Acoustic Analysis of the Interaction of Choral Arrangements, Musical Selection, and Microphone Location

Acoustic differences were evaluated among three choral arrangements and two choral textures recorded at three microphone locations. A choir was recorded when singing two musical selections of different choral texture, one homophonic and one polyphonic. Both musical selections were sung in three choral arrangements: block sectional, sectional-in-columns, and mixed. Microphones were placed at the level of the choristers, the conductor, and the audience. The recordings at each location were analyzed using long-term average spectrum (LTAS). The LTAS from the mixed arrangement exhibited more signal amplitude than the other arrangements in the range of 1000-3500Hz. When considering the musical selections, the chorus produced more signal amplitude in the region of 1800-2200Hz for the homophonic selection. In addition, the LTAS produced by the choir for the homophonic selection varied across the microphone locations. As for the microphone location, the LTAS of the signal detected directly in front of the chorus had a greater slope than the other two locations. Thus, the acoustic signal near the choristers differed from the signals near the conductor and in the audience. Conductors may be using acoustic information from the region of the second and third formants when they decide how to arrange a choir for a particular musical selection.     

Synthesizing choir singing

Analysis by synthesis is a method that has been successfully applied in many areas of scientific research. In speech research, it has proven to be an excellent tool for identifying perceptually relevant acoustical properties of sounds. This paper reports on some first attempts at synthesizing choir singing, the aim being to elucidate the importance of factors such as the frequency scatter in the fundamental and the formants. The presentation relies heavily on sound examples.     

Hearing myself with others: Sound levels in choral performance measured with separation of one's own voice from the rest of the choir

The choir singer has two acoustic signals to attend to: the sound of his or her own voice (feedback), and the sound of the rest of the choir (reference). The balance in loudness between feedback and reference is governed mainly by the room acoustics. Although earlier experiments have shown that singers have a fairly large tolerance for imbalance, with references ranging from −23 to +5 dB, experience suggests that, when singers are given control over this parameter, their preferences are much narrower. A quantification of the optimum balance would be useful in the design of concert stages and rehearsal halls. A method is described for measuring the feedback and reference levels as experienced by singers under live performance conditions. Recordings were made using binaural microphones worn by choir singer subjects. With the given combination of choir and room, it was possible to achieve adequate separation of the feedback and reference signals with simple signal processing. The feedback-to-reference ratio averaged over the 12 singers was found to be +3.9 dB, with extremes of +1.5 and +7.3 dB.     

Perceptual evaluations of voice scatter in unison choir sounds

The preferences of experienced listeners for pitch and formant frequency dispersion in unison choir sounds were explored using synthesized stimuli. Two types of dispersion were investigated: (a) pitch scatter, which arises when voices in an ensemble exhibit small differences in mean fundamental frequency, and (b) spectral smear, defined as such dispersion of formants 3 to 5 as arises from differences in vocal tract length. Each stimulus represented a choir section of five bass, tenor, alto, or soprano voices, producing the vowel [u], [a], or [w]. Subjects chose one dispersion level out of six available, selecting the “maximum tolerable” in a first run and the “preferred” in a second run. The listeners were very different in their tolerance for dispersion. Typical scatter choices were 14 cent standard deviation for “tolerable” and 0 or 5 cent for “preferred.” The smear choices were less consistent; the standard deviations were 12 and 7%, respectively. In all modes of assessment, the largest dispersion was chosen for the vowel [u] on a bass tone. There was a vowel effect on the smear choices. The effects of voice category were not significant.