Acoustic analyses of normal voiced and whispered Mandarin Chinese reveal significant differences in duration and intensity among the four lexical tones, differences that are moreover similar across the two speech genres. In contrast to previous claims, however, these differences among the tones are found to shrink in whisper rather than being exaggerated to facilitate perception. Furthermore, i...

Proficient bilingual listening requires differential processing of sound variation in each language context. We considered context-based processing of pitch information by Mandarin-English bilinguals, for whom pitch indicates word distinctions in one language but not the other. In an eye-tracked word-learning experiment, 58 bilinguals and 28 English monolinguals each learned English-like and Ma...

Functional Data Analysis (FDA) is used to investigate Tone 3 sandhi in Taiwan Mandarin. Tone 3 sandhi is a tone change phenomenon that arises when two low tones occur in succession resulting in the first tone being realised as a rising tone. Tone dyads T2T3 and T3T3 were compared in terms of their F0 contours and velocity profiles. No difference was found between the F0 contours of the two tone...

The current paper proposes a slack-based version of the Super SBM, which is an alternative superefficiency model for the SBM proposed by Tone. Our two-stage approach provides the same superefficiency score as that obtained by the Super SBM model when the evaluated DMU is efficient and yields the same efficiency score as that obtained by the SBM model when the evaluated DMU is inefficient. The p...

This study investigated hemispheric lateralization of Mandarin tone. Four groups of listeners were examined: native Mandarin listeners, English–Mandarin bilinguals, Norwegian listeners with experience with Norwegian tone, and American listeners with no tone experience. Tone pairs were dichotically presented and listeners identified which tone they heard in each ear. For the Mandarin listeners, ...

Previous studies have demonstrated that the second language (L2) learners’ linguistic backgrounds and L2 proficiency have an effect on their perception of L2 sounds. This paper attempts to investigate the assimilation patterns of Mandarin and Tibetan tones, and the influences of first language (L1) backgrounds and Mandarin proficiency on the perception of Mandarin tones. A total of 46 Tibetan p...

Palettized or indexed color images are the dominant type of image used on the “information super highway.” In spite of this, most image compression algorithms are designed for continuous-tone images. This paper presents a method of loss less ly compressing these palettized images, which treats each color independently. This color-plane method provides substantial compression improvement over ot...

This study examines and formulates the effect of acoustic parameters on the identification of Mandarin Tone2 and Tone3, which are timing of turning point (T_t), falling of F0 (F0_△), duration (T_d), and precursor register (Register). Three stimulus matrixes of tone tokens are set up with T_t and F0_△. Subjects identify tone tokens. In Experiment 1, matrixes are different in T_t or T_t/Td. In Ex...

This study investigates how native prosodic systems as well as L2 learning experience shape non-native tone production in terms of tone movement, a primary cue to tone identity. In an imitation task, the six Cantonese tones were produced by four speaker groups: native Mandarin speakers (tonal), native English speakers (non-tonal), native English speakers with Mandarin learning experience (L2 to...

We propose an Encoder-Classifier framework to model the Mandarin tones using recurrent neural networks (RNN). In this framework, extracted frames of features for tone classification are fed in to the RNN and casted into a fixed dimensional vector (tone embedding) and then classified into tone types using a softmax layer along with other auxiliary inputs. We investigate various configurations th...