Purpose
Assess diffuse reverberation (RT60, clarity) and modal ringing (waterfall ridges, modal analysis). In small domestic rooms, RT60 below ~300Hz is not meaningful — that's modal decay, not diffuse field. Topt is more reliable than T20/T30 there.
Numbered Workflow
- Measure: Full-range sweep, 256k+, with IR truncation long enough for room (increase in Analysis preferences for large rooms). Need decay to 60dB below peak or noise floor corrupts RT60.
- RT60: Open RT60 → select octave or 1/3-octave → filter order 6. View EDT, T20, T30, Topt, and correlation. Topt (optimized fit) is often best in small rooms. Compare to targets: <50m³ ≈0.3s, 50–200m³ 0.4–0.6s, fairly uniform vs frequency.
- Clarity: Open Clarity → C50 (speech), C80 (music), D50. Higher = more early energy = clearer. Low C50 = too reverberant or strong late reflections.
- RT60 Decay: Open RT60 Decay → Generate → Calculate RT60 model. This is frequency-domain exponential fit (high resolution). View T60M overlay — shows T60 vs frequency continuously.
- Spectral Decay / Waterfall: Generate. Look for horizontal ridges that persist in time = resonances. Choose Fourier (linear time) or Burst Decay (periods axis — better for comparing Q across frequencies).
- Spectrogram: Time-frequency color map. Modes: Fourier, Wavelet, Airy CWT, Morlet CWT, Burst Decay. Wavelet/CWT give better time-frequency tradeoff than pure Fourier.
- Modal Analysis: EQ → Modal Analysis → enter room dimensions → Find Resonances. Lists resonance frequencies, SPL, T60. Correlate with waterfall ridges.
Worked Example — Assess Room Reverb & Decide Treatment
Measurement: 4×5×2.4m living room. RT60 Topt: 0.65s at 500Hz, 1.1s at 63Hz. Waterfall shows ridge at 43Hz and 68Hz lasting 800ms. C50 = 2dB.
Interpretation: Mid/high RT60 too high (needs absorption/diffusion). LF long decay is modal — not RT60, but ringing. EQ can cut 43Hz peak but won't shorten 800ms decay as effectively as bass trapping. C50 low confirms poor clarity.
Action: Broadband absorption + bass traps in corners, then re-measure and apply modest modal EQ.
Outputs & Interpretation
- RT60 per band: Should be uniform. Rising at LF = modal, not diffuse. Don't chase RT60 below Schroeder frequency.
- T60M: Continuous T60 vs frequency — reveals narrow resonances RT60 bands average over.
- Waterfall ridge: Frequency of ridge = mode. Length = T60. Use Burst Decay to compare Q: same periods = same Q regardless of frequency.
- CSD mode: Anchors right window, moves left — shows how response builds as more time is included. Different from normal decay.
Deeper Detail — Schroeder, Lundeby & Why RT60 Fails at LF
Schroeder integral: Backwards integration of squared IR: E(t) = ∫t∞ h²(τ)dτ. Slope of log E(t) = decay rate. Lundeby method finds optimal truncation point to exclude noise floor.
Why RT60 is invalid at LF in small rooms: RT60 assumes diffuse field (many overlapping modes). Below Schroeder frequency (~200Hz in domestic rooms), modes are sparse and decay is modal, not statistical. Use waterfall/modal analysis instead.
Burst Decay: X-axis in periods (cycles) rather than ms. A resonance with Q=20 will show same period extent at 50Hz and 500Hz — direct Q comparison.
Frequency-domain RT60: Fits exponential decay + noise model per frequency bin — higher resolution than octave bands, good for finding narrow resonances.
Check Your Understanding
Small bedroom RT60 T30 at 63Hz reads 1.4s, but waterfall shows a single ridge at 58Hz. What's the correct interpretation?