Abstract
A single session of vocalizing through a straw (straw phonation) has been associated with consistent or increased acoustic output and chorister-reported improved choral sound and vocal efficiency. Some choristers, however, report discomfort from the intense pressure of vocalizing through a stirring straw (2.5–3.0 mm opening). It is unclear whether participants acclimate to this pressure over time. Two mixed and two treble choirs sang a unison melody, participated in a straw protocol, and sang the melody again. Choristers used the protocol during warmups in four subsequent rehearsals and repeated the same pretest-posttest protocols on the final day. Survey results indicated (a) at least 89.7% of participants self-reported improved choral sound and more efficient/comfortable individual voicing after the protocols on both data collection days, and (b) participants reported greater impact on their warmup on the final day. Analyses revealed (a) choirs sang with 1.35 to 3.37 dB SPL greater mean spectral energy after the protocols on the first day and 1.63 to 2.65 dB SPL on the final day, and (b) one choir evidenced less change on the final day, one evidenced more change, and two evidenced no statistical difference. These results may represent a modest acclimation effect for choristers using straw phonation.
Semi-occluded vocal tract exercises (SOVTEs) have become the centerpiece of many warmup and treatment protocols for professional singing and speaking coaches as well as voice and speech therapists. These techniques include many common voicing exercises: lip or tongue trills, raspberries, voiced fricatives (e.g. v, z), nasal consonants (e.g. m, n, ng), and vocalizing through tubes of various sizes (i.e. straw phonation) in air or suspended in water. Researchers have found that SOVTEs create an impedance in the vocal tract that can reduce the breath pressure needed to initiate and sustain voicing (Titze, 2006) and lower vocal fold collision forces during vibration (Titze, 2004) while increasing the acoustic energy of each collision in subsequent voicing (Dargin & Searl, 2015; Titze, 2004). One set of researchers (Laukkanen et al., 1996) also found that participants experienced a reduction in neck muscle engagement while producing spoken vowel sounds after engaging in an SOVTE. Laukkanen et al. (1998) also found diminished glottal resistance due to increased airflow after exercising with a SOVTE. Speech and voice therapists have used these exercises in specific voice therapy methods (Bassiouny, 1998; Laukkanen, 1992; Simberg & Laine, 2007; Stemple et al., 1994). Researchers have argued that the available evidence indicates that such protocols lead to improved singer or speaker “vocal economy” (Guzman et al., 2013; Titze & Laukkanen, 2007), allowing similar or increased vocal output with decreased effort and fatigue.
Until somewhat recently, research on SOVTEs has been limited to studies with individual singers or speakers. In a 2009 survey, Chorus America (2009) estimated that roughly 43 million people in the United States were participating in a chorus of some kind. Although many of these singers may seek private voice instruction, the highest rate of participation in this study was in volunteer or community choruses. It seems likely that many of these amateur singers experience the majority of their singing instruction in a choral context.
Similar to voice therapists and individual vocal pedagogues, choral teacher-conductors, and music teacher educators have long advocated the use of various SOVTEs in some of the most utilized texts in choral music teacher education (Brinson & Demorest, 2014; Ehmann & Haasemann, 1981; Nesheim & Noble, 1995; Phillips, 2016). In their book, Vocal Technique: A Guide for Conductors, Teachers, and Singers, Davids and LaTour (2012) specifically recommended straw phonation for entraining efficient vocal fold closure. Dr. Ingo Titze, a voice scientist, and Dr. Alan Henderson, the president of the National Association of Teachers of Singing (Titze & Henderson, 2015) also advocated straw phonation at an interest session of the American Choral Directors Association (ACDA) National Conference as a way to stretch and unpress the vocal folds during warmups. Such claims align with experimental results in extant research (Titze et al., 2002).
Although choral teacher-conductors and individual voice instructors have both recommended SOVTEs, a body of research has suggested that solo and choral singing are two related but distinct endeavors. For example, choristers stand in a group formation, which affects their ability to hear their own voice in relation to others (“self-to-other ratio,” Ternström, 2003). When they are unable to sing with enough “self” feedback, they may make unconscious changes to their own voicing. This phenomenon, referred to as the “Lombard effect” (Tonkinson, 1994), may lead choral singers to sing with “greater tension” (p. 25) or to “push or force their voices to enhance feedback” (p. 24).
In addition, soloists frequently seek to produce a “resonance” or “ring” that projects the voice over an orchestra. This “ring” corresponds to a boost in the 2 to 4 kHz region of the acoustic spectrum and is frequently referred to as the “singer’s formant” cluster (Sundberg, 1987). This acoustic region also corresponds with the area in which the human ear is most sensitive (Fletcher & Munson, 1933). Rossing et al. (1986) found that male singers utilized less acoustic energy in the “singer’s formant” region when asked to sing in “choral mode.” In addition, listeners have tended to prefer the sound of a chorus that sang with less energy in this region, either because of singer conscious choices (Ford, 2003) or when the spectral energy in this region was dampened concomitantly with inter-chorister spacing (e.g. Daugherty et al., 2013). Therefore, solo vocal techniques may not directly translate to the choral rehearsal.
A growing line of inquiry involves straw phonation with choirs of varied composition (Manternach & Clark, 2018; Manternach et al., 2017; Manternach & Daugherty, 2019; Manternach et al., 2019). For each investigation, the choristers participated in a 4- to 5-minute straw phonation protocol found on a YouTube video created by Titze, professor emeritus at the University of Iowa and executive director of the National Center for Voice and Speech (NCVS). In each investigation, a choir performed a piece, took part in the researcher-led straw phonation protocol with a small stirring straw (2.5–3.0 mm opening), and then sang the piece again.
Results in these studies have seemed to align with extant solo singer literature. An auditioned collegiate SATB (soprano, alto, tenor, bass) choir in one investigation (Manternach & Daugherty, 2019) sang with no significant acoustic change, but choristers (N = 48) reported a preference for the group sound (78.3%) and that they sang more efficiently/comfortably (73.9%) after the straw protocols. Another less experienced high school/collegiate SATB choir (N = 15 choristers) experienced small, significant boosts in the overall sound pressure level (the psychoacoustical correlate of volume) on two motets after the protocols (Manternach et al., 2017). A community barbershop chorus sang with 1 dB SPL (decibel sound pressure level) boost after the straw protocols (Manternach & Clark, 2018), and choristers reported improved group sound (85.7%) and individual voicing efficiency (62.5%) after the protocols. One final investigation of two matched SSAA choirs included a control group that sang the protocol using [a:] (“ah”) rather than through a straw (Manternach et al., 2019). The straw group experienced a boost of 1.35 dB SPL compared to 0.40 dB SPL for the control group. The difference between these groups is perhaps meaningful, as Howard and Angus (2006) have posited a “general rule of thumb” that a change of around 1 dB SPL would likely constitute a “just noticeable difference” (p. 98) when listening to complex sound. Results of this series of investigations seem to support the use of straw phonation as a way to increase or maintain chorister vocal output with perceptions of easier vocal production. Such results may imply an increase of vocal efficiency.
Summary and purpose statement
All of the ensembles in these studies have seemed to enjoy benefits from straw phonation after their typical warmups (Manternach & Clark, 2018; Manternach et al., 2017; Manternach & Daugherty, 2019), after a portion of rehearsal (Manternach & Clark, 2018; Manternach et al., 2017), or after performing the piece two times to ensure memorization (Manternach et al., 2019). None of the ensembles has utilized straw phonation as a stand-alone vocal warmup (i.e. before any other group vocalizing). These investigations have also involved pretest-posttest designs in which the choir had little or no experience using group straw phonation. Anecdotally, some choristers initially reported difficulty voicing through a small stirring straw. Nix and Simpson (2008) have also noted that the intense pressure caused by the narrow opening of the small stirring straw might initially be difficult for singers to negotiate. It is possible, however, that further exposure to the protocols could help acclimate participants to the intense pharyngeal pressure associated with voicing through a stirring straw. Such acclimation through an intentional training period may lead to different results.
Therefore, the purpose of this investigation was to measure the effect of a straw phonation protocol as a stand-alone vocal warmup on acoustic and perceptual changes of choral sound prior to and after a four-rehearsal acclimation period. The following questions guided the investigation.
Will there be acoustic differences in the overall spectrum (0–10 kHz) and the area in which the human ear is most sensitive (2–4 kHz), as measured by long-term average spectra (LTAS), of four choirs after taking part in a straw phonation protocol?
Will singer participants perceive differences in the choral sound, their individual vocal production, or their individual warmup level after taking part in the protocol?
Will acoustic or perceptual changes evoked by the straw phonation protocol differ between the first day of data collection and following a four-rehearsal acclimation period?
Method
Participants
Participants (N = 257) for this study were members of four choral ensembles at two medium-sized universities—one midwestern (n = 2 choirs) and one southeastern (n = 2 choirs). Each university had one treble choir (Soprano, Soprano, Alto, Alto voicing—i.e. SSAA) and one mixed choir (Soprano, Alto, Tenor Bass voicing—i.e. SATB). The treble choirs had 50 (midwestern) and 100 (southeastern) choristers, and the mixed choirs had 49 (midwestern) and 58 (southeastern) choristers.
Of these 257 choristers, 203 (79.0%) completed the singer survey on both days of data collection, which allowed for comparison of perceptions prior to and after the acclimation period. On the survey, these participants self-reported zero to 26 years of choral singing experience (M = 10.1, SD = 3.9) and indicated whether they typically sang the scored notes for soprano (n = 87, 42.9%), alto (n = 77, 37.9%), tenor (n = 12, 5.9%), or bass (n = 20, 9.9%). Seven participants (3.4%) did not report a voice part.
Procedures and equipment
On the respective first days of data collection, each choir convened in their normal rehearsal room and assembled in their choral formation. We then used a spacer to ensure consistent distances between their shoulders during the procedures. Choristers at the midwestern university used their binders to ensure an 8.5-inch inter-singer spacing. Choristers at the southeastern university used pre-cut, 6-inch, wooden shims to measure their inter-singer spacing. The researchers then marked the chorister locations on the floor with tape and used the same markings on both the first and final day of data collection to ensure consistent choral placement and formation. The spacing between choristers was not consistent between the two universities due to idiosyncrasies in the rooms’ configuration and dimensions. The spacing was, however, consistent with the approximate inter-chorister spacing for typical rehearsals in each space.
Protocols
After completing the spacing procedures, choristers watched a pre-recorded conductor on a projection screen who conducted a slightly abbreviated version of the “Star Spangled Banner” (omitting the second repetition of the initial melody, “Whose broad stripes. . .”). This initial viewing served to familiarize them with the tempo and the video recording while they audiated (i.e. thought the melody) the version of the melody we chose for the investigation. We then instructed the choristers to follow the pre-recorded conductor while they sang a unison rendition of the abbreviated “Star Spangled Banner” in the key of A Major.
To create the stimulus video, the director of the southeastern university choirs created a video in which she conducted the piece with a metronome set at MM = 84 beats/minute. The director of the midwestern university choirs then viewed this video as she created a recording of herself conducting the piece. Both directors wore black formal attire, which was their normal dress for conducting concerts at their institutions. The videos resulted in an approximate singing time of 53 seconds.
Straw protocol
After the first repetition of the “Star Spangled Banner,” we distributed a small stirring straw (3.0 mm internal diameter × 13.5 cm length) to each chorister. We then led the singers in a straw phonation protocol based on the same YouTube video (NCVS456, 2010) used in multiple studies in this line of inquiry, as it (a) was created by researchers who have investigated SOVTEs and (b) is widely accessible to researchers and practitioners for replication or adoption. This protocol included a brief tutorial on straw phonation basics. First, we instructed choristers to maintain a seal around the straw with no air escaping though the sides of the mouth or the nose. Second, we asked them to minimize neck muscle activity and engage “belly” energy (NCVS456, 2010) during the procedures.
After this tutorial, we led the chorus in a roughly 4-minute protocol of voicing through the straw. First, choristers performed glides from the lowest note to the highest note and back down again for 90 seconds. After the glides, we instructed the choristers to perform 90 seconds of “accents.” These exercises included a series of vocal pulses that varied in pitch in volume as they ascended through the range before then gliding downward toward the lowest note again. Finally, participants sang the same unison rendition of the “Star Spangled Banner” through the straw. Following these protocols, participants checked their standing positions, directed their attention to the projection screen, and followed the conductor in a second unison rendition of the “Star Spangled Banner.”
Acclimation period
Following the initial data collection, we instructed the participants to keep their straws. Over the subsequent four rehearsals, which varied in timing depending on the schedule of each ensemble, the choir assembled as they normally would. Prior to any other vocalizing on each day, we led them in the same straw phonation protocol. Each rehearsal then continued normally.
Posttest
After the four-rehearsal acclimation period, we replicated our procedures from the first day of data collection. Choristers arrived for rehearsal, took their marked place in the choral formation, viewed the pre-recorded conductor while audiating the melody, then sang the abbreviated “Star Spangled Banner” while following the pre-recorded conductor. They then completed the straw phonation protocols and sang the piece again while following the pre-recorded conductor.
Acoustic analyses
In each rehearsal space, a Zoom H6 (Tokyo, Japan) device with MSH-6 MS microphone capsules stood approximately where the conductor would stand during rehearsals in the room. The resulting distance was 7.5 feet (midwestern) and 6 feet (southeastern) from the first row of choristers. As sound pressure level (i.e. dB SPL) is highly dependent on distance (Zahorik & Kelly, 2007), we maintained these placements during both data collection sessions. Although we did not calibrate the instruments for exact SPL measurements, the consistent distance from choir to recording device allowed us to accurately collect each choir’s change in SPL (i.e. Δ dB SPL) on pre- and posttest recordings between the first and final days. The Zoom devices recorded at a 96 kHz sampling rate (24 bits) in a .wav format. We used KayPENTAX Multi-Speech 3700 software (version 3.4.1, Montvale, New Jersey, USA) to analyze the long-term average spectrum (LTAS, no preemphasis or smoothing, window size of 512 points, bandwidth of 187.5 Hz, and a Blackman window), which is commonly used in acoustics to assess choral timbre (e.g. Daugherty et al., 2013; Manternach & Clark, 2018; Manternach et al., 2017; Manternach & Daugherty, 2019; Manternach et al., 2019; Morris et al., 2010). The bandwidth created 54 data points in the entire spectrum and 11 data points in the 2 to 4 kHz region, the two areas on which we focused our acoustic analyses.
Singer survey
After completing the sung tasks, we asked willing participants to complete a follow-up survey. The IRB-approved survey collected demographic data (e.g. years of choral singing experience, voice part) and asked for the singers’ perceptions regarding the protocols. First, they noted whether the “choir as a whole” sounded best during the first (pretest) or second (posttest) excerpt. Next, they indicated which excerpt corresponded with their “most efficient/comfortable” vocal production.
Participants then indicated whether they believed the warmup protocols affected the sound of the choir. If they circled “yes,” they could place a vertical mark on a 10-cm horizontal visual analog scale (VAS) indicating the extent of the effect. They responded similarly to inquiries of the effect the warmup protocols had on their own vocal sound and how much the protocols helped them to be “more warmed up” for the second excerpt. We measured the vertical tick marks for each affirmative response, which resulted in a measurement of 0.0 (very little) to 10.0 (very much) cm.
Results
Acoustic differences
Overall results
For each recording, we calculated the LTAS across the entire spectrum (54 data points, 0–10 kHz). We then compared the difference at each data point on the first day to the difference on the final day to determine whether the acclimation period caused the ensemble to react differently to the straw protocols on the 2 days. Graphic representations of the overall results appear in Figures 1 and 2 and comparisons appear in Table 1. We ran six paired t-test comparisons for each choir—pretest versus posttest on the first and final days across the entire spectrum, pretest versus posttest on the first and final days in the 2 to 4 kHz region, and the change of the entire spectrum and 2 to 4 kHz region (i.e. Δ dB SPL) between days. In order to maintain a familywise error rate of 0.05 within each choir, we used a Bonferroni correction to determine statistical significance (.05/6 = .0083).
LTAS results on the first and final days for the mixed (SATB) choirs.
LTAS results on the first and final days for the two treble (SSAA) choirs.
Mean dB SPL boost across the spectrum and in the 2 to 4 kHz region for all four choirs on the first and final day of data collection.
Note. SMC = southwestern mixed choir; MMC = midwestern mixed choir; STC = southwestern treble choir; MTC = midwestern treble choir.
Indicates a statistically significant difference, p < .0083.
All the choirs had robust, statistically significant differences between the pretest and posttest on both the first and final days of data collection in the entire spectrum and in the 2 to 4 kHz region. In every case, the posttest had higher mean sound pressure level than the pretest. Below, we report these differences and compare the magnitude of the differences on the first and final days.
Southeastern mixed choir (SMC)
On the first day of data collection, the SMC mean sound pressure level was 1.79 dB SPL higher across the entire spectrum, t(53) = 13.661, p < .001, d = 1.86, and 2.06 dB SPL higher in the 2 to 4 kHz region, t(10) = 24.687, p < .001, d = 7.44, during the posttest. On the final day, the mean sound pressure level was 1.63 dB SPL higher across the entire spectrum, t(53) = 22.956, p < .001, d = 3.12, and 1.85 dB SPL higher in the 2 to 4 kHz region, t(10) = 23.829, p < .001, d = 7.18, during the posttest. The lower magnitude of sound pressure level boosts on the final day (0.16 and 0.21 dB SPL) were not statistically significant differences between the 2 days in the entire spectrum, t(53) = 2.089, p = .042, d = 0.28, or the 2 to 4 kHz region, t(10) = 2.432, p = .035, d = 0.73.
Midwestern mixed choir (MMC)
On the first day of data collection, the MMC mean sound pressure level was 1.35 dB SPL higher across the entire spectrum, t(53) = 27.879, p < .001, d = 3.79, and a mean of 1.50 dB SPL higher in the 2 to 4 kHz region, t(10) = 19.752, p < .001, d = 5.96, during the posttest. On the final day, the mean sound pressure level was 1.66 dB SPL higher across the entire spectrum, t(53) = 22.843, p < .001, d = 3.11, and a mean of 1.91 dB SPL higher in the 2 to 4 kHz region, t(10) = 12.944, p < .001, d = 3.90, during the posttest. The increased sound pressure level boosts on the final day (0.31 and 0.41vdB SPL) represented statistically significant differences between the 2 days in the entire spectrum, t(53) = 5.636, p < .001, d = 0.77, and in the 2 to 4 kHz region, t(10) = 3.387, p = .007, d = 1.02.
Southeastern treble choir (STC)
On the first day of data collection, the STC mean sound pressure level was 2.06 dB SPL higher across the entire spectrum, t(53) = 48.826, p < .001, d = 6.64, and a mean of 2.08 dB SPL higher in the 2 to 4 kHz region, t(10) = 34.771, p < .001, d = 10.48, during the posttest. On the final day, the mean sound pressure level was 2.14 dB SPL higher across the entire spectrum, t(53) = 34.131, p < .001, d = 4.64 and a mean of 2.68 dB SPL higher in the 2 to 4 kHz region, t(10) = 43.059, p < .001, d = 12.98, during the posttest. The increased boost on the final day across the entire spectrum (0.08 dB SPL) did not represent a statistically significant difference between the 2 days, t(53) = 1.6479, p = .105, d = 0.22. The larger boost in the 2 to 4 kHz region (0.60 dB SPL) however, was a statistically significant difference compared to the first day, t(10) = 9.418, p < .001, d = 2.84.
Midwestern treble choir (MTC)
On the first day of data collection, the MTC mean sound pressure level was 3.37 dB SPL higher across the 0 to 10 kHz spectrum, t(53) = 46.171, p < .001, d = 6.28, and a mean of 3.92 dB SPL higher in the 2 to 4 kHz region, t(10) = 47.356, p < .001, d = 14.28, during the posttest. These differences represent the most robust changes in the investigation. On the final day, the mean sound pressure level was 2.65 dB SPL higher across the entire spectrum, t(53) = 37.876, p < .001, d = 5.15, and 2.99 dB SPL higher in the 2 to 4 kHz region, t(10) = 34.390, p < .001, d = 10.37, during the posttest. The decreased sound pressure level boosts on the final day (0.72 and 0.93 dB SPL) represented statistically significant differences between the two days in the entire spectrum, t(53) = 15.365, p < .001, d = 2.09, and in the 2 to 4 kHz region t(10) = 12.472, p < .007, d = 3.76.
Survey results
Choir’s sound
Two-hundred three participants completed the survey on both days. Among that group, 200 (98.5%) thought the “choir as a whole” sounded better after the straw protocols on the first day of data collection. Nearly the same number (n = 199, 98.0%) believed the choir sounded better after the protocols on the final day of data collection. Three participants (1.5%) believed the choir sounded better before the protocols on both the first and final days. One participant (0.5%) did not respond to the question on the final day.
Individual vocal production
One-hundred eighty-two (89.7%) respondents reported more “efficient/comfortable” vocal production after the straw protocols on the first day. Twenty (9.9%) believed their vocal production was better before the protocols and one participant (0.5%) did not respond to the question. On the final day, 189 (93.1%) reported more “efficient/comfortable” vocal production after the protocols, an increase of seven responses from the first day. Fourteen participants (6.9%) believed their vocal production was better before the protocols, a decrease of six from the first day.
Effects of the protocol
The first two questions in the survey related to choral sound and vocal efficiency; however, they did not specifically refer to the straw protocols as the cause of these changes. The next three questions related specifically to the effect of the straw protocol on (a) the choir’s sound, (b) each chorister’s individual vocal sound, and (c) whether the protocols helped them to be more warmed up. At least 93.6% of the participants thought the protocols affected each parameter on the first and final days (see Table 2). These percentages were largely consistent between the first and final days of data collection.
Perceived effect of the protocol on the choir’s sound, individual vocal sound, and level of warmup.
Indicates a statistically significant difference between days.
Participants who selected “yes” in the previous questions could indicate on the visual analog scale how much of an effect the straw protocols had on each parameter. Mean responses ranged from 6.44 to 7.20 cm. Although means for all parameters were slightly higher on the final day, these increases were not statistically significant for differences in the choir’s sound and individuals’ vocal sound between the first and final days. However, choristers reported small, statistically significant increases in how much the straw protocol helped them to be warmed up on the final day (M = 7.20, SD = 1.99) compared with the first day (M = 6.62, SD = 6.62), t(186) = 3.877, p < .001, d = 0.28.
Summary
These four choirs sang with robust mean differences after the straw protocols compared to before them on both the first (r = 1.35–3.37 dB SPL) and the final days (r = 1.66–2.99 dB SPL). The changes across the full spectrum on the final day were statistically greater for one choir (MMC), lower for one choir (MTC), and were not statistically significant for two choirs (SMC and STC) compared to the first day. The choirs tended to have slightly more boost in the 2 to 4 kHz region than across the entire spectrum (0.02–0.55 dB SPL), though those small differences may not have been noticeable (Howard & Angus, 2006). The changes across this region on the final day were statistically greater for two choirs (MMC and STC), lower for one choir (MTC), and were not statistically different for one choir (SMC) compared to the first day. Neither region displayed a clear consistent trend after the acclimation period.
Discussion
In previous investigations, choirs utilizing straw phonation have done so in a one-day pretest-posttest study design and after some group singing on the day of data collection. These groups have experienced boosts of up to 1.35 dB SPL. All the groups in this investigation, who participated in the protocols prior to any other voicing in their rehearsal, displayed boosts of at least 1.35 dB SPL; differences were as great as 3.37 dB SPL across the spectrum. Choristers also reported at the highest rates to date that the choir sounded better (98.0%–98.5%) and that they sang more efficiently (89.7%–93.1%) after the straw protocols. One data point seemed to evidence a clear trend after the acclimation period—choristers reported feeling slightly more warmed up by the straw protocols on the final day (7.20) than on the first day (6.62). Taken together, the protocols seemed to evoke largely positive, robust changes in all four ensembles on both days. Though results are specific to the four choirs in this investigation, they seem to align with data in previous investigations related to possible benefits of straw phonation in choral settings.
As mentioned previously, choristers in previous investigations have evidenced less robust acoustic changes after straw phonation when they had already engaged in some group singing prior to the study protocols. We began our protocols immediately upon having participants enter the room. Rehearsals for these groups began between noon and 3:30 in their respective time zones. Presumably, at least some of the choristers had vocalized (either singing or speaking) throughout the day. However, we did not control for participant warmup level prior to their class meeting time. It seems reasonable that the robust changes might be because the groups had not sung together prior to data collection. In a previous investigation (Manternach et al., 2019), a control group performed the same protocol using an “ah” vowel rather than through a straw. Future researchers might consider using a straw phonation group and a control group who sing prior to any other voicing on a given day.
Choristers in this line of inquiry have generally stood with 24 inches between their shoulders, conforming to the “lateral spacing” that choristers have preferred in studies related to choral spacing (e.g. Daugherty et al., 2013). Due to the unique configurations of each of the rehearsal rooms, the choirs in this study stood with 6 inches (southeastern choirs) or 8.5 inches (midwestern choirs) between their shoulders. This spacing may have affected the ability of the choristers to hear themselves. Though the repeated-measures design ensured that spacing was consistent between all recordings for each choir, it is conceivable that the varied spacing of each group might have contributed to the differences after the straw protocols on both days of data collection. On the other hand, the results represent an ecologically valid data collection in each rehearsal space. These spaces are the typical rehearsal rooms for the choirs of these two institutions. As such, though the results are not necessarily generalizable across all situations, they may be transferrable to similarly constituted choirs who rehearse in similar spaces.
Each choir had a slightly different composition on the first and final days of data collection. These differences related to the normal fluctuation of attendance due to multiple factors (e.g. sickness, appointments, students withdrawing from courses) that take place over the course of an academic semester. There were also inconsistencies in the individual choristers between the 2 days. For example, some singers may have been in vocal distress or experiencing illness or vocal fatigue on one or both days. Others may have been working on certain vocal techniques over the weeks of the study. We were not able, nor did we attempt, to control for these inevitable occurrences. We may have ameliorated these history and maturation effects by the repeated measures design in which we compared the choirs to themselves on each day. However, an ideal study would have the same vocalists in similar states of vocal health. Future researchers may more carefully monitor attendance and might give a brief voice health survey (e.g. Singing Voice Health Handicap Index, Evaluation on the Ability to Sing Easily) to ascertain choristers’ perceived vocal health on the days of data collection.
In addition, we collected only acoustic data and perceptual data from the choristers themselves. Future researchers may wish to survey listeners (either expert or novice) to determine their preferences for various aspects of choral sound (e.g. timbre, blend, energy, intonation). Such data would further illuminate any changes that might have taken place specifically due to the straw protocols.
Finally, the choristers in this line of inquiry have utilized only small stirring straws (2.5–3.0 mm opening). We specifically designed this investigation to determine whether choristers might become acclimated with the intense pressure associated with stirring straws. Anecdotally, many choristers reported easier voicing, a fuller and clearer sound, and tuning improvements. Other singers, however, still reported feeling tightness in neck muscles, straining while voicing through the straw, and hindered sound after removing the straw. We also gave instructions on the straw protocol without individual coaching to determine whether singers are using the straw as is intended. It is possible that some choristers would feel differently about straw phonation with such individual instruction. Future researchers may incorporate individual tutorials with straws of varied straw sizes to tease out any acoustic or perceptual differences that may arise.
Conclusion
Choristers in this investigation experienced robust acoustic boosts across the 0 to 10 kHz spectrum after participating in straw phonation protocols. They also overwhelmingly reported that they sang more efficiently and that the choir sounded better after the straw protocol. These acoustic and chorister-reported changes were largely similar between the first and final days of data collection. There was a modest increase, however, in how much choristers reported that straw phonation helped them to be warmed up on the final day compared to the first day. Future researchers can expand on these findings to determine if this difference was the result of an acclimation effect.
