Choose the excitation signal for your impulse response measurement.
The room impulse response (RIR) is the most fundamental measurement in architectural acoustics and sound system engineering. It captures the complete acoustic signature of a space by recording how sound energy decays over time after an impulsive excitation. From this single measurement, virtually all standardized room acoustic parameters can be derived, including reverberation time (RT60), clarity (C50 and C80), definition (D50), centre time (Ts), and speech transmission index (STI).
An impulse response h(t) describes how a linear time-invariant (LTI) system responds to a Dirac delta function input. In room acoustics, this means exciting a space with an idealized instantaneous burst of energy and recording the result. The captured signal contains the direct sound, early reflections from nearby surfaces, and the late reverberant field that decays exponentially. SonaVyx uses the WebAudio API at 48 kHz with echo cancellation and noise suppression disabled to ensure measurement integrity, and processes the signal through a Rust WASM DSP engine for high-accuracy analysis.
Modern impulse response measurement overwhelmingly favors the logarithmic sine sweep method developed by Angelo Farina. A sweep spanning 20 Hz to 20 kHz is played through a loudspeaker, and the recording is deconvolved with an inverse filter to recover the impulse response. This method offers superior signal-to-noise ratio (SNR) compared to impulsive sources like starter pistols or balloon pops, and it separates harmonic distortion products from the linear impulse response. Maximum Length Sequences (MLS) are an alternative pseudo-random technique with good noise rejection but sensitivity to time variance -- if conditions change during the measurement, artifacts appear.
ISO 3382-1:2009 is the international standard governing the measurement of reverberation and other acoustic parameters in performance spaces. It specifies that the impulse-to-noise ratio (INR) must be at least 35 dB for T20 measurements and 45 dB for T30 measurements to ensure reliable results. The standard also defines the Schroeder backward integration method for computing the energy decay curve (EDC), from which reverberation times are extracted by linear regression. SonaVyx implements these calculations directly in its Rust DSP engine, providing ISO-compliant parameter extraction from measured impulse responses.
The energy decay curve (EDC) is obtained by backward-integrating the squared impulse response (Schroeder method). From the EDC, reverberation time is measured as the time for the curve to decay by 60 dB. In practice, T20 (extrapolated from the -5 to -25 dB range) and T30 (from -5 to -35 dB) are used due to background noise limitations. Clarity C50 and C80 compare early energy (within 50 or 80 ms) to late energy, indicating how clearly speech or music is perceived. Definition D50 is the ratio of early-to-total energy, while centre time Ts is the energy-weighted average arrival time. These parameters together characterize a room for speech intelligibility, musical clarity, and overall acoustic quality.
SonaVyx brings professional-grade impulse response measurement to the browser by combining the WebAudio API for signal generation and capture with a compiled Rust WASM module for DSP processing. The log sine sweep is generated at the specified duration and sample rate, played through the device speakers, and simultaneously captured through the microphone. The Farina deconvolution method then extracts the impulse response with excellent noise immunity. All processing -- including FFT-based deconvolution, Schroeder integration, and parameter computation -- runs locally in the browser with no data sent to external servers. Results can be exported as 24-bit WAV files for further analysis in tools like Room EQ Wizard (REW) or EASERA.
The logarithmic sine sweep (Farina method) is recommended for most situations. It provides the best signal-to-noise ratio and separates harmonic distortion from the linear response. Use a 5-second sweep for typical rooms and 10-15 seconds for very reverberant spaces like churches or concert halls. Balloon pops are convenient for quick measurements but have lower SNR and limited low-frequency energy. MLS is useful in noisy environments but requires time-invariant conditions.
The INR is the ratio between the peak impulse energy and the background noise floor, expressed in decibels. ISO 3382-1 requires at least 35 dB INR for T20 measurements and 45 dB for T30 measurements. A low INR means that the noise floor truncates the decay curve before it reaches the required dynamic range, leading to inaccurate reverberation time estimates. Use longer sweeps and reduce background noise to improve INR.
C50 is the ratio of early energy (first 50 ms) to late energy, expressed in dB. It is primarily used for speech: values above +2 dB indicate good speech clarity. C80 uses an 80 ms boundary and is used for music: values between 0 and +6 dB are considered optimal for concert halls. Higher values mean more direct/early energy relative to reverberation, which improves clarity but may feel acoustically dry if too high.
Built-in laptop microphones can be used for qualitative assessments and learning purposes, but they have significant limitations: non-flat frequency response, limited dynamic range, and directional characteristics that affect the measurement. For professional measurements per ISO 3382-1, an omnidirectional measurement microphone (such as the miniDSP UMIK-1 or Behringer ECM 8000) is strongly recommended.
The energy decay curve is computed using the Schroeder backward integration method: the squared impulse response is integrated from the end of the signal backward to each time point, then normalized and expressed in decibels. This produces a smooth, monotonically decreasing curve from which reverberation parameters (T20, T30, EDT) are extracted by fitting a linear regression to specific segments. The EDC is more reliable than the raw impulse response decay because it averages out fluctuations.