Directional Early-to-Late Energy Ratios to Quantify Clarity: A Case Study in a Large Auditorium

Campos A.; Sakamoto S.; Salvador C.D. · 2021 Immersive and 3D Audio: From Architecture to Automotive, I3DA 2021 (2021) · DOI: 10.1109/i3da48870.2021.9610935

Resumen

Early-to-late energy ratios (ELER) are used to quantity speech intelligibility and music clarity in acoustic spaces from measurements of omnidirectional room impulse responses (RIR). Nowadays, the capture of directional RIRs is possible with spherical microphone arrays and the spherical Fourier transform. These tools are thus motivating the enhancement of omnidirectional metrics and the search for new metrics to quantity directional features of sound. This research explores a directional metric of intelligibility and clarity based on ELERs of directional RIRs. The early-to-late transition times are chosen according to the content: 50 ms for speech and 80 ms for music. The proposed metrics can therefore be interpreted as directional versions of the standard clarity indexes of speech (C50) and music (C80). Directional RIRs were captured at many seats in a large auditorium using a first-order ambisonics microphone. Supporting acoustic simulations of a cuboid room with a second-order ambisonics microphone were also used. Directional ELERs were calculated in the octave bands within the operation range of the microphones. Three directional ELER patterns were identified: an omnidirectional pattern, a dipole pointing forward and backward, and a beam pointing towards the source.