Skip to main content
NCBI home page
Environment & Health logoLink to Environment & Health
. 2025 Jun 22;3(9):998–1007. doi: 10.1021/envhealth.5c00096

Speaking Different Languages Impacts Size-Resolved Exhaled Breath Aerosol Particle Emission

Xinyue Li †, Chenyu Zhu †, Qisong Xing †, Huaying Liu †, Yimeng Wang †, Maosheng Yao †,‡,*
PMCID: PMC12455340  PMID: 40995480

Abstract

Aerosol transmission plays an important role in airborne-spread diseases. The transmission variations across language-usage regions were observed during COVID-19, however the potential differences from languages on aerosol transmission are poorly understood. Here, fifty-one multilingual volunteers were recruited to speak same-semantic dialogues in three languages selected from eight different languages studied to investigate the emission characteristics of exhaled aerosol across languages. The findings revealed that the size of exhaled aerosol particles generated by speaking was predominantly concentrated below 1 μm. The emission loads of exhaled aerosols during speaking and the associated potential risk of aerosol transmission across languages showed notable discrepancies. Additionally, the individual physiological factors such as age, gender and body mass index (BMI) also jointly influenced the exhaled aerosols during speaking. The machine learning model of random forest regression further revealed that language differences had a considerably greater impact on size-resolved exhaled aerosol emission concentrations than gender, but not than BMI. Thus, different language usages can influence the emission concentrations of exhaled aerosol during speaking, thereby impacting the potential for aerosol transmission across languages. This linguistic-induced diversity of transmission potentials could have played a non-negligible role in the disparate global dissemination patterns observed in aerosol-transmitted pandemics including COVID-19.

Keywords: language usage, exhaled breath aerosols, pathogen transmission potential, size-resolved aerosol

graphic file with name eh5c00096_0005.jpg

graphic file with name eh5c00096_0003.jpg

Introduction

Pathogen aerosols can be generated from diverse settings, including wastewater treatment, municipal landfills , and industrial parks and also directly by humans. − Since COVID-19 pandemic, aerosol transmission has been extensively investigated as a primary cause of global outbreaks. Various pathogens, including influenza and SARS-CoV-2 virus, have been confirmed to cause widespread epidemic outbreaks through aerosol transmission. − Airborne transmission of infectious diseases from humans has led to significant global public health crises, such as the 1918 influenza pandemic, the SARS epidemic, and the current COVID-19 pandemic. However, notable differences exist in both the rates and features of the transmissions within different regions or nations. − These variances were not solely due to control measures and healthcare resources but also obviously impacted by other factors such as strain infectivity, population density, contact patterns, temperature, and humidity. Due to challenges in quantification and representation, demographic factorsincluding language usage, physiological conditions, and activity patternshave been overlooked, limiting the analysis and prediction of epidemic spread and progression. ,− Therefore, a comprehensive investigation of factors influencing pathogen aerosol transmission is urgently needed.

Human respiratory activities, such as breathing, speaking, sneezing, and coughing, generate and release substantial quantities of exhaled aerosols. These aerosols can transport pathogens into the ambient air, elevating both pathogen concentrations and infection risk in the environment. ,,,,− As a result, respiratory diseases may be transmitted via aerosols. ,, The production of aerosols can be unequivocally attributed to specific vocalizations. Nevertheless, most studies on exhaled aerosol emission patterns have focused on isolated syllables, shorter written expressions, or respiratory activities like coughs or sneezes. − Although speaking at normal volume (40–60 dB sound pressure level, dB-SPL) produces fewer transient exhaled aerosols than paroxysmal activities such as sneezing and coughing, the cumulative transmission risk from prolonged speaking may increase with louder volumes and specific articulation manner. ,− This phenomenon allows “superemitters” to trigger superspreading events through breathing and speaking alone. ,, Currently, significant uncertainties remain regarding interspeaker variability in aerosol emissions across languages, hindering the assessment of transmission potential and outbreak prediction in multilingual regions. Such oversights may stem from overlooked factors like language-specific influences in certain areas. Previous studies have analyzed exhaled particle size distributions and found that speech generates peak concentrations around 3.5 and 5 μm, with three dominant modes at 1.6, 2.5, and 145 μm. , However, breath particle emissions during speech in different languages remain poorly characterized, particularly at the nanoscale. ,, Consequently, the role of linguistic diversity in global disease spread remains unclear.

Building on the above findings, we investigated the differences in size-resolved exhaled aerosol emissions and the associated phonetic variations for conveying identical semantics across eight official languages from diverse countries. Simultaneously, we employed a random forest regression machine learning model to analyze how human physiological parameterssuch as age, height, weight, gender, and body mass index (BMI)influence size-resolved aerosol emissions during speech. Additionally, we compared the aerosol transmission potential of eight languages (expressing identical semantics) based on total exhaled particle emission concentrations. These findings provide critical insights into the role of exhaled aerosol in diseases transmission across linguistically diverse populations.

Methods and Materials

Volunteers’ Information

The sample size of subjects predicted by the mixed-effects regression model should be 60, approximately 10 subjects were needed for each language except Chinese and English, with a sex ratio of 1:1 (α = 0.05, the statistical power 1-β = 0.8, the effect size Cohen’s f = 0.25, medium effect). However, due to the constraints imposed by epidemic control policies and language proficiency screening criteria, a total of 51 healthy multilingual Chinese volunteers (19 males and 32 females, aged 19–25 years, mean BMI = 21.07 ± 3.01, SD) who majored in foreign languages were recruited. All participants demonstrated native-like fluency in both Chinese and English language, in addition to at least one of the other foreign languages (Russian/French/Arabic/Spanish/Japanese/Korean). The language-specific breakdown of participants was as follows: 11 Russian language participants (4 males, 7 females), 9 French language participants (3 males, with one male participant also participating in the Russian testing; 6 females), 9 Arabic participants (6 males, 3 females), 4 Spanish participants (1 male, 3 females), 11 Japanese participants (5 males, 6 females), and 8 Korean participants (1 male, 7 females). Prior to the experiment, all participants provided written informed consent and completed a baseline demographic questionnaire, including their gender, age, height, and weight. The specific information is shown in Table S1. Prior to the commencement of the experiment, the participants were thoroughly acquainted with their test text and had received a concise explanation of the experiment’s purpose and safety protocols. Ethical approval for this research was obtained from the Centers for Disease Control and Prevention (CDC) of Chaoyang District, Beijing, China (Approval No. CYCDPCiRB-20220214_I). Conduct of all studies adhered to the relevant regulations and guidelines set forth by the Institutional Review Board.

Language Speaking Materials and Experiment Setup

To compare variances in exhaled aerosol emissions during the expressing of identical semantics across different languages in daily conversations, ten scenario-based dialogues involving business travel were selected as test texts from the Cambridge BEC (Business English Certificate) textbook (total English word count ≈1500; see Supporting Information S1). Volunteers were invited to translate the English text into the equivalent semantic texts of other languages. The translations were appropriately modified and corrected based on the idiomatic expressions and grammar of the corresponding languages, aiming to be as close as possible to the colloquial expressions of each language. The translated texts are provided in Supporting Information S2–S8, with an average duration of 8–15 min for each dialogue. During the testing, the volunteers were asked to naturally retell the aforementioned language materials in Chinese, English, and another foreign language they were proficient in. Their voices were recorded in the testing process using a digital voice recorder (Sony, ICD-PX470) placed 5 cm away from their mouths in a quiet conference room.

To detect exhaled aerosols produced during daily conversations in different languages with the same semantics, we used a portable nanoparticle spectrometer (NanoScan SMPS Nanoparticle Sizer 3910, TSI Inc.) and a portable optical particle counter (OPC 1.108, Grimm Inc.) jointly in a Class 100 cleanroom. The particle size spectra of exhaled aerosols from participants were observed simultaneously, as displayed in Figure S1. The NanoScan SMPS Nanoparticle Sizer 3910 is a portable spectrometer for measuring nanoparticle size. It features 13 channels and can measure particles ranging from 11.5 to 365.2 nm in a cyclic scanning mode. The upper concentration limit is 1 × 106 particles/cm3, and it has a sample flow rate of 1 L/min with a time resolution of 1 min. An integrated system comprising a nonradioactive corona-jet monopole charger, a radial differential mobility analyzer (rDMA), and a condensation particle counter (CPC) with isopropanol as the working solution was used to quantify particulate matter. Since the temporal resolution of NanoScan, each subject completed 8–15 scans depending on the duration of the testing of each language and the dwell time for each particle size channel was approximately 4–5 s. The OPC is a micrometer-sized particle size spectrometer, based on the Mie scattering theory. Operating at a flow rate of 1.2 L/min, this system detects and records the real-time size and concentration of aerosol particles. The instrument comprises 15 measuring channels spanning from 0.35 to 22.5 μm, with the highest limit for measured concentration being 2 × 106 particles/L and a time resolution of 6 s. Given that the refractive index of human respiratory mucus (1.76) is proximate to that of the polystyrene microsphere solution employed for usually calibration (1.60), the optical particle sizes of exhaled aerosol particles are analogous to the aerodynamic equivalent volume diameters, thereby obviating the necessity for specific calibration. To suppress the potential particle interference in the ambient air during the measurement of exhaled aerosol emissions when speaking, we selected a Class 100 clean room for carrying out the experiments. The clean room was equipped with a high efficiency air filter (HEPA) and supplied with outside air under positive pressure to maintain a minimal concentration of ambient particles. During the measurement, the indoor air aerosol particle level was significantly lower than those from the exhaled breath as shown in Figure (b). The temperature and humidity were regulated at 20 °C and 50% RH in accordance with experimental standards.

1.

1

Open in a new tab

Total concentration (a) and the size-resolved concentration (b) of exhaled aerosol particle during speech in eight languages (Chinese, English, Russian, French, Arabic, Spanish, Japanese, Korean) with the same semantics. Bars labeled with the same letter within Figure (a) are not statistically significantly different (p > 0.05), while bars without the same letter or unlabeled bars within the same group are statistically significantly different (p < 0.05). The letter “n” represents the number of volunteers for each language and gender. All the participants were trilingual speakers, can speak both Chinese and English, and at least one additional language tested in this study.

To measure exhaled aerosol particle size spectra at the nano-to micrometer scale and reduce ambient airflow interference dilution without accumulating exhaled aerosol, we used a bowl-shaped transparent polyethylene funnel that was roughly 10 cm wide and 12 cm long. The funnel covered the participants’ mouths and noses slightly, without impeding their facial movements. Two 15 cm-long rubber hoses were inserted into the funnel’s end to collect the aerosols produced during the participant’s speech. The aerosols were drawn into the NanoScan and OPC, respectively, in real-time. The methodology is illustrated in Figure S1. It should be noted that the rubber hose used in the study was not conductive silicone and may have caused minor aerosol loss of aerosol particles to electrostatic adsorption. The inner surface of the polyethylene funnel was not perfectly smooth and may have caused some perturbation in the exhaled airflow into the monitoring instrument.

Data Processing and Statistical Analysis

The language-speaking audios recorded were parsed and converted into frequency distribution curves based on sound pressure level formula via Adobe Auditon software. The recorded mp3 audio files (bit rate: 128 kbps) were parsed using Adobe Auditon software to obtain the frequency distribution curves of the decibel digital signals (dB Full Scale) for each audio segment. The voltage-class sound pressure level formula

Lp=20×log10(Prms/Pref)

was applied to calculate the audio digital signal amplitude (Prms), then

Prms=Pref×10(Lp/20)

where Pref, the reference sound pressure, was set to 16 bit bit depth as a reference value (Pref = 2 16 ). The exported digital signal sound pressure level data from Audition were then substituted into the formula to approximate the actual amplitude of the audio’s digital signal. This method was used to analyze the frequency-resolved loudness differences across languages at the same semantics.

Statistical analysis was conducted on the differences in size-resolved aerosol emission concentrations and total concentrations corresponding to eight country official languages. The particle size distributions of exhaled aerosol emissions when speaking the same meaning in the eight kinds of languages were acquired by calculating each participant’s mean exhaled aerosol emission concentration at the time of speaking. The particle size distributions of exhaled aerosol emitted by the participants did not exhibit normal distribution. Therefore, the differences in size-resolved concentration and total emission concentrations among individuals speaking different languages were analyzed using the nonparametric Kruskal–Wallis one-way ANOVA. A p-value < 0.05 was considered statistically significant at the 95% confidence level. The statistical tests were performed using online Python compiler.

The Random Forest Regressor model, a machine learning approach, was utilized to construct a predictive model comparing the effects of language type and individual physiological factors on exhaled aerosol emissions during speech. The Random Forest Regressor model predicts the target variables by integrating multiple decision trees, exhibiting good fitting ability and strong robustness, making it suitable for data sets with multiple features and complex nonlinear relationships. The feature parameters in this study included language, gender, age, height, weight, and BMI, and the target parameter was the size-resolved speaking exhaled aerosol emission concentration. The model was constructed through the coding of the categorical variables as numeric variables. The data set was randomly partitioned into a training set (80%) and a test set (20%). The random division of the data was implemented to ensure that the training and test sets were representative of the overall data distribution during model training and evaluation. The hyperparameter n_estimators in the model was set to 100, signifying that the model utilizes 100 decision trees for prediction; random_state was set to 42 to ensure the reproducibility of the model results. The importance score of each feature was extracted by the feature_importances attribute of the random forest regression model, allowing the contribution of each feature to the predictive power of the model to be determined. The importance of each feature is indicative of its impact on the model’s prediction objectives; a higher score on this attribute indicates a greater influence on the prediction goal, providing insights into the key factors affecting aerosol emissions.

Results and Discussion

Size-Resolved Exhaled Aerosol Concentration and Transmission Potentials of Speaking Different Languages

Human exhalation activities (breathing, speaking, coughing, sneezing, etc.) generate aerosol particles of varying size, these particles consist of subdroplets and supermicron droplets formed by the evaporation of droplets generated by respiratory activities. ,,, They can carry 102 –1011 copies/mL of viruses with particle sizes ranging from 0.02 to 10 μm. − ,− The study here quantified the emission characteristics of exhaled aerosols from 51 participants while speaking different languages with the same semantic meaning. Figure shows the total concentrations (a) and particle size distribution of aerosol emissions (b) from these eight languages with same semantics. The analysis of the speaking-exhaled aerosol revealed that the majority of the particles were concentrated below 1 μm, which can remain suspended in the air for long periods. Therefore, aerosol transmission is more likely to occur when pathogens are carried in exhaled aerosols generated during speaking. And the typical nonwater content in an exhaled particle generated in the respiratory tract is reported to be 1%–10%. The number of viruses present in exhaled aerosol particles is contingent upon particle size. The probability that a particle size within the range of 3–50 μm contains at least one virus is reported to range from 0.01% to 37%. This factor should be considered when assessing the risks of airborne transmission. The size distribution and concentrations of exhaled aerosols from different languages varied significantly even within the same semantic. As Figure a shows, when the same group of people spoke different languages, notable variation in both size-resolved particle emissions and total concentrations were observed, especially for Japanese and Korean. The study found the following ranking of aerosol particle emission levels in descending order: Korean, Japanese, English, Chinese, Spanish, French, Arabic, Russian. Table S2 shows the detailed aerosol concentrations for each language.

The diameter of the virus is generally 20–300 nm, so only aerosol particles larger than 20 nm in exhaled aerosols contain the virus and have the potential to cause aerosol transmission. Consequently, in order to make a more accurate comparison of the aerosol transmission potential of different languages under the same semantics, recalculations of the aerosol emission concentration larger than 20 and 90 nm in different languages were conducted, considering the minimum particle size of the aerosol transmission viruses, and the particle size of SARS-Cov-2. , As shown in Figure a, the ranking of exhaled aerosol concentrations at >20 nm or >90 nm among the different languages was generally consistent with the ranking of total concentrations, with Japanese, Korean and English emitting significantly higher concentrations than Arabic, Spanish, Russian and French. However, the total and >20 nm concentrations of Chinese were higher, second only to English, but the >90 nm concentration of Chinese was relatively lower, even lower than that of Russian at >90 nm (as Table S2 shown). This finding indicates that the risk of aerosol-transmission pathogens varies among different languages, despite speaking the same semantic content. Furthermore, the aerosol transmission risk is contingent upon the particle size of specific pathogens. Nevertheless, disparities in speech speed and information density can lead to variations in speaking duration, which can also lead to significant differences in the risk of aerosol transmission caused by different languages. In this study, the durations required for eight languages to express 10 dialogues with the same semantic are listed in Table S2. Due to the inability to test the speaking airflow of the volunteers’ speaking durations in the experiment, it is assumed that the average airflow of the volunteers while speaking was 12 L/min, which is used to calculate the cumulative aerosol emissions. As demonstrated in Table S2, the cumulative emissions of exhaled aerosols in the eight languages are largely consistent with the total concentration ranking. However, when considering exhaled aerosols with particle sizes greater than 20 and 90 nm, the cumulative emissions of Arabic and Russian are higher due to longer speaking durations, and are almost equivalent to the cumulative emissions of Japanese and Korean within the particle size >90 nm. Conversely, Chinese and English exhibit comparatively reduced cumulative emissions due to their comparatively shorter speaking durations. Notably, Chinese exhibits the lowest cumulative emissions among the eight languages in the particle size range greater than 20 and 90 nm. The results suggest that, in terms of emission concentration and the risk of instantaneous aerosol exposure, Korean and Japanese languages demonstrate the highest aerosol transmission potential, while Arabic, Spanish, Russian, and French languages exhibit relatively lower aerosol transmission potential. However, when the cumulative effect of speaking duration is taken into consideration, Russian and Arabic languages exhibit a significantly increased aerosol transmission potential, attributable to the longer time required to articulate the same semantic content. It is noteworthy that Chinese language has the lowest aerosol transmission potential among the eight languages due to its lower emission concentration of aerosols with particle sizes greater than 20 and 90 nm, and its shortest speaking duration. Consequently, there are substantial discrepancies in both the instantaneous and cumulative potential of aerosol transmission between languages with equivalent semantics, which are influenced by the manner of speech and language characteristics.

Japanese and Korean had comparable levels of total exhaled aerosol emissions, which were significantly higher than those of the other six languages in the particle size range of 10–154 nm, as illustrated in Figure b. This unusually high emission rates of exhaled aerosols with particle size smaller than 154 nm compared to the other languages may stem from the frequent use of vowels and more syllables required to convey the same semantic meaning in Japanese and Korean. Japanese and Korean languages were found to have much higher vowel-use frequencies than English. − Vowels are generally louder and emit more aerosols than consonants even at the same loudness. , Additionally, Japanese and Korean exhibited significantly higher speed levels and much lower information density. These suggest they have to express the same meaning with more syllables and at a faster speaking rate than other languages. As a consequence, Japanese and Korean exhibit significantly higher exhaled aerosols concentrations in the particle size range of 10–154 nm than the other languages analyzed. Although there were differences observed for individuals speaking the same language, they were smaller compared to those caused by speaking different languages with the same semantics, especially for Japanese and Korean.

All 51 participants were native Chinese speakers and proficient in English. As Figure b illustrates, the overall levels of exhaled aerosol emissions in Chinese and English are comparable. However, the emission concentration of English in the particle size range of 86.6–205.4 nm was higher than that of the other seven languages, while the emission concentration of Chinese in the particle size range of 350–575 nm was lower than that of the other seven languages. The proportion of vowels in Chinese is similar to that of Japanese, and much higher than that of English. However, Chinese has greater information density and a slower speech rate than English. , Under the combined influence of the above and various potential factors, the aerosol emission levels of English and Chinese here were found to be relatively close. However, notable differences in emission concentrations were observed within specific particle size ranges.

The emission concentrations of Arabic, Spanish, Russian, and French in the range of 10–64.9 nm were considerably lower than those of Korean, Japanese, English, and Chinese. However, in the particle size range of 450 nm to 22.5 μm, the exhaled aerosol emission concentrations in Arabic, French, and Russian were significantly higher than those of Korean, Japanese, English, and Chinese. Additionally, Spanish was only marginally elevated in comparison to other languages in the size of 205.4–365.2 nm, with lower concentrations observed across all other particle size ranges compared to the other seven languages. The variation in the size-resolved exhaled aerosol emission concentrations among different languages may be attributable to linguistic differences, including vowel-consonant usage frequency, distinctive pronunciation patterns, and other linguistic characteristics. For example, Russian and Arabic are consonant-based languages due to their phonetic nature, with little vowel proportion in daily speech compared to vowel-based languages like English and Chinese. Furthermore, the phenomenon of vowel weakening shortens the duration of vowels in Russian, resulting in a decrease in the concentration of exhaled aerosol emissions. , Arabic has no vowel letters and is dominated by oral sounds, such as consonants and guttural sounds. French and Arabic languages exhibit distinctive phonetic characteristics, including retroflex consonants.

Although exhaled aerosol concentrations were monitored in a Class 100 cleanroom, the ambient background aerosol concentration should not be disregarded. The background aerosol concentration in the cleanroom is illustrated in Figure b. The background aerosol concentration is at its zenith at 115.5 nm, with approximately 100 particles/L. The background aerosol in the environment can introduce some errors in the exhaled aerosol monitoring. Nevertheless, the background aerosol concentration levels were found to be significantly lower than the exhaled aerosol concentrations across the particle size range that were monitored in the study, and the interference with human exhaled aerosol concentration monitoring is minimal.

Acoustic Phonetics Diversities Among Different Kinds of Languages

The primary difference between languages at physical level stems from the variations in voice vibration frequency. The vibration and rupture of respiratory mucus are major mechanisms for the production of exhaled aerosols. ,, Thus, voice frequency bridges the gap between language use and exhaled aerosol emissions during speech, with its characterization primarily influenced by language type and expressive conventions. − Moreover, the effect of voice frequency differences on exhaled aerosol emissions was found more powerful compared to physiological parameters such as gender, height and weight of the speaker.

The human body can control the respiratory tract and airflow to produce voice with varying frequencies and loudness (usually expressed in dB-SPL). Meanwhile, mechanical forces act on various fluids lining the respiratory tract to produce size-resolved exhaled aerosols. As a result, the frequency distribution of speaking is likely to determine the exhaled aerosol size characteristics. Previous studies have typically reported a positive correlation between the loudness of speech and exhaled aerosol emission loads. , However, research exploring the influence of different sound frequencies remains limited. Although acoustic and phonological studies have investigated differences in frequency distributions between different languages or variations in aerosol emissions for specific words and phrases; , there is a relative lack of experimental data comparing aerosol emissions from multiple languages during everyday conversations with the same semantics. ,

To address this gap, we recorded participants’ speaking audio during the detection of exhaled aerosol emissions, and analyzed the frequency-resolved loudness distributions. It is important to note that sound pressure level (dB-SPL) is a logarithmic unit used to record loudness from an auditory perspective. Consequently, it is not suitable for directly comparing the energy differences of frequency-resolved sounds. Therefore, in this study, the voltage class sound pressure level formula was used to process the audio recordings of volunteers speaking. Furthermore, the dB Full Scale (dB-FS) of the recordings was converted into linear amplitude units. This conversion enabled intuitive comparison of frequency-resolved differences in response during speech.

As illustrated in Figure , the peak values and amplitudes of frequencies are positively correlated with the exhaled aerosol concentrations shown in Figure . It is evident that languages such as French, Spanish and Russian, which exhibit lower levels of emissions, are characterized by bimodal distributions, with frequency amplitudes marginally greater than 400. In contrast, Arabic, despite having a peak amplitude that approaches 400, displays a unimodal distribution. As illustrated in Figure , speaking Chinese or English with the same semantic meaning have minor discrepancies in exhaled aerosol emission concentration but differences in particle size distribution of concentrations. Correspondingly, Figure a,b show slight variations in the frequency distributions for speaking Chinese and English: the peak amplitude values are similar, but the peak frequencies differ slightly as shown in Table S3. The frequency distributions of Japanese and Korean in Figure c,d are multimodal, with peak amplitudes exceeding 500, which are significantly higher than those of other languages and consistent with the high emission concentrations of Japanese and Korean. Audio frequency distributions during speaking French, Spanish, and Russian follow similar patterns of aerosol emission characteristics as shown in Figure . The audio features of Russian and Arabic suggest that exhaled aerosol emissions are not solely affected by amplitude, but that the precise frequency distribution of the sound may also play a significant role. Frequency-specific sounds may generate greater concentrations of aerosol particles under identical energy conditions. − ,, Hence, direct correlation between the amplitude of the peak frequency of a sound and the concentration of exhaled aerosol emissions cannot be simply established. Further investigation is needed to understand the mechanisms of voice generation at different frequencies and the corresponding exhaled aerosol production potential.

2.

2

Open in a new tab

Frequency distribution of participants’ speech when speaking eight different languages (Chinese, English, Russian, French, Arabic, Spanish, Japanese, Korean) with the same semantic meaning. “n” represents the number of volunteers for each language and gender.

Influencing Factors of Exhaled Aerosols Emission by Speaking

For different languages, there are noticeable differences in the size-resolved concentration and cumulative exhaled aerosol volume when expressing the same semantic content, as illustrated in Figure . This disparity is influenced not only by linguistic characteristics across languages, such as vowel proportions, frequency of special consonant usage, expression styles, and idiomatic expressions, but also by individual physiological factors of the speakers. For example, factors such as gender and body weight can impact aspects such as respiratory volume, speaking frequency, and manner, thereby affecting the aerosol emission characteristics of certain languages. Research has shown that speech volume, gender, body weight, and other individual traits can affect exhaled aerosol emissions. ,, Although this study minimizes the impact of individual variations on aerosol emissions for each language by selecting multilingual participants, inherent differences in exhaled aerosol emissions caused by nonlinguistic factors (e.g., physiological characteristics) still require validation.

The Random Forest Regression model in Table indicates that the importance of influencing factors of language varieties and individual physiological factors (e.g., BMI, height, weight, gender, age) on exhaled aerosol emission concentrations varies according to the range of particle size. In the range of particle sizes studied (11.5 nm-22.5 μm), gender was found to be the least important factor, suggesting that gender may not play a substantial role in influencing exhaled aerosol emissions. Instead, gender may exert an indirect influence on these emissions, primarily acting through BMI. ,, In the 10–154 nm particle size range, BMI exhibited the highest importance, with language differences demonstrating a marginal decrease in importance relative to BMI, yet remaining comparable to height or weight and significantly higher than age. In the 575–3500 nm range, the importance of age gradually increased, with BMI following closely. However, in the 4500–12,500 nm range, the importance of BMI reappeared to be higher than that of age. The importance of language differences in the 570–12,500 nm range also increased with particle size. While the importance of language differences remained lower than that of BMI and age, it was substantially higher than the individual importance of height or weight. It can be revealed that the influence importance of language variety was lower than that of BMI or age in the majority of the particle size ranges in this study, but it still made a notable contribution to exhaled aerosol emission.

1. Importance of Physiological Factors and Language on the Size-Resolved Exhaled Aerosols Calculated by the Random Forest Regressor Model.

feature names 11.5 15.4 20.5 27.4 36.5 48.7 64.9 86.6 115.5
BMI 0.350 0.373 0.353 0.356 0.345 0.317 0.304 0.249 0.245
height 0.201 0.199 0.183 0.224 0.198 0.171 0.169 0.222 0.256
language 0.186 0.163 0.163 0.117 0.144 0.192 0.190 0.226 0.194
weight 0.153 0.152 0.188 0.188 0.192 0.229 0.252 0.191 0.196
age 0.100 0.090 0.095 0.102 0.110 0.072 0.063 0.091 0.084
gender 0.010 0.022 0.017 0.012 0.011 0.019 0.021 0.021 0.025
MSE 4,637,543 4,940,307 1,896,676 1,253,371 1,114,223 520,282 234,910 108,037 74,888
R 2 0.318 0.365 0.366 0.480 0.397 0.320 0.314 0.281 0.495
feature names 154 350 365.2 575 725 900 1300 1800 2500
BMI 0.285 0.313 0.165 0.335 0.273 0.290 0.248 0.221 0.219
height 0.287 0.159 0.186 0.119 0.106 0.100 0.125 0.146 0.129
language 0.143 0.131 0.260 0.152 0.150 0.179 0.186 0.183 0.186
weight 0.191 0.114 0.273 0.073 0.086 0.098 0.125 0.143 0.125
age 0.070 0.269 0.096 0.309 0.370 0.312 0.297 0.288 0.311
gender 0.024 0.014 0.021 0.011 0.015 0.021 0.019 0.021 0.030
MSE 57,381 13,942 23,387 1539 134 59 194 39 50
R 2 0.550 0.131 0.048 0.528 0.613 0.614 0.563 0.544 0.632
feature names 3500 4500 6250 8750 12,500 17,500 total particles >20 nm particles >90 nm particles
BMI 0.258 0.357 0.356 0.333 0.351 0.170 0.351 0.299 0.261
height 0.113 0.068 0.060 0.045 0.090 0.134 0.192 0.212 0.282
language 0.208 0.234 0.278 0.302 0.223 0.324 0.164 0.161 0.174
weight 0.109 0.093 0.081 0.085 0.132 0.142 0.180 0.215 0.167
age 0.286 0.231 0.213 0.216 0.181 0.210 0.088 0.086 0.105
gender 0.027 0.017 0.012 0.019 0.023 0.021 0.024 0.027 0.011
MSE 2.09 0.47 0.60 0.05 0.04 0.01 74,106,805 14,231,597 514,638
R 2 0.560 0.574 0.599 0.507 0.568 0.011 0.447 0.475 0.492
Open in a new tab
a

MSE (mean square error). A smaller the MSE indicates better model performance (lower prediction error) and improved fitting accuracy. R 2 quantifies the proportion of variance in the dependent variable explained by the model. Values closer to 1 indicate stronger explanatory power. While negative R 2 values suggest that the model performs worse than random guessing. The R 2 values for particle sizes of 205.4 nm, 273.8 nm, 450 and 22,500 nm are negative; therefore, these values are excluded from the table. While For particle sizes of 350 nm, 365.2 and 17,500 nm, the R 2 values are close to 0, indicating negligible explanatory power, and are not discussed further in this study.

The total aerosol emissions for most participants were concentrated below 10,000 particles/L when speaking (Table S1). However, a notable 17.65% of participants (7 females and 2 males) exhibited significantly higher concentrations (>20,000 particles/L). These individuals, whose exhaled aerosol emission concentrations far exceeded those of others, are classified as “super-emitters” in previous studies. ,, Superemitters can release aerosol particles well above the mean value and are more prone to cause superspreading events of aerosol-borne illnesses. The causes of superemitters remain controversial, , highlighting the need to investigate additional potential factors contributing to this interindividual variation. In our study, participants with emission intensities exceeding 20,000 particles/L were classified as superemitters. However, the results in this study indicated that BMI alone did not explain superemitters status, as only one superemitter exceeded the normal BMI (BMI >24). Interestingly, superemitter occurrence and emission levels exhibited linguistic dependence: all the superemitters were identified from Japanese- (4/11) and Korean-speaking (5/8) participants. Furthermore, substantial variations in total aerosols concentrations were observed when the same superemitter spoke three different languages. For instance, three Korean and one Japanese superemitters did not exhibit superemitter behavior when speaking Chinese or English (Table S1). This pattern indicates that superemitter emergence depends not only on individual physiological factors but also on linguistic characteristics. This can provide a supplementary explanation the repeated superspreading outbreaks in Japan and South Korea during COVID-19 epidemic, despite the early implementation of rigid control measures. − Conversely, no superemitters were found among participants speaking French, Spanish, Russian, and Arabic.

Current research lacks studies examining disparities in aerosol-borne epidemic transmission rates and potentials (e.g., SARS-CoV-2) across linguistic regions. Some studies have explored exhaled aerosol emissions during daily conversation, considering population differences and causes. Prior research measured 0.5–20 μm particles from singers and bilingual speaker using an APS (TSI model 3321), , finding that singing emits more exhaled aerosol particles than speaking due to the sustained vocalization, higher sound pressure, frequency and airflow. While speech loudness increases aerosol emissions, no clear language-dependent differences were observed. Another research used condensation sampling of 13 COVID-9 participants repeating English phrases and songs, detecting more viral nucleic acids during speech than singing, suggesting speech may pose a higher and undervalued potential for disease transmission. The absence of observed language effect in the aforementioned studies may stem from equipment limitations in particle size detection and the comparatively shorter test material, complicating cross-linguistic comparisons. Existing studies have concentrated on activity-based differences (speaking, singing and sneezing), but few studies examine language-specific disparities in submicron and nanometer size-resolved exhaled aerosol particles. This situation poses a significant challenge to the investigation of the language-related potential for aerosol transmission of diseases in diverse countries or language-used regions.

To address these gaps, our study analyzed size-resolved exhaled aerosol emissions and transmission potential of airborne epidemics during daily conversations with identical semantics across multiple languages, with particular focus on nanoscale speaking aerosol particles. The larger cohort (over 50 participants) and the selection of trilingual speakers minimized individual variability. Previous work has suggested that detecting 2–3-fold differences requires a cohort of 25 or more participants due to interpersonal variation. Our mixed-effects model recommended 10 gender-balanced volunteers per language (all proficient in both Chinese and English), totaling 60 participants. However, potential limitations include the participants’ native Chinese background, the limited sample sizes for some languages, and gender imbalance.

Conclusions

In conclusion, this study demonstrates that although exhaled aerosol emissions during speech are primarily influenced by individual physiological factors (e.g., BMI and age), linguistic differences also contribute non-negligibly to exhaled aerosol emission concentrations. These linguistic variations likely played a notable role in the considerable differences observed in the transmission features across countries or different language-use regions during COVID-19 pandemic. Future epidemiological forecasting and modeling of aerosol transmission should incorporate language usage data to improve the assessment of transmission risks associated with aerosol-borne disease, thereby enabling more targeted implementation of necessary preventive and control measures or policies.

Supplementary Material

eh5c00096_si_001.pdf (649.4KB, pdf)

Acknowledgments

This research was supported by the National Natural Science Foundation of China (NSFC) Creative Research Group Funds (22221004), Guangzhou National Lab Grant (SRPG22-007), National Key Research & Development Program of China (2022YFC3702801 and 2023YFC3708200) and the Young Scientists Fund of the National Natural Science Foundation of China (No. 22406008).

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/envhealth.5c00096.

  • Equipment used for the experiment; basic information about the subjects; calculation results of aerosol transmission potential in different languages; frequency peaks and corresponding amplitudes in different languages; test texts in eight semantically identical languages (PDF)

The authors declare no competing financial interest.

Published as part of Environment & Health special issue “Grand Environmental Challenge: Indoor Air Pollution, Health Effects, and Mitigation”.

References

  1. Li X., Chen H., Yao M.. Microbial emission levels and diversities from different land use types. Environ. Int. 2020;143:105988. doi: 10.1016/j.envint.2020.105988. [DOI] [PubMed] [Google Scholar]
  2. Liang Z., Yu Y., Wang X., Liao W., Li G., An T.. The exposure risks associated with pathogens and antibiotic resistance genes in bioaerosol from municipal landfill and surrounding area. J. Environ. Sci. 2023;129:90–103. doi: 10.1016/j.jes.2022.09.038. [DOI] [PubMed] [Google Scholar]
  3. Zhang S., Liang Z., Wang X., Ye Z., Li G., An T.. Bioaerosols in an industrial park and the adjacent houses: Dispersal between indoor/outdoor, the impact of air purifier, and health risk reduction. Environ. Int. 2023;172:107778. doi: 10.1016/j.envint.2023.107778. [DOI] [PubMed] [Google Scholar]
  4. Zhang T., Chen Y., Cai Y., Yu Y., Liu J., Shen X., Li G., An T.. Abundance and cultivable bioaerosol transport from a municipal solid waste landfill area and its risks. Environ. Pollut. 2023;320:121038. doi: 10.1016/j.envpol.2023.121038. [DOI] [PubMed] [Google Scholar]
  5. Johnson G. R., Morawska L.. The mechanism of breath aerosol formation. J. Aerosol Med. Pulm. Drug Deliv. 2009;22(3):229–237. doi: 10.1089/jamp.2008.0720. [DOI] [PubMed] [Google Scholar]
  6. Johnson G. R., Morawska L., Ristovski Z. D., Hargreaves M., Mengersen K., Chao C. Y. H., Wan M. P., Li Y., Xie X., Katoshevski D.. et al. Modality of human expired aerosol size distributions. J. Aerosol Sci. 2011;42(12):839–851. doi: 10.1016/j.jaerosci.2011.07.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Asadi S., Wexler A. S., Cappa C. D., Barreda S., Bouvier N. M., Ristenpart W. D.. Aerosol emission and superemission during human speech increase with voice loudness. Sci. Rep. 2019;9(1):1–10. doi: 10.1038/s41598-019-38808-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Asadi S., Gaaloul ben Hnia N. G., Barre R. S., Wexler A. S., Ristenpart W. D., Bouvier N. M.. Influenza A virus is transmissible via aerosolized fomites. Nat. Commun. 2020;11(1):4062. doi: 10.1038/s41467-020-17888-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Asadi S., Wexler A. S., Cappa C. D., Barreda S., Bouvier N. M., Ristenpart W. D.. Effect of voicing and articulation manner on aerosol particle emission during human speech. PLoS One. 2020;15(1):e0227699. doi: 10.1371/journal.pone.0227699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Morawska L., Buonanno G., Mikszewski A., Stabile L.. The physics of respiratory particle generation, fate in the air, and inhalation. Nat. Rev. Phys. 2022;4(11):723–734. doi: 10.1038/s42254-022-00506-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Peng Y., Yao M.. Quantitatively Visualizing Airborne Disease Transmission Risks of Different Exhalation Activities through CO2 Imaging. Environ. Sci. Technol. 2023;57(17):6865–6875. doi: 10.1021/acs.est.2c08503. [DOI] [PubMed] [Google Scholar]
  12. Kang M., Wei J., Yuan J., Guo J., Zhang Y., Hang J., Qu Y., Qian H., Zhuang Y., Chen X.. et al. Probable evidence of fecal aerosol transmission of SARS-CoV-2 in a high-rise building. Ann. Int. Med. 2020;173(12):974–980. doi: 10.7326/M20-0928. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Klompas M., Baker M. A., Rhee C.. Airborne transmission of SARS-CoV-2: theoretical considerations and available evidence. JAMA. 2020;324(5):441–442. doi: 10.1001/jama.2020.12458. [DOI] [PubMed] [Google Scholar]
  14. Liu Y., Ning Z., Chen Y., Guo M., Liu Y., Gali N. K., Sun L., Duan Y., Cai J., Westerdahl D.. et al. Aerodynamic analysis of SARS-CoV-2 in two Wuhan hospitals. Nature. 2020;582(7813):557–560. doi: 10.1038/s41586-020-2271-3. [DOI] [PubMed] [Google Scholar]
  15. Ma J., Qi X., Chen H., Li X., Zhang Z., Wang H., Sun L., Zhang L., Guo J., Morawska L.. et al. Coronavirus disease 2019 patients in earlier stages exhaled millions of severe acute respiratory syndrome coronavirus 2 per hour. Clin. Infect. Dis. 2021;72(10):e652–e654. doi: 10.1093/cid/ciaa1283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Zhou L., Yao M., Zhang X., Hu B., Li X., Chen H., Zhang L., Liu Y., Du M., Sun B.. et al. Breath-, air-and surface-borne SARS-CoV-2 in hospitals. J. Aerosol Sci. 2021;152:105693. doi: 10.1016/j.jaerosci.2020.105693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Johnson T. J., Nishida R. T., Sonpar A. P., Lin Y. C. J., Watson K. A., Smith S. W., Conly J. M., Evans D. H., Olfert J. S.. Viral load of SARS-CoV-2 in droplets and bioaerosols directly captured during breathing, speaking and coughing. Sci. Rep. 2022;12(1):3484. doi: 10.1038/s41598-022-07301-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Moreno T., Gibbons W.. Aerosol transmission of human pathogens: From miasmata to modern viral pandemics and their preservation potential in the Anthropocene record. Geosci. Front. 2022;13(6):101282. doi: 10.1016/j.gsf.2021.101282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Li Y., Huang X., Yu I. T., Wong T. W., Qian H.. Role of air distribution in SARS transmission during the largest nosocomial outbreak in Hong Kong. Indoor Air. 2005;15(2):83–95. doi: 10.1111/j.1600-0668.2004.00317.x. [DOI] [PubMed] [Google Scholar]
  20. Tang S., Mao Y., Jones R. M., Tan Q., Ji J. S., Li N., Shen J. S., Lv Y., Pan L., Ding P.. et al. Aerosol transmission of SARS-CoV-2? Evidence, prevention and control. Environ. Int. 2020;144:106039. doi: 10.1016/j.envint.2020.106039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Sera F., Armstrong B., Abbott S., Meakin S., O’Reilly K., von Borries R., Schneider R., Royé D., Hashizume M., Pascal M.. et al. A cross-sectional analysis of meteorological factors and SARS-CoV-2 transmission in 409 cities across 26 countries. Nat. Commun. 2021;12(1):5968. doi: 10.1038/s41467-021-25914-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Favas C., Jarrett P., Ratnayake R., Watson O. J., Checchi F.. Country differences in transmissibility, age distribution and case-fatality of SARS-CoV-2: a global ecological analysis. Int. J. Infect. Dis. 2022;114:210–218. doi: 10.1016/j.ijid.2021.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. World Health Organization (WHO) . WHO Coronavirus (COVID-19) Dashboard. https://data.who.int/dashboards/covid19/cases?n=c (accessed Oct 10, 2024).
  24. Merow C., Urban M. C.. Seasonality and uncertainty in global COVID-19 growth rates. Proc. Natl. Acad. Sci. U.S.A. 2020;117(44):27456–27464. doi: 10.1073/pnas.2008590117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Yao M., Zhang L., Ma J., Zhou L.. On airborne transmission and control of SARS-Cov-2. Sci. Total Environ. 2020;731:139178. doi: 10.1016/j.scitotenv.2020.139178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Dabisch P., Schuit M., Herzog A., Beck K., Wood S., Krause M., Miller D., Weaver W., Freeburger D., Hooper I., Green B.. et al. The influence of temperature, humidity, and simulated sunlight on the infectivity of SARS-CoV-2 in aerosols. Aerosol Sci. Technol. 2021;55(2):142–153. doi: 10.1080/02786826.2020.1829536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kimhi S., Marciano H., Eshel Y., Adini B.. Resilience and demographic characteristics predicting distress during the COVID-19 crisis. Soc. Sci. Med. 2020;265:113389. doi: 10.1016/j.socscimed.2020.113389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Stirland L. E., González-Saavedra L., Mullin D. S., Ritchie C. W., Muniz-Terrera G., Russ T. C.. Measuring multimorbidity beyond counting diseases: systematic review of community and population studies and guide to index choice. BMJ. 2020;368:m160. doi: 10.1136/bmj.m160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Coleman K. K., Tay D. J. W., Tan K. S., Ong S. W. X., Than T. S., Koh M. H., Chin Y. Q., Nasir H., Mak T. M., Chu J. J. H.. et al. Viral load of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in respiratory aerosols emitted by patients with coronavirus disease 2019 (COVID-19) while breathing, talking, and singing. Clin. Infect. Dis. 2022;74(10):1722–1728. doi: 10.1093/cid/ciab691. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Xie X., Li Y., Sun H., Liu L.. Exhaled droplets due to talking and coughing. J. R. Soc., Interface. 2009;6(6):S703–S714. doi: 10.1098/rsif.2009.0388.focus. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Holmgren H., Ljungström E., Almstrand A. C., Bake B., Olin A. C.. Size distribution of exhaled particles in the range from 0.01 to 2.0 μm. J. Aerosol Sci. 2010;41(5):439–446. doi: 10.1016/j.jaerosci.2010.02.011. [DOI] [Google Scholar]
  32. Fan H., Li X., Deng J., Da G., Gehin E., Yao M.. Time-dependent size-resolved bacterial and fungal aerosols in Beingg subway. Aerosol Air Qual. Res. 2017;17(3):799–809. doi: 10.4209/aaqr.2016.03.0114. [DOI] [Google Scholar]
  33. Zheng Y., Chen H., Yao M., Li X.. Bacterial pathogens were detected from human exhaled breath using a novel protocol. J. Aerosol Sci. 2018;117:224–234. doi: 10.1016/j.jaerosci.2017.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Bake B., Larsson P., Ljungkvist G., Ljungström E., Olin A. C.. Exhaled particles and small airways. Respir. Res. 2019;20(1):1–14. doi: 10.1186/s12931-019-0970-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Gregson F. K., Watson N. A., Orton C. M., Haddrell A. E., McCarthy L. P., Finnie T. J., Gent N., Donaldson G. C., Shah P. L., Calder J. D.. et al. Comparing aerosol concentrations and particle size distributions generated by singing, speaking and breathing. Aerosol Sci. Technol. 2021;55(6):681–691. doi: 10.1080/02786826.2021.1883544. [DOI] [Google Scholar]
  36. Lindsley W. G., Blachere F. M., Beezhold D. H., Thewlis R. E., Noorbakhsh B., Othumpangat S., Goldsmith W. T., McMillen C. M., Andrew M. E., Burrell C. N.. et al. Viable influenza A virus in airborne particles expelled during coughs versus exhalations. Influenza Other Respir. Viruses. 2016;10(5):404–413. doi: 10.1111/irv.12390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Alsved M., Matamis A., Bohlin R., Richter M., Bengtsson P. E., Fraenkel C. J., Medstrand P., Löndahl J.. Exhaled respiratory particles during singing and talking. Aerosol Sci. Technol. 2020;54(11):1245–1248. doi: 10.1080/02786826.2020.1812502. [DOI] [Google Scholar]
  38. Ding S., Teo Z. W., Wan M. P., Ng B. F.. Aerosols from speaking can linger in the air for up to nine hours. Build. Environ. 2021;205:108239. doi: 10.1016/j.buildenv.2021.108239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Lloyd-Smith J. O., Schreiber S. J., Kopp P. E., Getz W. M.. Superspreading and the effect of individual variation on disease emergence. Nature. 2005;438:355–359. doi: 10.1038/nature04153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Frieden T. R., Lee C. T.. Identifying and interrupting superspreading eventsimplications for control of severe acute respiratory syndrome coronavirus 2. Emerging Infect. Dis. 2020;26(6):1059–1066. doi: 10.3201/eid2606.200495. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Morawska L., Johnson G. R., Ristovski Z. D., Hargreaves M., Mengersen K., Corbett S., Chao C. Y. H., Li Y., Katoshevski D.. Size distribution and sites of origin of droplets expelled from the human respiratory tract during expiratory activities. J. Aerosol Sci. 2009;40(3):256–269. doi: 10.1016/j.jaerosci.2008.11.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Wynne A. C., Abbott B. S., Niazi R., Foley K., Walters K. B.. Experimental and modeling approaches to determine drug diffusion coefficients in artificial mucus. Heliyon. 2024;10(20):e38638. doi: 10.1016/j.heliyon.2024.e38638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Chien C. H., Theodore A., Wu C. Y., Hsu Y. M., Birky B.. Upon correlating diameters measured by optical particle counters and aerodynamic particle sizers. J. Aerosol Sci. 2016;101:77–85. doi: 10.1016/j.jaerosci.2016.05.011. [DOI] [Google Scholar]
  44. Gralton J., Tovey E., McLaws M., Rawlinson W.. The role of particle size in aerosolised pathogen transmission: a review. J. Infect. 2011;62:1–13. doi: 10.1016/j.jinf.2010.11.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Buonanno G., Stabile L., Morawska L.. Estimation of airborne viral emission: quanta emission rate of SARS-CoV-2 for infection risk assessment. Environ. Int. 2020;141:105794. doi: 10.1016/j.envint.2020.105794. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Hirose R., Daidoji T., Naito Y., Watanabe Y., Arai Y., Oda T., Konishi H., Yamawaki M., Itoh Y., Nakaya T.. Long-term detection of seasonal influenza RNA in faeces and intestine. Clin. Microbiol. Infect. 2016;22(9):813 e1–e7. doi: 10.1016/j.cmi.2016.06.015. [DOI] [PubMed] [Google Scholar]
  47. Wei J., Li Y.. Airborne spread of infectious agents in the indoor environment. Am. J. Infect. Control. 2016;44(9):S102–S108. doi: 10.1016/j.ajic.2016.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Pan Y., Zhang D., Yang P., Poon L. L., Wang Q.. Viral load of SARS-CoV-2 in clinical samples. Lancet Infect. Dis. 2020;20(4):411–412. doi: 10.1016/S1473-3099(20)30113-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. To K. K., Tsang O. T., Leung W., Tam A. R., Wu T., Lung D. C., Yip C. C., Cai J., Chan J. M., Chik T. S.. et al. Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study. Lancet Infect. Dis. 2020;20(5):565–574. doi: 10.1016/S1473-3099(20)30196-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Smith S. H., Somsen G. A., Van Rijn C., Kooij S., Van Der Hoek L., Bem R. A., Bonn D.. Aerosol persistence in relation to possible transmission of SARS-CoV-2. Phys. Fluids. 2020;32(10):107108. doi: 10.1063/5.0027844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Pan M., Lednicky J. A., Wu C. Y.. Collection, particle sizing and detection of airborne viruses. J. Appl. Microbiol. 2019;127(6):1596–1611. doi: 10.1111/jam.14278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Oh, Y. M. Linguistic Complexity and Information: Quantitative Approaches. Ph D. Dissertation; Université de Lyon: Lyon, FR, 2015; . http://cyberphon.ddl.cnrs.fr/fulltext/Yoonmi/Oh_2015_1.pdf (accessed 2024-09-10. [Google Scholar]
  53. Zuurbier M., Hoek G., Hazel P. V. D., Brunekreef B.. Minute ventilation of cyclists, car and bus passengers: an experimental study. Environ. Health. 2009;8:1–10. doi: 10.1186/1476-069X-8-48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Hayden R. E.. The relative frequency of phonemes in General-American English. Word. 1950;6(3):217–223. doi: 10.1080/00437956.1950.11659381. [DOI] [Google Scholar]
  55. Delattre P.. Comparing the vocalic features of English, German, Spanish and French. Int. Rev. Appl. Linguist. Lang. Teach. 1964;2:71–98. doi: 10.1515/iral.1964.2.1.71. [DOI] [Google Scholar]
  56. Tamaoka K., Makioka S.. Frequency of occurrence for units of phonemes, morae, and syllables appearing in a lexical corpus of a Japanese newspaper. Behav. Res. Meth. Instrum. Comput. 2004;36:531–547. doi: 10.3758/BF03195600. [DOI] [PubMed] [Google Scholar]
  57. Shin, J. Vowels and Consonants. The Handbook of Korean Linguistics; John Wiley & Sons, 2015; pp 1–21. [Google Scholar]
  58. Van Mersbergen, M. ; Marchetta, J. ; Foti, D. ; Pillow, E. ; Dasgupta, A. ; Cain, C. ; Morvant, S. . Comparison of aerosol emissions during specific speech tasks. J. Voice, 2023, in Press 10.1016/j.jvoice.2023.05.004. [DOI] [PubMed] [Google Scholar]
  59. Lin H., Wang Q.. Mandarin rhythm: an acoustic study. J. Chin. Lang. Comput. 2007;17(3):127–140. [Google Scholar]
  60. Erbaugh, M. S. The Acquisition of Mandarin. The Crosslinguistic Study of Language Acquisition, 1st ed.; Psychology Press, 1993; pp 373–455. [Google Scholar]
  61. Dryer, M. S. ; Haspelmath, M. . The World Atlas of Language Structures Online; München: Max Planck Digital Library, 2013. http://wals.info/ (accessed Aug 10, 2024).
  62. Oh E. J.. Differences in vowel duration due to the underlying voicing of the following coda stop in Russian and English: native and non-native values. Speech Sci. 2006;13(3):19–33. [Google Scholar]
  63. Barnes, J. Phonetics and Phonology in Russian Unstressed Vowel Reduction: A Study in Hyperarticulation. MD. Dissertation; Boston University: Boston, MA, 2007; . https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=5e6a3b9ec0eab69183db72903ce6c24227af84c7 (accessed 2024-10-10). [Google Scholar]
  64. Altenberg E. P., Ferrand C. T.. Fundamental frequency in monolingual English, bilingual English/Russian, and bilingual English/Cantonese young adult women. J. Voice. 2006;20(1):89–96. doi: 10.1016/j.jvoice.2005.01.005. [DOI] [PubMed] [Google Scholar]
  65. Keating P., Kuo G.. Comparison of speaking fundamental frequency in English and Mandarin. J. Acoust. Soc. Am. 2012;132(2):1050–1060. doi: 10.1121/1.4730893. [DOI] [PubMed] [Google Scholar]
  66. Mennen I., Schaeffler F., Docherty G.. Cross-language differences in fundamental frequency range: A comparison of English and German. J. Acoust. Soc. Am. 2012;131(3):2249–2260. doi: 10.1121/1.3681950. [DOI] [PubMed] [Google Scholar]
  67. Gautam, S. ; Singh, L. . Developmental pattern analysis and age prediction by extracting speech features and applying various classification techniques International Conference on Computing, Communication & Automation; IEEE: Greater Noida, India, 2015, pp 83–87. [Google Scholar]
  68. Good N., Fedak K. M., Goble D., Keisling A., L’Orange C., Morton E., Phillips R., Tanner K., Volckens J.. Respiratory Aerosol Emissions from Vocalization: Age and Sex Differences Are Explained by Volume and Exhaled CO2. Environ. Sci. Technol. Lett. 2021;8(12):1071–1076. doi: 10.1021/acs.estlett.1c00760. [DOI] [Google Scholar]
  69. Lee J., Yoo D., Ryu S., Ham S., Lee K., Yeo M., Min K., Yoon C.. Quantity, size distribution, and characteristics of cough-generated aerosol produced by patients with an upper respiratory tract infection. Aerosol Air Qual. Res. 2019;19(4):840–853. doi: 10.4209/aaqr.2018.01.0031. [DOI] [Google Scholar]
  70. Edwards D. A., Ausiello D., Salzman J., Devlin T., Langer R., Beddingfield B. J., Fears A. C., Doyle-Meyers L. A., Redmann R. K., Killeen S. Z.. et al. Exhaled aerosol increases with COVID-19 infection, age, and obesity. Proc. Natl. Acad. Sci. U.S.A. 2021;118(8):e2021830118. doi: 10.1073/pnas.2021830118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Shen Z., Ning F., Zhou W., He X., Lin C., Chin D. P., Zhu Z., Schuchat A.. Superspreading SARS events, Beijing. Emerging Infect. Dis. 2004;10(2):256–260. doi: 10.3201/eid1002.030732. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Chun B. C.. Understanding and modeling the super-spreading events of the Middle East respiratory syndrome outbreak in Korea. Infect. Chemother. 2016;48(2):147–149. doi: 10.3947/ic.2016.48.2.147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Ryu S., Ali S. T., Noh E., Kim D., Lau E. H., Cowling B. J.. Transmission dynamics and control of two epidemic waves of SARS-CoV-2 in South Korea. BMC Infect. Dis. 2021;21(1):1–9. doi: 10.1186/s12879-021-06204-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Kawahara T., Ueki Y., Nawa N., Miyamae S., Hanafusa M., Goto Y., Tohda S., Fujiwara T.. Characteristics of SARS-CoV-2 super-spreaders in Japan. J. Infect. 2022;84(2):e6–e9. doi: 10.1016/j.jinf.2021.12.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Shim E., Tariq A., Choi W., Lee Y., Chowell G.. Transmission potential and severity of COVID-19 in South Korea. Int. J. Infect. Dis. 2020;93(93):339–344. doi: 10.1016/j.ijid.2020.03.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Lee H., Han C., Jung J., Lee S.. Analysis of superspreading potential from transmission clusters of COVID-19 in South Korea. Int. J. Environ. Res. Publ. Health. 2021;18(24):12893. doi: 10.3390/ijerph182412893. [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Majra D., Benson J., Pitts J., Stebbing J.. SARS-CoV-2 (COVID-19) superspreader events. J. Infect. 2021;82(1):36–40. doi: 10.1016/j.jinf.2020.11.021. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

eh5c00096_si_001.pdf (649.4KB, pdf)

Articles from Environment & Health are provided here courtesy of American Chemical Society

RESOURCES