Acoustic Influence in Mandarin

As an investigation of Mandarin voice preference, this section discusses the similarities and differences between acoustic influence in Mandarin and in other languages.

First, jitter and shimmer posed negative effect on voice preference in Mandarin (see the first and second columns in Table 14), indicating that unstable voice quality was detrimental to one’s impression of the speaker. To be more detailed, jitter was particularly found to reduce the perception of emotion, while shimmer diminished the clarity of a voice (Coelho et al., 2009).

This is similar to what has been reported in European and Brazilian Portuguese (Coelho et al., 2009), where vocal perturbation was likely to be an indication of poor health (S. M. Hughes et al., 2014) or excessive smoking and drinking habits (Babel et al., 2014). However, some cultures regard vocal perturbation as an advantage, where huskier female voices created a sexier tone in American culture (Karpf, 2006), and hoarse male voices were also considered warm and sexy in English culture (S. M. Hughes et al., 2008).

Table 14. The Effects of Acoustic Features in Each Text Emotion

Text Emotion Jitter Shimmer PitchRange PitchMean HNR Duration

Neutral – – +

Positive – + +

Negative – – + +

Note: The sign (“–”) indicated that this feature had negative influence on voice preference;

The sign (“+”) indicated that this feature had positive influence on voice preference;

The blanks indicate that such this feature had insignificant effect on voice preference.

Next, pitch range had a positive correlation with voice preference in Mandarin (see the third column in Table 14), indicating that a speaker’s intonation can be beneficial to the perception of voice. According to Pisanski et al. (2018), European female increased their pitch variability while speaking to an ideal male mate. Similarly, European male (from Italy, UK, and Czech Republic) was found to speak with more intonation as they tried to become more

attractive (Anolli & Ciceri, 2002; Leongómez et al., 2014). It is suggested that more intonation might enhance a speaker’s vocal attractiveness, where listeners are more likely to make favorable impression on the speaker.

Furthermore, pitch mean generally had a positive influence on voice preference in

emotional (positive and negative) context in Mandarin (see the fourth column in Table 14). Since most research made an investigation of pitch mean with neutral utterance (English vowels), it

and negative) content was found to be significantly affected by pitch mean. With a normal F0 range of male (from 90 to150 Hz) and female (from 190 to 250 Hz) speakers, the highest preference ratings were respectively at 130 Hz for male and 250 Hz for female. Listeners had preference for relatively “higher male and female pitch mean (F0)” in emotional content.

However, previous investigation on pitch mean (F0) used “neutral content” as their materials, where they found that “lower male pitch” (Sarah A Collins, 2000; Feinberg et al., 2005; B. C.

Jones et al., 2010) and “higher female pitch” (Bryant & Haselton, 2009; Feinberg et al., 2008; B.

C. Jones et al., 2010; Re et al., 2012) were more preferred.

Since there is an obvious difference between the preferred male pitch, we are concerning whether that comes from our choice of materials. To better compare our results with previous studies, we further calculated the averaged F0 of speakers saying neutral content in our study and listed the corresponding preference ratings beside (see Table 15). In Table 15, the information of speakers, pitch mean, and preference ratings, were arranged in the order of pitch mean (from high to low). Speakers with the highest pitch mean (F0) (see the first row in Table 15) and lowest pitch mean (F0) (see the last row in Table14) pitch were not rated with the highest preference

ratings. The most preferred F0 fell into the middle of the entire F0 range from all speakers (female: 222.13 Hz; male: 117.86 Hz). This can be referred back to the previous claim that either too high or too low of pitch would pose a negative effect on voice attractiveness (Borkowska &

Pawlowski, 2011; Re et al., 2012). However, our findings do not support other previous claims

that lower male pitch (Sarah A Collins, 2000; Feinberg et al., 2005; B. C. Jones et al., 2010) and higher female pitch (Bryant & Haselton, 2009; Feinberg et al., 2008; B. C. Jones et al., 2010; Re et al., 2012) were more preferred. For instance, both higher female pitch (e.g., Speaker F5 and Speaker F4) and lower female pitch (e.g., Speaker F2) were preferred, with their preference ratings greater than 3. Similarly, Mandarin listeners had preference for some speakers with high male pitch (e.g., Speaker M2) and low male pitch (e.g., Speaker M1), since they were also rated at nearly 3. Therefore, it is suggested that Mandarin speakers have no particular preference for relatively high or low pitch, which is slightly different from the findings in English languages (Bryant & Haselton, 2009; Sarah A Collins, 2000; Feinberg et al., 2008; Feinberg et al., 2005; B.

C. Jones et al., 2010; Re et al., 2012). However, it was consistent that both English and Mandarin speakers disliked the voice that contained extremely high and extremely low pitch.

Table 15. Female and Male Pitch Mean in Neutral Content

The last significant acoustic feature is HNR, which showed a relatively small contribution to voice preference since it only posed its contribution in negative content (see the fifth column in Table 14). Previous investigation of HNR failed to reach a consensus on its influence on voice preference. Some claimed that vocal breathiness (lower HNR) could enhance our preference for voice due to the softer vocal texture (Pisanski et al., 2016; Šebesta et al., 2017; Titze, 1989; Van Borsel et al., 2009), and created a more alluring impression for male listeners (Guzman et al., 2013). Others introduced an opposing view that breathiness (lower HNR) was detrimental to voice preference because of the production of noise in phonation. Our findings in Mandarin were

Female Pitch Male Pitch

Speaker

ID PitchMean Preference Speaker

ID PitchMean Preference

Note: The bold number is the most preferred voice; The underlined is the highest pitch;

The dotted underlined is the lowest pitch.

more consistent to the latter, where HNR enhanced our preference for the voice. From both statistical and qualitative response, sonorous vocalization (i.e., less noise emission or higher HNR) was well-preferred by listeners, where they claimed that noises and breathy sounds

prompted them to grade the voices with lower preference (see the second column in Table 16).

Table 16. Qualitative Response: Voice Quality

Q: what makes you dislike the voice?

creaky noise weak others Total

26 23 10 2 61

43% 38% 16% 3% 100%

Note: Forty participants were allowed to choose more than one voice quality.

Others include “nasal voice” or “pronunciation of words.”

Aside from the aforementioned acoustic influences, duration, on the contrary, did not show any effect on voice preference. It is likely that our stimuli were designed short, so that the length of each utterance would not vary much. Only a few listeners reported that an obvious pause between each word (e.g., 他-pause-是-pause-個-pause-工-pause-程-pause-師) would

greatly lower their preference for the voice. Aside from these responses, no participant was consciously affected by the duration of the speaker’s utterance.

In sum, the acoustic influence in Mandarin were similar to other languages, as to how voices with less disturbance (low jitter and shimmer) and more intonation (high pitch range)

could increase our preference for the voices. However, the difference appeared in pitch mean (F0) and HNR. As far as F0 was concerned, a more neutral F0 of both male and female speakers in Mandarin could be more preferentially perceived than “lower male pitch” (Feinberg et al., 2005; B. C. Jones et al., 2010) or “higher female pitch” (Feinberg et al., 2008; Re et al., 2012).

Besides, a more sonorous vocalization was preferred in Mandarin, which is different from how English (Xu et al., 2013) or African (Šebesta et al., 2017) people perceived breathy voices as beneficial voice qualities.