Schroeder Backward Integration

The energy decay curve (EDC) is computed by backward integration of the squared impulse response h(t):

EDC(t) = ∫ₜ^∞ h²(τ) dτ

Expressed in dB, the EDC shows the decay of reverberant energy over time. This method, proposed by Schroeder in 1965, provides a smoother and more reliable decay curve than the original interrupted-noise method. The upper integration limit must extend sufficiently beyond the point where the impulse response merges with the noise floor — this is where INR quality matters.

T30 — The Primary Reverberation Time

T30 is defined as the time for the EDC to decay 60 dB, estimated by fitting a linear regression line to the portion of the EDC between -5 dB and -35 dB below the initial level. The slope of this line, extrapolated to a 60 dB range, gives T30. This avoids measuring the full 60 dB decay, which would require very high SNR.

Requirement: the regression correlation coefficient r² must be ≥0.995. If r² is below this threshold, the measurement is unreliable — increase source level, reduce background noise, or use a longer averaging time.

T20 — Reduced Evaluation Range

T20 uses the -5 dB to -25 dB portion of the EDC, requiring only 20 dB of usable decay range. This means T20 can be reliably measured with an INR of just 35 dB (vs 45 dB for T30). T20 is useful in noisy environments or when using low-power sources like phone speakers. However, T20 has higher uncertainty because it samples a smaller portion of the decay.

EDT — Early Decay Time

EDT is calculated from the 0 dB to -10 dB portion of the decay curve. It characterizes the early part of the sound decay, which correlates more closely with the subjective perception of reverberance in occupied rooms. EDT is typically shorter than T30 in spaces with good early reflections and can vary significantly with position, unlike T30 which is more spatially uniform.

Octave-Band Analysis

ISO 3382-1 requires reverberation time to be measured in octave bands from 125 Hz to 4 kHz (extended to 63 Hz – 8 kHz for precision grade). The impulse response is filtered into octave bands per IEC 61260-1 before computing the EDC. Each band has its own T30, T20, and EDT value. The single-number "RT60" is often reported as the average of the 500 Hz and 1 kHz octave bands.

Typical Values

Recording studio: 0.2 – 0.4 s
Classroom: 0.4 – 0.7 s
Conference room: 0.5 – 0.8 s
Concert hall: 1.5 – 2.2 s
Cathedral: 3.0 – 8.0 s

SonaVyx RT60 Measurement

The SonaVyx RT60 tool computes T20, T30, and EDT using Schroeder backward integration with Lundeby truncation to handle the noise floor. Results include per-octave-band values, regression r² quality indicators, and comparison to room-type targets. The impulse response tool shows the EDC plot directly. For clarity and definition parameters derived from the same impulse response, see the next section.