Longer Reverberation Research Articles

An experiment on the installation of small calming spaces in childcare rooms - Frequency of use with internal sound absorption

In this study, a field experiment was carried out in a kindergarten classroom for 5-year-olds where a small spaces for calming down were installed. The small space has a box shape for one person with a width, depth, and height of 42, 57, and 74 cm, and two units were installed with different inside sound-absorbing conditions. The noise attenuation due to the small space was about 6 to 7 dB above 500 Hz bands with additional internal sound absorption. The experiment was performed for two months in a room with short and long reverberation, respectively, during which the use of the space was automatically observed using a mat sensor and the room sound was recorded at the same time. As a result, the frequency and duration of use of the spaces increased in the room with longer reverberation. Also, the duration of use of spaces with internal sound absorption tended to be longer.

Early reflection detection using autocorrelation to improve robustness of speaker verification in reverberant conditions

A speech signal captured by a distant microphone is generally smeared by reverberation, which severely degrades automatic speaker recognition performance. To improve system performance, an effective and robust method is proposed to extract features for speech processing. In this paper, a room impulse response is presumed to comprise of three parts: a direct-path response, early reflections and late reverberations. Since late reverberations are known to be a major cause of system performance degradation, this paper focuses on dealing with the effect of early reflection because the early reflections and their properties play a necessary role within the acoustics of an enclosure. The proposed method first estimates the early reflection using autocorrelation function from the presentation of speech signals in the first stage, the estimates are combined with an anechoic signal for use into training the system in the second stage. The employed method looks to be promising, achieving a substantial improvement in system performance relating to reduced equal error rate and detection trade-off, especially at longer reverberation time.

Amplified music and variable acoustics—Shorter reverberation times at low frequencies

A survey from 2010 among professional musicians and sound engineers revealed that a long reverberation time at low frequencies in halls during concerts of reinforced music such as pop and rock, is a common cause for an unacceptable sounding event. Lower frequency sounds are, within the genre of popular music, rhythmically very active and loud, and a long reverberation leads to a situation where the various notes and sounds including vocals cannot be clearly distinguished. This reverberant bass sound rumble often partially masks even the direct higher pitched sounds. In an article from 2011, it was hypothesized that mid- and high-frequency sound is seldom a reason for lack of clarity and definition due to a 6 times higher absorption by audience compared to low frequencies, and a higher directivity of speakers at these frequencies. A survey from December 2016 among 25 professional musicians and sound engineers confirms that a longer reverberation at higher frequencies in the empty hall can advantageously be projected and that the 125 Hz octave band is probably the single most important band to control. Details from this survey and results regarding the author’s most recent developments in the field of variable and mobile absorption in favor of both amplified and classical musical genres are presented.

Subjective Evaluation of Acoustical Environments for Solo Performance

A subjective experiment was conducted involving assessment of acoustical environments for solo performance by two different instruments, the cello and the xylophone. The sound fields were produced by digital convolution of binaural impulse responses with recorded “dry” music. A group of professional listeners and a group representing the general public were asked to evaluate simulated music via headphones. A forced-choice comparison was used. Participants picked their preferred sound field for each of 66 pairs of test signals. Multiple regression analyses yielded 3 orthogonal measures. They were early level (G80), reverberation time (RT) or early decay time (EDT), and treble ratio of early level (T(G80)) according to their respective significance of correlation with preference score. The minimum acceptable overall level (G) was 5.6 dB and the preferred reverberation time (RT) was 1.4 sec. Both groups of listeners preferred a 0.19-sec longer reverberation time for the cello music than for the xylophone music.

Speech Recognition in a Noisy and Reverberant Environment with and without Earmuffs

The effect of earmuffs on speech recognition in conditions comparable to an actual environment was studied with the use of sentences, words and nonwords. The speech level was adjusted to 60 and 85 dBA. White noise was used at signal-to-noise ratios (S/N) of 0, +5, and +10 dB. The reverberation times were 2.1 and 1.6 s. In quiet conditions at normal speech level (60 dBA), the perception was better without earmuffs than with them. At a high noise level, equal or better scores were achieved with ear protectors. At high speech and noise levels (85 dBA, S/N 0), the benefit of earmuffs was clearly demonstrated. Also with longer reverberation, the use of ear protectors increased the perception scores, especially at high speech and noise levels.

Monaural and Binaural Free-Field Speech Reception under Noise and Reverberation

Speech reception was studied in a single sound-treated room at reverberation times T=0.3 and T=0.6 sec, to compare binaural and monaural reception in the presence of an impulsive noise (16 imp/sec) and a quasisteady noise (a babble of eight voices). The speech and noise sources were fixed at locations, relative to the listener, to give the largest advantage for binaural reception. Binaural gain is taken as the difference in speechto-noise ratios for 60% correct speech reception comparing binaural and monaural results. At T=0.6 sec, binaural gain was 3 dB for both types of noise. At T=0.3 sec the gain was 5 dB for the babble and 4 dB for the impulsive noise. The longer reverberation caused a reduction in reception in both binaural and monaural conditions. These effects were larger for the impulsive noise and for low S/N ratios. The largest reverberation effect reduced reception from 77% to 53% words correct. [Supported by Gallaudet College.]