• Corpus ID: 61598998

Room acoustics and speech perception Prepared for Seminars in Hearing

@inproceedings{Boothroyd2002RoomAA,
  title={Room acoustics and speech perception Prepared for Seminars in Hearing},
  author={Arthur Boothroyd},
  year={2002}
}

• A. Boothroyd
• Published 2002
• Physics

The acoustic speech signal received by a listener is a function of the source, distance, early reverberation, late reverberation, and noise. Specifically, it depends on the Speech Audibility Index, which is defined, here, as the proportion of the combined direct speech and early reverberation (also known as early reflections) whose level is above that of the combined noise and late reverberation. Speech Audibility Index rises from 0 to 100% as the effective signal-to-noise ratio rises from -15 to +15 dB. Both reverberation and ambient noise need to be low in order to maintain Speech Audibility Index at an optimal level. Speech Audibility Index can be used to predict various measures of speech perception, but the results are highly dependent on the complexity of the language and the characteristics of the listener. Conditions that are tolerable for normally hearing adults in casual conversation can be difficult for adults and children in learning situations, and intolerable for persons with deficits of hearing, language, attention or processing. Sound-field amplification can improve Speech Audibility Index for all listeners in a noisy room. It offers less benefit when the primary problem is reverberation and, if improperly installed, can make the reverberation problem worse. There is no good substitute for reverberation control. Audiologists have an important contribution to make in the identification and resolution of continuing inadequacies of classroom acoustics. Introduction Room acoustics have a major effect on the transmission of speech sounds from talker to listener. Four principal factors are involved: distance, early reverberation, late reverberation, and noise. The present paper outlines the effects of these factors on the reception and perception of speech. The initial speech signal Before examining what happens to speech in a room, it is important to define the original acoustic signal. i) Long-term average level For present purposes, I will consider the original acoustic signal to be that measured at 1 foot from the lips. At this distance, the long-term speech level of a typical talker, averaged over 10 or 20 seconds, is around 70 dB SPL. It is important to remember, however, that this value is summed across frequency and averaged over time. ii) Long-term average spectrum The heavy line in Figure 1 shows the long-term level of a 12 second speech sample measured in 1/3-octave bands. This is the Long-TermAverage Speech Spectrum or LTASS. The level is highest in the lowfrequency bands, and falls at the rate of around 6 dB per octave at frequencies above 500 Hz (see, also, Cox and Moore, 1988; Boothroyd, Erickson, and Medwetsky, 1994). It is a characteristic of the acoustic Room Acoustics. Boothroyd, 2002. Page 3 of 18 Frequency in Hz 125 250 50