key: cord-0726420-4dsj78c7
authors: Richter, B.; Hipp, A. M.; Schubert, B.; Axt, M. R.; Stratmann, M.; Schmoelder, C.; Spahn, C.
title: From classic to rap: Airborne transmission of different singing styles, with respect to risk assessment of a SARS-CoV-2 infection
date: 2021-03-26
journal: nan
DOI: 10.1101/2021.03.25.21253694
sha: b35b51cd46d0c59617d4b470ff05ae6fb319c8e8
doc_id: 726420
cord_uid: 4dsj78c7

Since the Covid-19 virus spreads through airborne transmission, questions concerning the risk of spreading infectious droplets during singing and music making arose. To contribute to this question and to help clarify the possible risks, we analyzed 15 singing scenarios (1) qualitative, by making airflows visible, while singing, and (2) quantitative, by measuring air velocities in three distances (1m, 1.5m and 2m). Air movements were considered positive, when lying above 0.1 m/s, which is the usual room air velocity in venue, such as the concert hall of the Bamberg Symphony, where our measurements with three professional singers (female classical style, male classic style, female popular music styles) took place. Our findings highlight, that high measurements for respiratory air velocity while singing, are comparable to measurements of speaking and, by far, less than coughing. All measurements for singing stayed within a reach of 1.5m, only direct voiceless blowing achieved measurements at the 2m sensor. Singing styles, which use plosive sounds i.e. consonants more often, such as rap, produced the highest air velocities of 0.17 m/s at the 1m sensor. Also, singing, while wearing a facemask, produces no air movements over 0.1 m/s. On the basis of, our recent studies on measurements of airflows and air velocities of professional singers and wind instrument players, as well as further studies on CO2; measurements in room settings of music activities, we publish our results, in consideration of further up-to-date research, in our frequently updated risk assessment (first published in April 2020). On this behalf, we suggest 2m radial distances for singers, especially in choirs.

Since the Covid-19 virus spreads through airborne transmission, questions concerning the risk 25 of spreading infectious droplets during singing and music making has arisen. 26

To contribute to this question and to help clarify the possible risks, we analyzed 15 singing 27 scenarios (1) qualitatively -by making airflows visible, while singing -and (2) quantitatively 28 -by measuring air velocities at three distances (1m, 1.5m and 2m). Air movements were 29 considered positive when lying above 0.1 m/s, which is the usual room air velocity in venue, 30 such as the concert hall of the Bamberg Symphony, where our measurements with three 31 professional singers (female classical style, male classic style, female popular music style) 32 took place. 33

The coronavirus pandemic has had and still has a great impact on singing, especially while 52 performing in groups, since the SARS-CoV-2 virus infection is transported through droplets 53 from the respiratory system and can emit from the mouth opening to the surrounding air [1] . 54

Hence, prevailing questions concern mostly distance regulations and the fundamental 55 questions as to whether and how individual lessons or ensemble and choir rehearsals (in 56 different settings, levels of profession, as well as amateur choirs) can take place. Since it is 57 known that music making in general, and therefore singing, has an extended influence on a 58 person's well-being [2, 3, 4], these questions concern 4 million people in Germany who sing 59 in a choir, of which 2.2 million amateur singers are registered in a choir union [5] . 60

In the following, we need to know at what distance air movements are being produced while 61 singing and how these air velocities can be evaluated in comparison to usual air movements, 62 such as speaking or coughing, as well as considering usual room air velocities. 63

The measurements of our study took place in the concert hall of the Bamberg Symphony (see 64 than speaking or breathing and that an increase of aerosol rates can be observed with an 93 increasing sound pressure level during singing, especially while holding long tones. 94

In consideration of their study, Mürbe et al. [9] , Kriegel & Hartmann [10] , and Hartmann et 95 al. [11] published various assessments on the risk of infection with the virus-loaded aerosols, 96 while singing indoors during the SARS-CoV-2-pandemic. They assume that different styles 97 of singing -e.g., singing vs. speaking -as well as different intensities of voice can lead to 98 various sizes and density of droplets and aerosols [9] and that room situations of choir 99 rehearsals have to be taken into account [10, 11] . 100 Furthermore, Mürbe et al. found out that child voices emit fewer aerosols during singing than 101 adults [12] . In this study they tested 8 children (four girls and four boys) of semiprofessional 102 children's choirs (Staats-und Domsingknaben Berlin and a girl's choir of the Berlin 103 Singakademie), who were all 13 years old (except one girl was 15 years old). The study was 104 conducted the same way as their previous study on professional adult singers as mentioned 105 above. Their mean measurements show emission rates of 16 -267 P/s for speaking 141 -106 1240 P/s for singing, and 683 -4332 P/s for shouting. 107

In 2019, before the Coronavirus infection started, the study by Asadi et al. [13] observed 108 different speaking situations regarding their particle rate of normal speech in correlation with 109 loudness. They support the hypotheses that coughing and sneezing are emphasizing airborne 110 infectious disease transmissions and point out that loudness of voice is related to the particle 111 rate of aerosols, but also assume that further factors have to be considered. 112

Asahi et. al furthermore emphasize that the generation of aerosols increases with the 113 amplitude of vocalization depending on loudness, but the language spoken (English, Spanish, 114

Mandarin, or Arabic) showed no difference. 115 6 A study conducted by the University of the German Armed Forces in Munich, which was pre-116 published in May 2020 [14] , analyzed larger droplets when singing and speaking, as well as 117 the flow-related small droplets when singing and playing wind instruments. The study was 118 conducted with a professional singer, two amateur choir singers, five professional musicians 119 (clarinet, flute, oboe, bassoon, and trumpet), and an amateur brass player (trumpet, trombone, 120 and euphonium). This study observed the motion of droplets and air leaving either mouth or 121 outlet during exhalation, which was then illuminated with laser and recorded with a digital 122 camera, producing a series of images that were quantitatively analyzed afterwards. The 123 analysis points out that while singing, airflows at a distance of 0.5m were no longer detectable 124 regardless of volume, pitch or whether the singer was a professional or an amateur. In 125 conclusion, they recommend a minimal radial distance of 1.5m between singers and/or wind 126

instrumentalists. 127

Another recent publication [15] conducted at the Bauhaus-Universität Weimar observed the 128 spread of exhaled air while playing wind instruments and singing using the schlieren imaging 129 with a schlieren mirror and the background oriented schlieren method (BOS) as a way to 130 make respiratory air visible. Two professional singers (baritone and soprano) and eleven wind 131 instruments (woodwinds: oboe, bassoon, Bb clarinet, bass clarinet, grand flute, piccolo, alto 132 flute; brass: Bb trumpet, tenor trombone, French horn, F tuba) of the orchestra Thüringen 133

Philharmonic Gotha, Eisenach, were positioned in front of the schlieren mirror while playing 134 or singing. The findings show that the spreading range, as well as the angle at which the air 135 escapes the mouth or outlet, varies strongly from instrument and player, depending on the 136 structure of the instrument, the structure of the mouthpiece, the way an instrument is blown 137 and individual blowing or breathing capacities. 138

Based on our own studies, we updated our risk assessment for the fifth time, which was first 139 published in April 2020 [17] . Herein we suggest a 2m radial distance between singers, as well 140 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. In order to exclude infectious persons, before the measurements the singers were questioned 158 as to whether they had or were still suffering from typical symptoms of the Covid-19 virus 159 disease such as a cough, loss of taste, fever or fatigue. Additionally, the temperatures of the 160 three singers were taken with an electronic fever thermometer directly before entering the 161 concert hall. No one reported suspicious symptoms and the measurements of the temperature 162

showed values under the cut-off of 37,5°C in all persons. Beyond measurements, singers 163 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 168 The measurements were conducted at the beginning of May 2020 and were initiated by the 169 Bamberg Symphony Orchestra during the first lockdown of the SARS-CoV-2-pandemic in 170

Germany. The three professional singers were singing different styles, speaking or blowing. 171

For the qualitative analysis, the measurements were also filmed on a digital camera, 172 producing videos that were qualitatively analyzed afterwards. 173

The different singing styles of the male classical baritone singer were an excerpt of a music 174 piece and performing utterances including plosives. He was also tested while wearing a 175 facemask, wearing no face mask, while speaking and while blowing directly at the sensors. 176

The female classical singer sang an excerpt of a classical music piece, tried out high pitches, 177 plosive tones and blew directly in the direction of the sensors. 178

The female non-classically trained singer sang an excerpt of a rock song, pop lyrical, musical 179 belting and rap. Furthermore, she was also measured while speaking and directly blowing at 180 the sensors. 181 which consists of water droplets and is usually used as stage fog [22] . The fog droplets have a 189 size smaller than 5µm (see Appendix 2) and can therefore be compared to the core droplets 190 potentially carrying the SARS-CoV-2 virus [1] . 191

The artificial fog was transported through a system called "Hydra" using a flexible tube to the 192 release spot of the airborne particles. Through the application tube installed on a stand, the 193 fog escaped into the free space of the room and created a cloud of fog. It was oriented towards 194 the mouth of the singers, who sang directly into the cloud. 195

The movement of the fog was filmed with a video camera and qualitatively analyzed 196 afterwards. 197 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Making use of the descriptive analysis, the measurements of the singing scenarios were also 248 put into a table (see Table 1 insight into the movement of airflows while singing, speaking and blowing. Aside from that, 263 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

282 Air velocity measurements were taken with three singers, performing a total of 15 different 283 scenarios of singing styles and speaking (see Table 1 ), with and without a facemask, while 284 singing an excerpt of a classical music piece, singing a rock song, pop lyrical, musical belting, 285 performing a rap song, using plosive tones or directly blowing at the sensors or speaking. The 286 numbers point out that directly blowing and aiming for one of the sensors, which is 287 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. to the measurements of speaking (see Table 1 and Appendices 5, 6). 293 294 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Blowing 305 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 0.12m/s, while no air movement was measured at 2m above 0.1m/s. 326

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Blowing 348 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. When comparing the measurements for singing and speaking, they show the same maximal 370 measurements of air velocities of up to 0.17 m/s at a distance of 1m, dropping to 0.12 -0.13 371 m/s at the 1.5m sensor, and staying below the usual room air velocity of 0.1m/s at the 2m 372 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Taking into account different styles of singing, we observed that they make almost no 383 difference at all in the generation of increasing airflows. A significant difference between 384 blowing and singing or speaking was observed, but not by use of different singing styles. In 385 this case, we compared mainly the highest merits of measurements of a sequence. 386

All qualitative observations were made close to the body of the singing person, which lead to 387 the presumption of low measurements at the sensors, especially in distances beyond 1.5m. 388

Quantitative measurements supported these observations, showing that respiratory air 389 movements while singing are not dependable on volume, pitch or style of singing, but rather 390 on individual singing habits and possibilities, which was also observed in the study of Becher 391 et al. [15] . 392

It is remarkable and interesting that we found differences in our assumptions after seeing the 393 qualitative observations and having the merits from the quantitative measurements. On the 394 basis of the qualitative observations only, it was assumed that plosive tones or rap singing 395 would produce higher air velocities than other singing styles. The measured results showed 396 differently: We could not find a difference if a singer is singing a song, speaking or using 397 plosive tones -all measurements stay at the same level for all three test persons. Other studies 398 found out otherwise, observing differences between different speaking and singing, especially 399 while holding long tones [8], or a difference between singing with or without text, but no 400 difference between singing or speaking text [6] . 401

Comparing the findings on respiratory air movements while singing with our own recent 402 study on woodwinds, it was found out that singers reach higher air velocities than woodwinds 403 With regard to our risk assessment [17], we suggest a radial distance regulation of 2m for 413 singers, especially when singing in groups. Only then, droplet infection due to direct contact 414 can be excluded. This recommendation also supports the fact that strong airflows can 415 transport heavy droplets even further than 1.5m. On the basis of our findings, though, 416 measurements of airflows beyond 1.5m while singing were not visible. We were furthermore 417 unable to measure jet-stream droplet effects of light droplets, which would have been visible 418 in the artificial fog. Furthermore, we suggest frequent ventilation breaks, as mentioned in our 419 study on CO measurements [18], to prevent a coronavirus infection while performing 420 musical activities. 421 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

The test results show that singing is less dangerous than coughing and comparably produces 443 the same air velocity as speaking. In our study, results do not depend on volume, pitch or the 444 style of singing. In this spectrum, the highest measurement of singing lies at 0.17m/s at a 445 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 26, 2021. ; https://doi.org/10.1101/2021.03.25.21253694 doi: medRxiv preprint distance of 1m to the singer. At 2m, measurements over the usual room air velocity of 0.1m/s 446 were no longer visible, while frontal and direct blowing, which is comparable to coughing, 447 rise up to 0.7 m/s at the most at 1m. 448

Since no respiratory air movements -of any singing style -were measured at the 2m sensor, 449 we find distance regulations of 2m to the front of the singing person to be maintainable. 450

In order to maintain a responsible risk assessment, we find it crucial that besides a regulated 451 distance setup (e.g., small or large ensemble), constant fresh air ventilation and social 452 behavior should also be taken into account. Starting from April 2020, the Freiburg Institute of (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

Size distribution and sites of origin of droplets expelled 465 from the human respiratory tract during expiratory activities

Health Promotion in Higher Music Education

Singing and Wellbeing. Ancient Wisdom, Modern Proof. London: 473 Routledge

Laienmusizieren in Zahlen -Ergebnisse 475 bundesweiter Studien und Bevölkerungsumfragen

Impulse dispersion of aerosols during singing and 479 speaking

Impulse dispersion 483 of aerosols during playing wind instruments. medRxiv

Aerosol emission is increased 486 in professional singing. DepositOnce

Beurteilung der Ansteckungsgefahr 489 mit SARS-CoV-2-Viren beim Singen

Aerosol emission of child voices during speaking, singing and shouting

Aerosol emission and superemission during human speech increase with voice 502 loudness. Sci Rep

Singing in choirs and making music with wind instruments 505 -Is that safe during the SARS-CoV-2 pandemic? ResearchGate

The spread of 508 breathing air from wind instruments and singers using schlieren techniques

Airflow and air velocity measurements while playing wind instruments, with respect 513 to risk assessment of a SARS-CoV-2 infection. medRxiv [preprint

Risk assessment of a coronavirus infection in the field of 516 music

CO measurements in 519 instrumental and vocal closed room settings as a risk reducing measure for a 520 Coronavirus infection

Experimental 525 Setup for Visualisation of Pulsating Turbulent Flows

45 Jahre SAFEX -Theaternebel

FlowMarker II -Technische Daten

Auf dem Weg zum Nullemissionsgebäude

Quelllüftung und ihre Anwendungsbereiche. Klimalüftung. 534 Fachjournal

Thermische Behaglichkeit -Komfort in Gebäuden

Analysis of thermal comfort and 539 indoor air quality in a mechanically ventilated theatre

No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted