Target Pitch Research Articles

Contextual influences on baseball ball-strike decisions in umpires, players, and controls

Baseball umpires, players, and control participants with no baseball experience were asked to call balls and strikes for video clips. In a basic judgement task, umpires and players were significantly better at calling pitches than controls. In a direct information task, borderline pitches were presented following clips of definite balls and definite strikes. Participants called target pitches closer to the strike end of the scale when viewed after definite balls than when they followed definite strikes. Similarly, when borderline pitches were shown in different pitch counts, participants called pitches more towards the strike end of the scale when there were three balls in the count (3–0, 3–2). These findings indicate that the standard for evaluation changes based on the context in which stimuli are processed. Moreover, the strength of the contextual factors is illustrated in that the effects were shown in observers with and without experience in the task. Overall, however, umpires had a greater tendency to call strikes, indicating that they may use a norm of “hastening the game”.

Dissociation of procedural and semantic memory in absolute-pitch processing

We describe two memory-retrieval systems in absolute-pitch (AP) processing and propose existence of a universal internal pitch template to which subpopulations of musicians selectively gain access through the two systems. In Experiment I, AP and control musicians adjusted the frequency of a pure tone to match the pitch of a visually displayed randomly selected musical note. In Experiment II the same subjects vocally produced within 2 s the pitch associated with a randomly selected musical note label. AP musicians, but not controls, were highly accurate in frequency matching. Surprisingly, both AP and non-AP groups were extremely accurate in voicing the target pitch as determined from an FFT of the recorded voiced notes (i.e., σ = 0.97, 0.90 semitones, respectively). Spectrogram analysis showed that notes voiced by non-AP musicians are accurate from onset of voicing suggesting that pitch accuracy does not result from an auditory-motor feedback loop. Findings support existence of two memory-retrieval systems for musical pitch: a semantic associative form of memory used by AP musicians, and a more widespread form of procedural memory which allows precise access to internal pitch representations through the vocal-motor system.

Effects of Increasing Time Delays on Pitch-Matching Accuracy in Trained Singers and Untrained Individuals

Trained singers (TS) generally demonstrate accurate pitch matching, but this ability varies within the general population. Pitch-matching accuracy, given increasing silence intervals of 5, 15, and 25 seconds between target tones and vocal matches, was investigated in TS and untrained individuals. A relationship between pitch discrimination and pitch matching was also examined. Thirty-two females (20-30 years) were grouped based on individual vocal training and performance in an immediate pitch-matching task. Participants matched target pitches following time delays, and completed a pitch discrimination task, which required the classification of two tones as same or different. TS and untrained accurate participants performed comparably on all pitch-matching tasks, while untrained inaccurate participants performed significantly less accurately than the other two groups. Performances declined across groups as intervals of silence increased, suggesting degradation of pitch matching as pitch memory was taxed. A significant relationship between pitch discrimination and pitch matching was revealed across participants.

Sound effecter, fundamental tone extraction method, and computer program

The present invention provides a technique for shifting pitch to target pitch without detecting the original pitch directly, and for extracting the pitch of the audio waveform exactly. A phase compensator extracts 2 or more frequency channels each having frequency components of a harmonic overtone whose frequency is 1 or more times as higher than frequency of a fundamental tone of the original sound, from the frequency channels from which the frequency components are extracted by fast Fourier transform. The phase compensator calculates a scaling value to be used for converting the fundamental tone to another target fundamental tone, and performs phase compensation in accordance with the scaling value. A pitch shifter performs pitch scaling in accordance with the scaling value onto the audio data resultant from inverse fast Fourier transform onto the phase compensated frequency components. Thus, audio data representing the target fundamental tone are generated.

Demonstration of an Inverted Stewart Platform Target-Suspension System Using Lightweight, High-Tensile Strings

This paper presents the design, development, and testing of an inverted Stewart platform for suspending and positioning targets during RF antenna and signature testing. Previous string target-support systems have used multiple-string-attachment-point configurations that do not allow the target roll or pitch to be modified during the azimuthal data collection. This presentation will discuss an in-house development of a scale model target-support system that allows for high-accuracy simultaneous target roll and pitch positioning. The inverted Stewart platform also offers unique stability of the target by damping out the torsional pendulum motion typically encountered in conventional string-support systems. In this paper, we will also discuss the advantages and disadvantages of the string-support concepts, and provide design guidance for building an inverted Stewart-platform support system.

Tone Features, Tone Perception, and Peak Alignment in Thai

This paper investigates the relationship between the phonological features of tone and tone perception in Thai. Specifically, it tests the hypothesis (proposed by Morén & Zsiga) that the principle perceptual cues to the five-way tonal contrast in Thai are high and low pitch targets aligned to moras. Results of four perception studies, one using natural speech and three using digitally-altered speech, are presented in support of the hypothesis. It is argued that, by associating tones to moras, a straightforward mapping from the abstract autosegmental features H and L to the production and perception of Thai tones, a heretofore elusive goal, can be accomplished. This result has consequences for theories of contour tone perception, the distinctive features of tone, and the alignment of pitch targets to the segmental string.

Interaction of word stress, focus, and sentence type in English

This study investigates prosodic differences between statements and declarative questions in American English by examining their interaction with focus, as manifested in the local and global pitch range variations, and in the underlying pitch targets of the stressed syllables in the pre-, post-, and on-focus content words. Five native speakers read 24 sentences eight times. Results of F0 analyses indicate that focus has no effect on the prefocus region for either statements or questions. In the on-focus region, the pitch range of the stressed syllable is raised in statements, but lowered in questions. The postfocus pitch range is compressed and lowered in statements, but compressed and raised in questions. Furthermore, the pitch target of the stressed syllable in a content word is high or falling in statements and low or rising in questions, depending on the focus condition and on the word’s stress and syllabic pattern. Comparisons of duration and spectral balance in each statement-question pair yielded no significant difference. These results suggest that a particular combination of word stress, focus, and sentence type in an English utterance largely determines its local and global pitch contours. Only after these communicative functions are identified can the associated forms be explained.

Determining the temporal interval of segments with the help of F 0 contours

The temporal interval of a segment such as a vowel or a consonant, which is essential for understanding coarticulation, is conventionally, though largely implicitly, defined as the time period during which the most characteristic acoustic patterns of the segment are to be found. We report here evidence for a need to reconsider this kind of definition. In two experiments, we compared the relative timing of approximants and nasals by using F 0 turning points as time reference, taking advantage of the recent findings of consistent F 0-segment alignment in various languages. We obtained from Mandarin and English tone- and focus-related F 0 alignments in syllables with initial [j], [w] and [ɹ], and compared them with F 0 alignments in syllables with initial [n] and [m]. The results indicate that (A) the onsets of formant movements toward consonant places of articulation are temporally equivalent in initial approximants and initial nasals, and (B) the offsets of formant movements toward the approximant place of articulation are later than the nasal murmur onset but earlier than the nasal murmur offset. In light of the Target Approximation model (TA) originally developed for tone and intonation [Xu & Wang. (2001). Pitch targets and their realization: Evidence from Mandarin Chinese. Speech Communication, 33, 319–337], we interpreted the findings as evidence in support of redefining the temporal interval of a segment as the time period during which the target of the segment is being approached, where the target is the optimal form of the segment in terms of articulatory state and/or acoustic correlates. This new definition may have implications for our understanding of many issues in speech, including, in particular, coarticulation and temporal coordination in speech motor control.

Expressive speech synthesis using multilevel unit selection

Generating natural sounding and meaningful prosody is a central challenge in text-to-speech synthesis, especially when generating expressive speech. Recently, we proposed a multilevel unit sequence synthesis approach, based on the general superpositional model of intonation, which describes a pitch contour as the sum of component curves that are each associated with different phonological levels, specifically the phoneme, foot, and phrase. During synthesis, segmental perturbation curves, accent curves, and phrase curves are extracted from the acoustic signal and are combined into target pitch curves; these target curves are then imposed on the acoustic unit sequences using standard pitch modification methods. This approach represents an attempt to combine the strengths of the two dominant approaches to speech synthesis: unit selection synthesis, which preserves all details of natural speech but struggles with coverage of the very large combinatorial space of phoneme sequences and prosodic contexts, and diphone synthesis, which addresses coverage by generating rule-based synthetic target prosody and imposing it on acoustic units using signal modification methods. This approach minimizes prosodic modification artifacts, optimizes the naturalness of the target pitch contour by using quasi-natural contours, yet avoids the combinatorial explosion of standard unit selection synthesis.

Prosody conversion from neutral speech to emotional speech

Emotion is an important element in expressive speech synthesis. Unlike traditional discrete emotion simulations, this paper attempts to synthesize emotional speech by using strong, medium, and weak classifications. This paper tests different models, a linear modification model (LMM), a Gaussian mixture model (GMM), and a classification and regression tree (CART) model. The linear modification model makes direct modification of sentence F0 contours and syllabic durations from acoustic distributions of emotional speech, such as, F0 topline, F0 baseline, durations, and intensities. Further analysis shows that emotional speech is also related to stress and linguistic information. Unlike the linear modification method, the GMM and CART models try to map the subtle prosody distributions between neutral and emotional speech. While the GMM just uses the features, the CART model integrates linguistic features into the mapping. A pitch target model which is optimized to describe Mandarin F0 contours is also introduced. For all conversion methods, a deviation of perceived expressiveness (DPE) measure is created to evaluate the expressiveness of the output speech. The results show that the LMM gives the worst results among the three methods. The GMM method is more suitable for a small training set, while the CART method gives the better emotional speech output if trained with a large context-balanced corpus. The methods discussed in this paper indicate ways to generate emotional speech in speech synthesis. The objective and subjective evaluation processes are also analyzed. These results support the use of a neutral semantic content text in databases for emotional speech synthesis

Production of Weak Elements in Speech – Evidence from F₀ Patterns of Neutral Tone in Standard Chinese

Many weak elements in speech, such as schwa in English and neutral tone inStandard Chinese, are commonly assumed to be unspecified or underspecifiedphonologically. The surface phonetic values of these elements are assumed toderive from interpolation between the adjacent phonologically specified elementsor from the spreading of the contextual phonological features. In the present study,we re-evaluate this view by investigating detailed F0 contours of neutral-tone syllablesin Standard Chinese, which are widely accepted as toneless underlyingly. Werecorded sentences containing 0-3 consecutive neutral-tone syllables at twospeaking rates with two focus conditions. Results of the experiment indicate thatneutral-tone syllables do have a target that is independent of the surroundingtones, which is likely to be static and mid. Furthermore, the neutral tone is found tobe different from the full lexical tones in the manner with which the underlyingtonal target is implemented: it is slow and ineffective both in overcoming the influenceof the preceding full lexical tone and in approaching its own target. Applyingthe recently proposed pitch target approximation model, we conclude that theneutral tone differs from the other lexical tones in Standard Chinese not only interms of its mid target, but also in terms of the weak articulatory strength withwhich this target is implemented. Finally, we suggest that this new understandingis potentially applicable to other weak elements in speech.

Remapping Auditory-Motor Representations in Voice Production

Evidence regarding visually guided limb movements suggests that the motor system learns and maintains neural maps between motor commands and sensory feedback. Such systems are hypothesized to be used in a feed-forward control strategy that permits precision and stability without the delays of direct feedback control. Human vocalizations involve precise control over vocal and respiratory muscles. However, little is known about the sensorimotor representations underlying speech production. Here, we manipulated the heard fundamental frequency of the voice during speech to demonstrate learning of auditory-motor maps. Mandarin speakers repeatedly produced words with specific pitch patterns (tone categories). On each successive utterance, the frequency of their auditory feedback was increased by 1/100 of a semitone until they heard their feedback one full semitone above their true pitch. Subjects automatically compensated for these changes by lowering their vocal pitch. When feedback was unexpectedly returned to normal, speakers significantly increased the pitch of their productions beyond their initial baseline frequency. This adaptation was found to generalize to the production of another tone category. However, results indicate that a more robust adaptation was produced for the tone that was spoken during feedback alteration. The immediate aftereffects suggest a global remapping of the auditory-motor relationship after an extremely brief training period. However, this learning does not represent a complete transformation of the mapping; rather, it is in part target dependent.

Speech melody as articulatorily implemented communicative functions

The understanding of speech melody, i.e., pitch variations related to tone and intonation, can be improved by simultaneously taking into consideration two basic facts: that speech conveys communicative meanings, and that it is produced by human articulators. Communicative meanings, as I will argue, are conveyed through a set of separate functions that are realized by an articulatory system with various biophysical properties. These properties make it unlikely that the melodic functions are encoded directly in terms of invariant surface acoustic forms. Rather, the encoding is likely done through the manipulation of a limited number of articulatorily operable parameters that may be considered as the phonetic primitives. Four such primitives can be recognized for speech melody: local pitch targets, pitch range, articulatory strength and duration. The values of the melodic primitives are specified by a set of encoding schemes, each associated with a particular communicative function. The encoding schemes are distinct from each other in the manner of controlling the melodic primitives, which allows multiple communicative functions to be conveyed in parallel. The communicative functions are ultimately converted to continuous, detailed surface acoustic patterns through an articulatory process of syllable-synchronized sequential target approximation, which takes the melodic primitives specified by the encoding schemes as the control parameters. This view of speech melody is summarized into a comprehensive model of tone and intonation, namely, the parallel encoding and target approximation (PENTA) model.

Synthesis of prosody using multi-level unit sequences

Generating meaningful and natural sounding prosody is a central challenge in text-to-speech synthesis (TTS). In traditional synthesis, the challenge consists of how to generate natural target prosodic contours and how to impose these contours on recorded speech without causing audible distortions. In unit selection synthesis, the challenge is the sheer size of the speech corpus that is needed to cover all combinations of phone sequences and prosodic contexts that can occur in a given language. This paper describes new methods that are being explored, based on the principle of superpositional prosody transplant. Both methods are based on the following procedure. In a recorded, prosodically and phonemically labeled corpus, the log pitch contours are additively decomposed into component curves according to a prosodic hierarchy, typically phrase curves (corresponding to phrases), accent curves (corresponding to feet), and segmental perturbation (or residuals) curves. During synthesis, the corpus is searched for multiple unit sequences: A unit sequence that covers the target phoneme string, and one or more unit sequences that cover the prosodic labels at a given phonological level (e.g., the foot or phrase) and are constrained by being matched to the phone match sequence in terms of the phonetic classes of the phonemes (or in terms of higher level entities, such as the number of feet and their sizes measured in syllables). The methods differ in terms of the level of detail of these constraints. A superpositional prosody transplant procedure generates a target pitch contour by extracting and recombining component curves from these sequences, and imposing this contour on the sequence that matches the phone string using standard speech modification methods. This process minimizes prosodic modification artifacts, optimizes the naturalness of the target pitch contour, yet avoids the combinatorial explosion of standard unit selection synthesis.

Phonetic realization of focus in English declarative intonation

The present study investigates how focus is phonetically realized in declarative sentences in American English. The goal is to test the hypotheses that, (a) focus is manifested in parallel rather than in alternation with other intonational functions, and (b) every syllable in a sentence is associated with a local pitch target. Eight native speakers of American English recorded short declarative sentences with narrow focus at different locations or without any narrow focus. Detailed f 0 analyses reveal that a narrow focus is realized by expanding the pitch range of the on-focus stressed syllables, suppressing the pitch range of postfocus syllables, and leaving the pitch range of prefocus syllables largely intact. Focus is not found, however, to determine the presence or absence of f 0 peaks. Data analyses also reveal evidence for the presence of a local pitch target in every syllable. These findings are incompatible with conventional theories of English intonation. As an alternative, the Parallel Encoding and Target Approximation (PENTA) model is considered. The model defines and organizes the intonational components in terms of function rather than form. It also assumes target approximation rather than interpolation as the basic articulatory mechanism of f 0 contour generation. It is argued that the approach used in the PENTA model, which takes account of both communicative functions and articulatory implementation, may provide a coherent account of detailed f 0 contours in English.

Neural-network simulation of tonal categorization based on F0 velocity profiles

Perception studies have shown that by the age of six months, infants show particular response patterns to tones in their native language. The present study focuses on how infants might develop lexical tones in Man- darin. F0 is generally considered the main cue in tone perception. However, F0 patterns in connected speech display extensive contextual variability. Since speech input to infants consists mainly of multi-word utterances, tone learning must involve processes that can effectively resolve variability. In this study we explore the Target Approximation model (Xu and Wang, 2001) which characterizes surface F0 as asymptotic movements toward underlying pitch targets defined as simple linear functions. The model predicts that it is possible to infer underlying pitch targets from the manners of F0 movements. Using production data of three of the speakers from Xu (1997), we trained a self-organizing neural network with both F0 profiles and F0 velocity profiles as input. In the testing phase, velocity profiles yielded far superior categorization than F0 profiles. The results confirm that velocity profiles can effectively abstract away from surface variability and directly reflect underlying articulatory goals. The finding thus points to one way through which infants can successfully derive at phonetic categories from adult speech.

Modeling Improved Prosody Generation from High-Level Linguistically Annotated Corpora

Synthetic speech usually suffers from bad F0 contour surface. The prediction of the underlying pitch targets robustly relies on the quality of the predicted prosodic structures, i.e. the corresponding sequences of tones and breaks. In the present work, we have utilized a linguistically enriched annotated corpus to build data-driven models for predicting prosodic structures with increased accuracy. We have then used a linear regression approach for the F0 modeling. An appropriate XML annotation scheme has been introduced to encode syntax, grammar, new or already given information, phrase subject/object information, as well as rhetorical elements in the corpus, by exploiting a Natural Language Generator (NLG) system. To prove the benefits from the introduction of the enriched input meta-information, we first show that while tone and break CART predictors have high accuracy when standing alone (92.35% for breaks, 87.76% for accents and 99.03% for endtones), their application in the TtS chain degrades the Linear Regression pitch target model. On the other hand, the enriched linguistic meta-information minimizes errors of models leading to a more natural F0 surface. Both objective and subjective evaluation were adopted for the intonation contours by taking into account the propagated errors introduced by each model in the synthesis chain.

Age-related differences in auditory event-related potentials during a cued attention task

Objective: To determine if aging is associated with differences in attentional regulation using behavioral and event-related potential (ERP) measures. Methods: Younger ( n=13; M=20 years) and older ( n=12; M=76 years) subjects performed an auditory cued attention task. Verbal cues correctly (valid) or incorrectly (invalid) predicted the ear receiving a target tone 1.5 s later, or were uninformative (neutral). Targets were either ‘high’ (2000 Hz) or ‘low’ (1000 Hz) pitch monaural tones. Subjects pressed one of 4 buttons to indicate target ear and pitch. ERPs following cues and targets (P50, N100, P200, slow waves), and negative slow potentials (CNV) between cues and targets were assessed. Results: Cue information had significant effects on reaction time for both groups (valid<neutral<invalid). Target N100 amplitude was significantly affected by cueing in younger (invalid>valid) but not older subjects. Target slow waves were also affected by cue information (invalid>valid), and the difference was larger and lasted longer in older subjects. Slow waves following cues were significantly larger in older subjects, but the subsequent CNV amplitudes were comparable among groups. Conclusions: When performing a cued attention task, age differences are present in transient ERPs following cues and targets. Significance: Age differences in ERPs associated with attentional regulation support the hypothesis that attentional changes contribute to cognitive aging.

Multiple effects of consonant manner of articulation and intonation type on F0 in English

In this study we examine how consonant manner of articulation interacts with intonation type in shaping the F0 contours in English. Native speakers of American English read aloud words differing in vowel length, consonant manner of articulation and consonant position in word. They produced each word in either a statement or question carrier. F0 contours of their speech were extracted by measuring every complete vocal period. Preliminary results based on graphic analysis of three speakers’ data suggest that there are three distinct consonantal effects: F0 interruption due to devoicing, a large but brief (10–40 ms) F0 raising at the onset of voicing, and a smaller but longer-lasting F0 raising throughout a large proportion of the preceding and following vowels. These effects appear to be imposed on a continuously changing F0 curve that is either rising-falling or falling-rising, depending on whether the carrier sentence is a statement or a question. Further analysis will test the hypothesis that these continuous curves result from local pitch targets that are assigned to individual syllables and implemented with them in synchrony regardless of their segmental composition. [Work supported by NIDCD Grant No. R01 DC03902.]

Tonal structure and pitch targets in Italian focus constituents

Aquest article investiga l’estructura tonal de l’accent focal de les oracions declaratives amb focus estret en italià de Nàpols. Les propietats formals de l’accent focal es podrien en principi analitzar de dues maneres possibles, bé com un accent tonal HL de tipus descendent o bé com un accent tonal LH de tipus ascendent. Les dues possibles anàlisis són avaluades en funció del comportament de l’accent tonal en constituents que contenen un nombre variable de mots. Els constituents focalitzats presenten una vall d’F0 que es comporta com un clar objectiu entonatiu. Aquest punt d’inflexió podria ser part també d’un accent HL o LH. Indicis tonals i temporals donen suport a la hipòtesi que es tracta d’un accent tonal ascendent. D’aquesta anàlisi se’n deriva el fet que el moviment tonal descendent final de les oracions declaratives s’ha d’analitzar com un gest tonal diferent de l’accent nuclear i que aquest es comporta de forma anàloga al to de frontera ascendent final de les oracions interrogatives.