The Impact of Auditory Roughness on Perceived Choral Blend: A Pilot Study
Relevant Previous Studies
Components of Blend
Methodology
Considerations
New Project Methodology
Weston Noble, "Achieving Choral Blend Through Standing Position" --> Acoustical Matching
Vibrato
Vowel
Upper Partial Energy
Spacing
"A primary finding of this investigation is that spread spacing among choristers in this university choral ensemble yields significantly preferred nuances in choral sound for both singers an auditors." Daugherty 2003
"80% of choristers in this investigation, including 100% of female singers experienced moderate to minor body tension and strain of vocal production when in close spacing" Daugherty 2003
Lombard effect: The tendency to increase one's speaking volume when in a loud environment in order to be better heard, Tonkinson (1990)
Self-to-Other Ratio: "the difference in decibels between the sound levels of Self and Other, as experienced by a given singer." Ternström, 2003
Pitch
"Listeners would tolerate a standard deviation in F0 of 14 cents, meaning that two-thirds of the ensemble would be within +/- 14 cents of each other" Ternström 1993b
Sample Recording
Survey Creation
click to edit
Acoustical Matching Process: Singers perform single phrase from "My Country 'Tis of Thee," and are rearranged until it matches conductors preference
Auditors perceive less difference between spacing conditions in comparison to singer assessment, Adams (2019)
Close formations are least preferred by auditors, with less extreme difference between moderate and spread spacing, Daughter (2003), Aspaas et. al (2004)
Blend Definition: "a product of sound where each element becomes unified or homogenized." Knutson (1987)
When singers are placed in a choral environment, there is an instinctual lessening of resonance in singers formant, Goodwin (1980)
Cultural Trends in Aesthetic Values
Western auditors tend to prefer less high frequency energy in choral sound, Ford (2003)
Eastern European choirs can achieve a high quality blend with significant high frequency energy in sound, Quist (2008)
Katherine et. al (2007) found that upper partial energy actually increased for opera singers participating in an opera chorus
Quantifying Blend
Common use of multiple auditors rating on a 7-point Likert scale
Many studies (Quist, Daugherty, Aspaas, etc.) only consider the composite sound of the choir when analyzing spectrographically or addressing auditors
Application of Bel Canto training techniques improves blend of choir while establishing a better balance of chiaroscuro, Fagnan (2005)
Trevor (1977) found that judges rated higher blend as differences in vibrato rate and extent decreased
"it was the alteration of vibrato extent, not the elimination of vibrato, that led to an apparent increased perception of blend" Basinger, 2006
"Successful vocal pedagogues throughout time have been concerned with the quality and healthy production of tone. The fact that vibrato presence and rate has changed over time is an indication that the occurrence of more than one degree of vibrato variation is within the realm of healthy production." Quist (2008)
"Singers with greater vibrato extent tend to blend less-effectively and the more years of lessons singers have, the greater their vibrato extent tends to be" Osinski, (2020)
Does the acoustic match between personnel contribute to blend? Yes: the more similar singers' vibrato extents are, the better the perceived blend, Osinski (2020)
Giardiniere (1991) --> one of the first written records studying acoustical matching. Suggested that vibrato and singers' formant resonance are two primary variables
"With only two voices singing together, similar strength in singers' formant may have been overruled by the unique structure of each vocalist's singers' formant in making his/her voice distinguishable" Osinski 2020 quoting Sundberg 1972
Distribution of voice parts is less influential than spacing of choral members Daugherty (1999), Lambson (1961)
Sample Collection
Goodwin (1980): "The results of this study suggest that choral singers may adjust their overall intensity not only to affect the perceived loudness of their tones, but also to create other acoustical changes in their vocal sounds, which the singers perceive as helpful for achieving vocal blend"
"...vibrato allows singers to vary the size of consonant intervals... Sundberg suggests that vibrato serves not as a deterrent, but as an aid to good intonation" Galante (2011)
"Ensemble blend is enhanced as all choristers learn to produce a similar, unified sound" Fagnan, (2005), in relation specifically to vowels
Vowels inhabit different balances of natural brightness and darkness. Fagnan, (2005)
"It is not yet clear to what degree singers' formant prominence is a liability in choral blend" Osinski (2020)
Many studies use LTAS of choral repertoire passages as basis of study - we need to simplify this to prove basic concepts
Many samples feature singers in homophony, meaning they are mixing the presence of different pitches while trying to spectrographically analyze
New Study: Collect data for both individual and composite sounds
New Study: Mix and match singers who are producing simple and consistent sounds, and examine vocal qualities from there
New Study: Assess singers who are producing the same pitch for a small length of time to determine perception of blend
In previous studies, auditors were not only limited to those with vocal training (positive and negatives here)
Think through most effective way to pair permutations of samples
Subjective quantification in questioning auditors
Objective quantification of blend using software?
What types of statistical analysis will show the most important relationships?
Control Parameters: Vowel [a] [i] [u], Pitch (D4/D5), Vibrato rate (Straight-tone, mediated, soloistic)
Variable: Upper partial energy --> Prompted through use of different dynamics
3 subjects to provide variation and statistically valid measures of comparison
Survey Creation
Normalizing volume of samples to retain discrepancies in registration/presence of upper harmonics while eliminating uneven presence in general volume
Use of 7-point Likert Scale
4 Questions per sample (written in methodology)
Samples of around 6 seconds
Composite pairs always have same vowel, same pitch, and same vibrato rate
Further Analysis
Survey results processed in RStudio to create charts identifying samples with highest rates of blend
TBD: Analyzing sample sets specifically in regards to vibrato, given previous research in the field
Identifying trends in blend perception: Hypothesis is that samples with discrepancy in use of upper partial energy will be rated with least blend
Survey given to auditors with vocal training, controlling as possible for headsets used to minimize discrepancies in perceptual experience
click to edit