Vowel Normalization Research Articles

The comparative study of resonance disorders for Vietnamese and Korean cleft palate speakers using nasometer

BackgroundNasalance is used to evaluate the velopharyngeal incompetence in clinical diagnoses using a nasometer. The aim of this study is to find the nasalance differences between Vietnamese cleft palate children and Korean cleft palate children by measuring the nasalance of five oral vowels.MethodsTen Vietnamese cleft palate children after surgery, three Vietnamese children for the control group, and ten Korean cleft palate children after surgery with the same age participated in this experimentation. Instead of Korean control, the standard value of Korean version of the simplified nasometric assessment procedures (kSNAP) was used.ResultThe results are as follows: (1) the highest nasalance score among the Vietnamese normal vowels is the low vowel /a/; however, that of Korean normal vowels is the high vowel /i/. (2) The average nasalance score of Korean cleft palate vowels is 18% higher than that of Vietnamese cleft palate vowels. There was a nasalance score of over 45% among the vowels /e/ and /i/ in Vietnamese cleft palate patients and /i/, /o/, and /u/ in Korean cleft palate patients.ConclusionThese different nasalance scores of the same vowels seem to cause an ethnic difference between Vietnamese and Korean cleft palate children.

A double-Nearey theory of vowel normalization: Approaching consensus

Terry Nearey has provided substantial data and insights into the realm of speech perception throughout his career. A major concern of his is to reconcile seemingly contradictory results. Here, his double-weak theory of perception and his probabilistic sliding template model (PSTM) are examined in relation to vowel normalization. Various algorithms achieve good results, but they must all confront the contradiction in behavior: We normalize the speech of others but we can also tell when they use a (slightly) different vowel than we do. Thus, listeners are (as in double-weak theory) fairly accommodating of different vocal tracts, but, contrarily, they can hear differences as well. PSTM may account for this difference, if deviations from theory-predicted vowel location are just those that speakers identity as being different; this remains to be tested. The various sources of vowel information (lower formants, F0, higher formants, spectral tilt) interact in complex ways, as shown in Nearey’s work and others’. No theory is comprehensive yet, as acknowledged by Nearey himself. He has also made the observation that our theories of speech perception ought to perform at least as well as automatic speech recognition systems. Although we have not reached that stage, Nearey’s work brings us closer.

What the f***: An acoustic-pragmatic analysis of meaning in The Wire

The award-winning television drama, The Wire, contains a famous scene in which Detectives Moreland and McNulty discuss the crime scene they are investigating using only variations of a single expletive—fuck. Despite limited vocabulary, viewers are able to extract meaning and interpret the scene. This study considers all of the expletives produced in the scene, and carries out an acoustic analysis of / ʌ /—F1~F3 (midpoint and averages) and vowel duration. These measurements are then used in combination with Gricean (Grice 1975) and neo-Gricean (e.g. Horn 1984) pragmatic analysis in an attempt to categorize the intended meaning of fuck as: intensification, confusion, dissatisfaction or suspicion (Fairman 2009). Vowel measurements are considered for Moreland and McNulty individually, and vowel normalization is carried out in order to determine pragmatic categorizations across both detectives. This paper argues that acoustic-phonetic analysis can augment Gricean and neo-Gricean pragmatic analysis by providing a means of determining whether what appears at a lexical level to be conversationally implicated meaning is in fact conventional implicature conveyed at the phonetic level.

Suitability of speaker normalization procedures for classifying vowels produced by speakers with dysarthria

Speaker normalization, a process whereby the perceptual system of a listener recalibrates to accommodate individual speakers, is proposed to account for the ease with which we understand speech produced by multiple speakers with different sized and shaped vocal tracts. A variety of vowel-, formant-, and speaker-intrinsic or extrinsic transforms have been proposed to model speaker normalization of vowels produced by multiple speakers (e.g., Mel, Bark, and Lobanov methods). Suitability of such normalization procedures has been examined extensively in non-disordered speaker populations. Unknown at this point, however, is the appropriateness of normalization procedures for transforming spectrally distorted vowels produced by speakers with dysarthria. Thus, we examined the suitability of two transforms, Bark and Lobanov, for normalizing vowels produced by a heterogeneous cohort of 45 speakers with dysarthria. Non-normalized (Hertz) and Bark transformed vowel tokens were classified via discriminant function analysis (DFA) with 55% and 56% accuracy, respectively. Classification accuracy of vowel tokens normalized using Lobanov’s method was 65%. The results of the DFAs were compared to perceptual data, which revealed listeners identified vowel tokens with 71% accuracy. These results suggest vowel-extrinsic and formant- and speaker-intrinsic normalization methods (e.g., Lobanov) are better suited to model speaker normalization of dysarthric vowels.

A general auditory bias for handling speaker variability in speech? Evidence in humans and songbirds.

Different speakers produce the same speech sound differently, yet listeners are still able to reliably identify the speech sound. How listeners can adjust their perception to compensate for speaker differences in speech, and whether these compensatory processes are unique only to humans, is still not fully understood. In this study we compare the ability of humans and zebra finches to categorize vowels despite speaker variation in speech in order to test the hypothesis that accommodating speaker and gender differences in isolated vowels can be achieved without prior experience with speaker-related variability. Using a behavioral Go/No-go task and identical stimuli, we compared Australian English adults’ (naïve to Dutch) and zebra finches’ (naïve to human speech) ability to categorize / I/ and /ε/ vowels of an novel Dutch speaker after learning to discriminate those vowels from only one other speaker. Experiments 1 and 2 presented vowels of two speakers interspersed or blocked, respectively. Results demonstrate that categorization of vowels is possible without prior exposure to speaker-related variability in speech for zebra finches, and in non-native vowel categories for humans. Therefore, this study is the first to provide evidence for what might be a species-shared auditory bias that may supersede speaker-related information during vowel categorization. It additionally provides behavioral evidence contradicting a prior hypothesis that accommodation of speaker differences is achieved via the use of formant ratios. Therefore, investigations of alternative accounts of vowel normalization that incorporate the possibility of an auditory bias for disregarding inter-speaker variability are warranted.

Non-distorting method of vowel normalization by transposition in the semitonal scale

Converting formant values from Hertz into semitones provides a remarkable “area agreement” between vowel systems in the F1/F2 plane without distorting intervals between the formants. Further simple transposition of a Sample system in relation to a Reference system by the difference between the mean of all converted F1s and F2s for the Sample speaker/group and the mean of those for the Reference speaker/group efficiently eliminates physiological mismatch between the speakers/groups and allows for convenient comparison of the vowel systems in a meaningful scale of pitch intervals. To convert formant n of vowel V from Hertz to semitones (expressed in MIDI keyboard numbers) use formula: [1] Fn[V]midi = midi(Fn[V]Hz), where [2] midi(x) = 12*log2(x/440) + 69. To normalize converted formants of a Sample speaker/group against those of a Reference speaker/group, deduct from the converted Sample formants (Fn[V]midi (Sample)) the difference between the mean of all converted F1s and F2s for the Sample speaker/group a...

A virtual environment for modeling the acquisition of vowel normalization

Vowel normalization is a computation that is meant to account for the differences in the absolute direct (physical or psychophysical) representations of qualitatively equivalent vowel productions that arise due to differences in speaker properties such as body size types, age, gender, and other socially interpreted categories that are based on natural variation in vocal tract size and shape. We present a virtual environment for vocal learning which provides the means to model the acquisition of vowel normalization, along with other aspects of vocal learning. The environment consists of models of caretaker agents representing five different language communities—American English, Cantonese, Greek, Japanese, and Korean—derived from vowel category perception experiments (Munson et al., 2010, Plummer et al., 2013) and models of infant agents (Plummer, 2012, 2013) that “vocally interact” with their caretakers. Moreover, we develop a model of caretaker social and vocal signaling in response to infant vowel productions, and of an infant's internalization of these signals and the internal computations over them. More broadly, we model the acquisition of vowel normalization within a developmental framework encompassing a suite of vocal learning phenomena, including language-specific caretaker vocal exchanges, perceptual warping, and multisensory matching and narrowing.

Analysis and Recognition of Vowels in SHAI`YÂNG MIRI Language using Formants

Speech is generated by the co-ordination of various anatomical articulations known collectively as the vocal organs. The waveforms thus produced are either voiced or unvoiced speech. All the vowel sounds are voiced and have the sound source, the glottis. An amplification caused by the vocal tract filter is called resonance, and in speech these resonance are known as formants. Since each vowel has a different vocal tract shape, it will have a different format pattern. This paper is an attempt to analyze the vowels in SHAI`YÂNG MIRI language using the formant analysis. Analysis of the eight cardinal vowel speech samples in SHAI`YÂNG MIRI language is performed using formants, spectrogram, 19-channel auditory filter bank and through normalization of vowels. General Terms Speech Processing and Analysis, Speech Synthesis

Normalization and Automatic Recognition of Assamese Vowels

Vowel normalization is a procedure of finding out the intermediate parameter from varied forms of utterance produced by different speakers of a particular language or same vowel token uttered by same speaker at different times. Every speaker has their own speech parameters and the variation of these parameters is the primary cause of speaker to speaker variation of speech sounds. The significant parameters that distinguish one vowel from the other are its formant values. In this paper an automatic vowel recognition procedure is described where the eight standard Assamese vowels are analyzed through Formant detection and Normalization. Here a standard approach is adopted for Normalization of vowels based on two Formants F1 and F2. A Vowel Recognition System based on Formant Estimation is developed and a preliminary recognition test for isolated vowels shows an expected recognition rate of 84-96% for arbitrarily selected speakers. General Terms Speech processing, Phonetics

Evidence for precategorical extrinsic vowel normalization

Three experiments investigated whether extrinsic vowel normalization takes place largely at a categorical or a precategorical level of processing. Traditional vowel normalization effects in categorization were replicated in Experiment 1: Vowels taken from an [I]-[ε] continuum were more often interpreted as /I/ (which has a low first formant, F(1)) when the vowels were heard in contexts that had a raised F(1) than when the contexts had a lowered F(1). This was established with contexts that consisted of only two syllables. These short contexts were necessary for Experiment 2, a discrimination task that encouraged listeners to focus on the perceptual properties of vowels at a precategorical level. Vowel normalization was again found: Ambiguous vowels were more easily discriminated from an endpoint [ε] than from an endpoint [I] in a high-F(1) context, whereas the opposite was true in a low-F(1) context. Experiment 3 measured discriminability between pairs of steps along the [I]-[ε] continuum. Contextual influences were again found, but without discrimination peaks, contrary to what was predicted from the same participants' categorization behavior. Extrinsic vowel normalization therefore appears to be a process that takes place at least in part at a precategorical processing level.

Vowel normalization and the perception of speaker changes: An exploration of the contextual tuning hypothesis

Many experiments have reported a perceptual advantage for vowels presented in blocked-versus mixed-voice conditions. Nusbaum and colleagues [Nusbaum and Morin (1992). in Speech Perception, Speech Production, and Linguistic Structure, edited by Y. Tohkura, Y. Sagisaka, and E. Vatikiotis-Bateson (OHM, Tokyo), pp. 113-134; Magnuson and Nusbaum (2007). J. Exp. Psychol. Hum. Percept. Perform. 33(2), 391-409] present results which suggest that the size of this advantage may be related to the facility with which listeners can detect speaker changes, so that combinations of less similar voices can result in better performance than combinations of more similar voices. To test this, a series of synthetic voices (differing in their source characteristics and/or formant-spaces) was used in a speeded-monitoring task. Vowels were presented in blocks made up of tokens from one or two synthetic voices. Results indicate that formant-space differences, in the absence of source differences between voices in a block, were unlikely to result in the perception of multiple voices, leading to lower accuracy and relatively faster reaction times. Source differences between voices in a block resulted in the perception of multiple voices, increased reaction times, and a decreased negative effect of formant-space differences between voices on identification accuracy. These results are consistent with a process in which the detection of speaker changes guides the appropriate or inappropriate use of extrinsic information in normalization.

Studies on inter-speaker variability in speech and its application in automatic speech recognition

In this paper, we give an overview of the problem of inter-speaker variability and its study in many diverse areas of speech signal processing. We first give an overview of vowel-normalization studies that minimize variations in the acoustic representation of vowel realizations by different speakers. We then describe the universal-warping approach to speaker normalization which unifies many of the vowel normalization approaches and also shows the relation between speech production, perception and auditory processing. We then address the problem of inter-speaker variability in automatic speech recognition (ASR) and describe techniques that are used to reduce these effects and thereby improve the performance of speaker-independent ASR systems.

Constraints on the processes responsible for the extrinsic normalization of vowels.

Listeners tune in to talkers’ vowels through extrinsic normalization. We asked here whether this process could be based on compensation for the long-term average spectrum (LTAS) of preceding sounds and whether the mechanisms responsible for normalization are indifferent to the nature of those sounds. If so, normalization should apply to nonspeech stimuli. Previous findings were replicated with first-formant (F1) manipulations of speech. Targets on a [pt]–[pɛt] (low–high F1) continuum were labeled as [pt] more after high-F1 than after low-F1 precursors. Spectrally rotated nonspeech versions of these materials produced similar normalization. None occurred, however, with nonspeech stimuli that were less speechlike, even though precursor–target LTAS relations were equivalent to those used earlier. Additional experiments investigated the roles of pitch movement, amplitude variation, formant location, and the stimuli's perceived similarity to speech. It appears that normalization is not restricted to speech but that the nature of the preceding sounds does matter. Extrinsic normalization of vowels is due, at least in part, to an auditory process that may require familiarity with the spectrotemporal characteristics of speech.

Probing the independence of formant control using altered auditory feedback.

Two auditory feedback perturbation experiments were conducted to examine the nature of control of the first two formants in vowels. In the first experiment, talkers heard their auditory feedback with either F1 or F2 shifted in frequency. Talkers altered production of the perturbed formant by changing its frequency in the opposite direction to the perturbation but did not produce a correlated alteration of the unperturbed formant. Thus, the motor control system is capable of fine-grained independent control of F1 and F2. In the second experiment, a large meta-analysis was conducted on data from talkers who received feedback where both F1 and F2 had been perturbed. A moderate correlation was found between individual compensations in F1 and F2 suggesting that the control of F1 and F2 is processed in a common manner at some level. While a wide range of individual compensation magnitudes were observed, no significant correlations were found between individuals' compensations and vowel space differences. Similarly, no significant correlations were found between individuals' compensations and variability in normal vowel production. Further, when receiving normal auditory feedback, most of the population exhibited no significant correlation between the natural variation in production of F1 and F2.

The influence of language background on the relative perception of vowels.

Individual differences in vowel production result in an extensive overlap of vowel categories in F1 × F2 vowel space, especially in languages with large vowel inventories. When perceiving vowel categories, listeners compensate for this variability by identifying vowels relative to the voice characteristics of individual speakers. The present study examines the effect of language background on the relative perception of vowels, i.e., whether listeners engage in vowel normalization differently in languages with vowel inventories of varied sizes and with different amounts of category overlap due to across‐speaker variability. To that end, we compare English and Dutch, both with 10+ vowels and a substantial category overlap, and Spanish, with only 5 vowels and less category overlap. Vowel targets from an /o/–/u/ continuum were embedded in sentences spoken by native speakers of all three languages. F1 of the precursor sentences was modified to be either high or low. Seventy‐two listeners participated in the vo...

Using systematic synthetic voice differences to probe the perception of mixed‐talker vowels.

Many experiments have reported a perceptual advantage for vowels presented in blocked versus mixed speaker conditions. Nusbaum and colleagues [Nusbaum and Morin (1992); Magnuson and Nusbaum (2007)] found that the size of this advantage varies with aspects of the similarity of the voices involved. The largest performance difference occurs when voices are of intermediate similarity (e.g, two similar sounding female voices with incongruent vowel spaces). Difference are smaller when voices are either very similar or very different. The current study will explore some potentially relevant dimensions of similarity using vowels from a set of synthetic virtual talkers whose voice characteristics differ in formant ranges, f0, and/or source characteristics. Perceptual accuracy and response times will be collected using a speeded monitoring task in which participants are asked to respond as soon as they hear a specific target vowel. Results will be evaluated in light of current theories of vowel normalization and speaker adaptation.

A Comparison of Vowel Normalization Methods and Their Applications

In this paper I discuss the necessity and possibility of the application of normalization methods in phonetic studies, especially in the acoustic analysis of vowels. I also compare and evaluate the vowel normalization methods listed in Adank et al. (2004), with the method of calculating the centrality of vowels after normalization. The results show that the Z-score method and CLIH1 method are the most effective in centralizing the data and eliminating gender differences.

A Comparison of Vowel Normalization Methods and Their Applications

In this paper I discuss the necessity and possibility of the application of normalization methods in phonetic studies, especially in the acoustic analysis of vowels. I also compare and evaluate the vowel normalization methods listed in Adank et al. (2004), with the method of calculating the centrality of vowels after normalization. The results show that the Z-score method and CLIH1 method are the most effective in centralizing the data and eliminating gender differences.

Teaching & Learning Guide for: The Study of Variation from Two Perspectives

Teaching & Learning Guide for: The Study of Variation from Two Perspectives

At which processing level does extrinsic speaker information influence vowel perception?

The interpretation of vowel sounds depends on perceived characteristics of the speaker [e.g., average first formant (F1) frequency]. A vowel between /ɪ/ and /ε/ is more likely to be perceived as /ɪ/ if a precursor sentence indicates that the speaker has a relatively high average F1. Behavioral and electrophysiological experiments investigating the locus of this extrinsic vowel normalization are reported. The normalization effect with a categorization task was first replicated. More vowels on an /ɪ/-/ε/ continuum followed by a /papu/ context were categorized as /ɪ/ with a high-F1 context than with a low-F1 context. Two experiments then examined this context effect in a 4I-oddity discrimination task. Ambiguous vowels were more difficult to distinguish from the /ɪ/-endpoint if the context /papu/ had a high F1 than if it had a low F1 (and vice versa for discrimination of ambiguous vowels from the /ε/-endpoint). Furthermore, between-category discriminations were no easier than within-category discriminations. Together, these results suggest that the normalization mechanism operates largely at an auditory processing level. The mismatch negativity (an automatically evoked brain potential) arising from the same stimuli is being measured, to investigate whether extrinsic normalization takes place in the absence of an explicit decision task.