. ✓ Delay design and directivity are the same fight — arrays aimed into the congregation and off the shell, delays kept inside the Haas window so downstream energy lands early rather than feeding the crown echo. The dome's ~150–200 ms return is the villain we starve of energy.
Dc = 0.057·√(Q·V/RT60), V = 11,441 m³.
| Q (DI) | Dc @ RT 1.6 | @ 1.5 | @ 1.4 | D/R at 11 m |
|---|---|---|---|---|
| 1 (0 dB) | 4.82 | 4.98 | 5.15 | −7 dB (fail) |
| 8 (9 dB) | 13.6 | 14.1 | 14.6 | ≈ +2 dB |
| 16 (12 dB) | 19.3 | 19.9 | 20.6 | ≈ +5 dB |
| 32 (15 dB) | 27.3 | 28.2 | 29.2 | ≈ +8 dB |
Design target: Q 8–16 (DI 9–12 dB) across 250 Hz–4 kHz → Dc 13.6–19.3 m → D/R +2 to +5 dB at the worst (11 m) seat. Net: Q 8–16 steered columns clear STI > 0.6 in the 11 m zones — and the same tight vertical pattern keeps energy off the dome. One spec, both problems. ⚑ The DI-holds-Q-8-16-across-band claim is exactly what the GLL gate (§9) must prove.
Uniformity: a single array set for 78 dBA at 11 m over-levels its near seats — mitigate by overlapping adjacent zones + amplitude-shaded/steered columns; verify ±3 dB by EASE before sign-off. Temperature: c = 331.3 + 0.606·T. A packed summer hall at ~28–30 °C → c ≈ 348–350 m/s, shortening every propagation time ~1.5 %. The DSP must expose per-zone delay for warm-field re-tuning; commission at worship temperature, not empty-and-cold.
| # | Zone | Class | Throw | SPL@1m (size to) | Delay | Confidence |
|---|---|---|---|---|---|---|
| 1 | Hall Z1 (0–11) | steered column | 11 m | 110.8 dB | 0 ms | delay SOLID |
| 2 | Hall Z2 (11–22) | steered column | 11 m | 110.8 dB | 47 ms | delay SOLID |
| 3 | Hall Z3 (22–33) | steered column | 11 m | 110.8 dB | 79 ms | delay SOLID |
| 4 | Hall Z4 (33–33.8) | column/fill | ~2–3 m | ~98 dB | 111 ms | delay SOLID |
| 5–6 | Extension A / B | steered column | ~9 m | 109 dB | ~114 / ~137 ms ⚑ | ⚑pos |
| 7–8 | Mezzanine rail + ceiling taps (×~30) | column + point | ~2.5–6 m | 105.6 / ~98 dB | ~81 / ~90–130 ms ⚑ | ⚑geom |
| 9 | Women's musalla (68 m²) | 4 fills | ~8 m | 108 dB | ~55–75 ms ⚑ | ⚑pos |
| 10–12 | Riwaq qibli / east / west | side-fill column | ~9–10 m | 106.5–107 dB | ~40–90 ms ⚑ | ⚑pos |
| 13–14 | Courtyard C1 / C2 | WR column | 11 m | 112.8 dB (80 dBA) | ~15–80 ms ⚑ | ⚑pos |
| 15 | Minaret adhan | LT horn | 200 m | ~40–48 @200 m* | own scene | SOLID* |
*Minaret: 48 dB @ 200 m is geometric spreading only; add ~5–15 dB HF air-absorption + ground/wind → real adhan level is lower. Confirm neighbourhood SPL bylaw. Rule for every ⚑ delay: delay_ms = (distance_of_array_from_imam_m ÷ 343) + 15, re-measured at occupied worship temperature. Lock only after the qibla-referenced site survey.
Chain: mic → preamp (−18…−20 dBFS) → A/D → DSP trim → matrix → per-zone delay → PEQ → limiter → output trim → networked amp. Preamp gain first, digital gain last. HPF 80–120 Hz on all speech inputs. No global presence boost — tune the 2.5–4 kHz consonant band per zone by measurement. Per-zone limiters (driver protection + ~90/92 dB SPL cap). Feedback margin ≥ 6 dB before ring; earset wired PRIMARY. ⚑ Feedback is computed from the REAL loudspeaker→imam-mic transfer path (measured), not inferred from the dome focus.
Primary recommendation: QSC Q-SYS (software-defined, Core-based). True input×output matrix + logic + UCI in one platform — the "Dars local-vs-broadcast" routing and all-call override are native logic, not bolt-ons (the client's #1 functional requirement).
| Platform | Fit | Trade-off |
|---|---|---|
| QSC Q-SYS Core ⭐ | Best matrix+logic+UCI+scale; strongest tablet control | Ecosystem lock-in; US-origin (export gate); needs a Q-SYS integrator |
| Symetrix (Radius NX / Prism) | Excellent value, rock-solid Dante DSP, simple to service | Control UI less rich; logic more manual; US-origin |
| BSS Soundweb London (BLU) | Proven mosque workhorse; UK-origin → survives if US licensing blocks US gear | Older paradigm, weaker modern tablet UCIs |
Inputs — provision 24, populate ~16: imam earset (WIRED, primary); imam wireless backup (SY-TPRA gated); khutbah gooseneck; recitation/adhan (hypercardioid, no LDC); Dars mics ×3; women's; paging; streaming; media; recording return; spares. Outputs — provision 24, populate ~18 (≈18 live / 24 provisioned, ~40 % future headroom). Transport: Dante primary backbone + AES67 interop; one PTPv2 grandmaster (the Core); Dante redundancy on the ground↔basement↔mezzanine trunk; audio on its own VLAN, managed PoE switches (QoS/DSCP for PTP).
Emergency/Paging (P1) > Adhan (P2) > Khutbah/all-call (P3) > local Dars (P4)
A sheikh mid-lesson can never block the adhan; the adhan can never be blocked by operator error. Scenes (per-zone delay/PEQ/limiter/level/mute/route): Daily الفرض · Jumu'ah الخطبة · Lesson/Dars-local درس · Dars-broadcast بث · Ramadan/Tarawih + Courtyard · Adhan · Courtyard-only · Emergency. Delays (0/47/79/111 ms + downstream) baked into scenes, not touched live.
Amps: ~5× 8-channel networked amps (QSC CX-Q / Powersoft Quattrocanali / Lab.gruppen) across ground/mezzanine/basement racks; mixed low-Z + 70V (70V ONLY where genuinely right — distributed soffit/women/arcade fills + minaret). Live subtotal ≈ 28 ch → +40 % → spec ~40 ch. Mics — feedback-first, RF-gated: imam wired earset PRIMARY (close-mic = max gain-before-feedback, no RF lifeline risk); RF-agile digital wireless backup only; NO large-diaphragm condenser on the PA path.
No wireless ordered until the band clears the Syrian regulator (confiscation/interference risk). The system MUST commission fully on the wired earset alone; wireless is additive. Wired fallback is a requirement, not a nicety.
A physically separate, DSP-independent path — a dedicated dynamic bypass mic at the mihrab (⚑ corrected: do not passively split the phantom-powered condenser earset — if the failed rack was supplying phantom, the split leg goes dead in the one scenario bypass exists for; use a self-powered dynamic mic) → standalone analog mixer-amp → core hall speakers, selected by a manual A/B relay on the rack face. Independent power leg (own small UPS). Tested monthly; technician trained.
Power: online double-conversion UPS on every rack; AVR/voltage stabiliser + conditioning ahead of the UPS (220 V/50 Hz, grid swings hard); sequenced power-up (Core & network first, amps last); UPS rides the generator transfer gap. Control: wall panel (operator/muezzin — scenes + health), Dars wall plates (sheikh — source + level in a locked range), tablet UCI (head tech — all scenes/trim/diagnostics), admin console (PIN — full DSP). Admin-vs-operator lock enforced in software (addresses the #1 mosque-system failure mode: well-meaning tampering).
Basement (Nawawi, Next Audiocom W6W) — do NOT rip out. ⚑ Corrected tie-in: a local-vendor rush install is almost certainly plain analog — there is no digital transport to "AES67-bridge." Land a Dante/AES67 I/O frame in the basement rack and feed the existing W6W amp from that frame's analog line output (and take an analog line in from the Nawawi Dars mic). Result: Nawawi becomes zone #N in the matrix — locally self-driven as today, now flaggable to all-call and recallable in scenes — without re-cabling the room. Describe it as an analog line I/O feed pending the existing-system inventory, not an AES67 bridge. ⚑ Reconcile the Nawawi 90 m² sched vs 49.3 m² deployed discrepancy before trusting it in overflow scenes.
Backbone: Cat6A intra-level and rack-to-rack < 90 m; OM3/OM4 fibre for long/inter-level trunks (ground↔basement, ↔minaret, ↔courtyard — beats copper distance, immune to ground-loop/EMI and minaret lightning). Redundant Dante trunk on separate switches. Add ~20 % spares + 10 % contingency to the infrastructure BOM (DSP core, 3× I/O frames, 5× 8-ch amps, 3× PoE switches, cabling/fibre, mics, wall/tablet control, 3× racks, per-rack UPS + AVR, analog bypass kit, spares kit, commissioning + training).
07 · INVESTMENTNo BOM or STI figure is frozen — and no PO is cut — until these clear. Present every pre-gate figure as indicative.
| Quantity | Current (estimate) | Confirmed | Source/date |
|---|---|---|---|
| Dome apex above floor | ~30–34 m ⚑EST | ______ | |
| Dome springing height | ~16 m ⚑EST | ______ | |
| Hall clear/cornice height | ~10 m ⚑EST | ______ | |
| Hall L × W | ~34 × ~33.6 m (33.6 derived) | ______ | |
| Column bay (scale cert.) | 4.00 m (cm-scale) | ______ | |
| RT60 occupied (measured) | Fitzroy est ~5.9 s carpeted | ______ | |
| Ambient / NC | TBD | ______ | |
| Confirmed EAW model + GLL | — | ______ | |
| SY-TPRA cleared bands | — | ______ |
| Phase | Gate to enter | Work | Deliverable |
|---|---|---|---|
| P0 · Mobilise | RFI-001 follow-up + client go | Site visit (Damascus): laser survey (apex/springing, heights, one bay), impulse RT60/STI at 6–10 points, ambient-NC, RF scan, existing-system inventory | Measured survey pack; frozen geometry |
| P1 · Confirm the dome | Frozen geometry | 3-D ray-trace on surveyed dome; confirm the ~150–200 ms echo + focal level; coupled-volume decay; STI from modeled IR; qibla-wall reflection model | Ray-trace report; STI baseline; hybrid-fix validation |
| P2 · Design v1.0 | Ray-trace done | Final zoning, array selection + GLL proof of dome-off-axis rejection, coverage/uniformity maps, DSP schematic, delay/EQ map, treatment set, feedback transfer-path check | Design v1.0 + BOM (now freezable) |
| P3 · Procure | Design v1.0 + budget + export/SY-TPRA gates | Vendor quotes, duties/freight, pre-order long-lead items (mineral-wool densities, acoustic plaster, arrays), commission local muqarnas/GRG craft | Final BOM + procurement spec |
| P4 · Install | Equipment on site | Heritage-safe mounting (non-invasive brackets), cabling, rack build, basement analog tie-in, treatment install (crown diffusion + drum band + rear + drapes) | Installed system |
| P5 · Commission & accept | System installed | Tune by measurement (0/47/79/111 ms + downstream, DI/Q verify, feedback margin ≥ 6 dB, STI grid, scene authoring), analog-bypass drill, operator/technician training | Commissioning report + acceptance; STI/SPL/uniformity sign-off |