Melodic Context Research Articles

Фторы и звукорядные модуляции в средневизантийском распеве (на материале классической редакции Стихираря XIII – XIV вв.)

The Middle Byzantine diastematic notation knows a number of particular signs – the phthorai – indicating modal changes. Being added to the interval neumes, they convey information that could not be seen from the neumes themselves, namely, about scale structure of Middle Byzantine hymns.Phthorai of archaic type are placed on the intonation of ascending forth g-c in a melodic element of the mode “nana”. Latest phthorai, which had been added to Middle Byzantine singing manuscripts in post-Byzantine period, are of different types. The paper deals with one of them, the phthora of the tritos mode (c). It is found in chants of all eight modes, in different melodic context, placed on c and one fifth downwards, on f. It indicates small interval, i.e. one half-step below the neume on which the phthora occurs. However, in few other cases phthorai “c” are found on another pitch. From these transposed phthorai we can learn about interval structure of melodic formulae and, more specifically, about scale modulations.The paper presents an analysis of two stichera: “To tritto tes eroteseos” to Saint Peter, in tetartos mode, and “Deute makarisomen apantes Iosef” for Holy Saturday, in plagios deuteros mode. The author draws conclusion that their modal structure contains a scale modulation. Some melodic configurations or more large melodic formulae should be transcribed to the modern notation system with accidentals, in these cases, f-sharp. Modulations may serve as an artistic mean to stress the structure of the hymn and to convey its author’s line of thought.

Final-note expectancy and humor: an empirical investigation

BackgroundMelodic expectations were manipulated to investigate the nature of tonally incongruent melodic final notes that may elicit humor in listeners. To our knowledge, this is the first experiment aiming at studying humor elicitation in music with the use of empirical, quantitative methods. To this aim, we have based the experiment on the incongruency/resolution theory of humor and the violations of expectations in music. Our goal was to determine the amount of change, that is, the degree of incongruency required to elicit humor.MethodsWe composed two simple, 8-bar long melodies, and changed their final notes so that they could randomly finish on any semitone between an octave upwards and downwards with respect to the original, tonic final note. This resulted in 25 versions for both melodies, including the original final notes, for each semitone. Musician and non-musician participants rated each version of each melody on five 7-point bipolar scales according to goodness of fit, humor, beauty, playfulness, and pleasantness.Results and conclusionsOur results showed that even a single change of the final note can elicit humor. No strong connection was found between humor elicitation and the level of incongruency (i.e., the amount of violation of expectation). Instead, changes to the major-mode melody were more likely to be found humorous than those to the minor-mode melody, implying that a so-called playful context is necessary for humor elicitation as the major melody was labelled playful by the listeners. Furthermore, final notes below the original tonic end note were also found to be less humorous and less fitting to the melodic context than those above it.

Musicianship and melodic predictability enhance neural gain in auditory cortex during pitch deviance detection.

When listening to music, pitch deviations are more salient and elicit stronger prediction error responses when the melodic context is predictable and when the listener is a musician. Yet, the neuronal dynamics and changes in connectivity underlying such effects remain unclear. Here, we employed dynamic causal modeling (DCM) to investigate whether the magnetic mismatch negativity response (MMNm)—and its modulation by context predictability and musical expertise—are associated with enhanced neural gain of auditory areas, as a plausible mechanism for encoding precision‐weighted prediction errors. Using Bayesian model comparison, we asked whether models with intrinsic connections within primary auditory cortex (A1) and superior temporal gyrus (STG)—typically related to gain control—or extrinsic connections between A1 and STG—typically related to propagation of prediction and error signals—better explained magnetoencephalography responses. We found that, compared to regular sounds, out‐of‐tune pitch deviations were associated with lower intrinsic (inhibitory) connectivity in A1 and STG, and lower backward (inhibitory) connectivity from STG to A1, consistent with disinhibition and enhanced neural gain in these auditory areas. More predictable melodies were associated with disinhibition in right A1, while musicianship was associated with disinhibition in left A1 and reduced connectivity from STG to left A1. These results indicate that musicianship and melodic predictability, as well as pitch deviations themselves, enhance neural gain in auditory cortex during deviance detection. Our findings are consistent with predictive processing theories suggesting that precise and informative error signals are selected by the brain for subsequent hierarchical processing.

Change detection and schematic processing in music

Research into vision has highlighted the importance of gist representations in change detection and memory. This article puts forth the hypothesis that schematic processing and gist provide an account for change detection in music as well, where a musical gist is an abstracted memory representation for schematically consistent tones. The present experiments illuminate the content of gist memory representations by testing when listeners can detect single-tone changes in pairs of melodies. In Experiment 1, musicians and non-musicians listened to melodies varying in tonal structure. Less structure resulted in compromised change detection in both groups. Most often, musicians displayed more accurate change detection than non-musicians, but, surprisingly, when schematic processing could not contribute to memory encoding, musicians performed worse than their untrained counterparts. Experiment 2 utilized a full-factorial design to examine tonality, interval of pitch change, metrical position, and rhythm. Tonality had a particularly large effect on performance, with non-scale tones generally aiding change detection. Listeners were unlikely, however, to detect schematically-inconsistent tones when only brief melodic context was available. The results uphold the hypothesis that memory for melodies relies on schematic processing, with change detection dependent upon whether the change alters the schematic gist of the melody.

“I know it when I hear it”

Listeners regularly judge the accuracy of musical performances. However, as is true for several types of judgments (e.g., beauty or obscenity), estimating the correctness of melodies is not based on a precise definition of the object/performance but rather follows arguments such as “I know it when I hear it”. In order to clarify the definition of correctness in melodies, participants identified parametrically manipulated sung melodies as in-tune or out-of-tune, using the method of limits procedure (Experiment 1). Listeners’ tolerance with regard to mistuning was compared across melodies (Experiment 2). The potential roots of correctness perception were investigated by testing the effect of familiarity, the influence of formal musical training (Experiment 3), and the task repetition effect (Experiment 4). The results highlight a surprisingly small tolerance with regard to mistuning (half of a quarter tone), whatever the melodic context, large individual differences, but high consistency over time. This high sensitivity was mainly modulated by musical training as well as by previous exposure. In addition to defining the boundary between in- and out-of-tune melodies, this study supports the implicit development of the normative notion of “correctness” as a category that might drive listeners’ appreciation of artistic performances.

Familiar Tonal Context Improves Accuracy of Pitch Interval Perception.

A fundamental feature of everyday music perception is sensitivity to familiar tonal structures such as musical keys. Many studies have suggested that a tonal context can enhance the perception and representation of pitch. Most of these studies have measured response time, which may reflect expectancy as opposed to perceptual accuracy. We instead used a performance-based measure, comparing participants’ ability to discriminate between a “small, in-tune” interval and a “large, mistuned” interval in conditions that involved familiar tonal relations (diatonic, or major, scale notes), unfamiliar tonal relations (whole-tone or mistuned-diatonic scale notes), repetition of a single pitch, or no tonal context. The context was established with a brief sequence of tones in Experiment 1 (melodic context), and a cadence-like two-chord progression in Experiment 2 (harmonic context). In both experiments, performance significantly differed across the context conditions, with a diatonic context providing a significant advantage over no context; however, no correlation with years of musical training was observed. The diatonic tonal context also provided an advantage over the whole-tone scale context condition in Experiment 1 (melodic context), and over the mistuned scale or repetition context conditions in Experiment 2 (harmonic context). However, the relatively small benefit to performance suggests that the main advantage of tonal context may be priming of expected stimuli, rather than enhanced accuracy of pitch interval representation.

Intonation detection in a melodic context

Listeners with a wide range of formal and informal musical experience were asked to listen to an eight-tone diatonic C Major scale, generated using a piano sample library, in which one of four notes (D4, F4, A4, or C5) would be mistuned in 13 different mounts between -32¢ and + 32¢. Listeners were told which note might be mistuned and were simply asked to indicate whether the scale was in-tune or not. Each listener was exposed to each degree of mistuning ten times. The frequency with which they said a scale was in-tune as a function of the degree of mistuning was plotted for each note and listener, to which a three-parameter pseudo-normal distribution (mean, standard deviation, height) was fitted. The standard deviation indicated the sensitivity of the listener to intonation in each case (large deviation implied low sensitivity to intonation). Listeners were then ranked based on their musical background, training, and experience. The effect of musical training on intonation sensitivity was a significant factor (p<0.001). There was also a significant effect of the particular note on the sensitivity of listeners, with the intonation of A4 being most difficult to detect across listeners (p<0.001).

Memory for musical tones: the impact of tonality and the creation of false memories.

Although the relation between tonality and musical memory has been fairly well-studied, less is known regarding the contribution of tonal-schematic expectancies to this relation. Three experiments investigated the influence of tonal expectancies on memory for single tones in a tonal melodic context. In the first experiment, listener responses indicated superior recognition of both expected and unexpected targets in a major tonal context than for moderately expected targets. Importantly, and in support of previous work on false memories, listener responses also revealed a higher false alarm rate for expected than unexpected targets. These results indicate roles for tonal schematic congruency as well as distinctiveness in memory for melodic tones. The second experiment utilized minor melodies, which weakened tonal expectancies since the minor tonality can be represented in three forms simultaneously. Finally, tonal expectancies were abolished entirely in the third experiment through the use of atonal melodies. Accordingly, the expectancy-based results observed in the first experiment were disrupted in the second experiment, and disappeared in the third experiment. These results are discussed in light of schema theory, musical expectancy, and classic memory work on the availability and distinctiveness heuristics.

Intonation Analysis of Rāgas in Carnatic Music

Intonation is a fundamental music concept that has a special relevance in Indian art music. It is characteristic of a rga and key to the musical expression of the artist. Describing intonation is of importance to several music information retrieval tasks such as developing similarity measures based on rgas and artists. In this paper, we first assess rga intonation qualitatively by analysing varṇaṁs, a particular form of Carnatic music compositions. We then approach the task of automatically obtaining a compact representation of the intonation of a recording from its pitch track. We propose two approaches based on the parametrization of pitch-value distributions: performance pitch histograms, and context-based svara distributions obtained by categorizing pitch contours based on the melodic context. We evaluate both approaches on a large Carnatic music collection and discuss their merits and limitations. We finally go through different kinds of contextual information that can be obtained to further improve the two approaches.

Pitch discrimination and melodic memory in children with autism spectrum disorders

Pitch perception is enhanced among persons with autism. We extended this finding to memory for pitch and melody among school-aged children. The purpose of this study was to investigate pitch memory in musically untrained children with autism spectrum disorders, aged 7-13 years, and to compare it to that of age- and IQ-matched typically developing children. The children were required to discriminate isolated tones in two differing contexts as well to remember melodies after a period of 1 week. The tasks were designed to employ both short- and long-term memory for music. For the pitch discrimination task, the children first had to indicate whether two isolated tones were the same or different when the second was the same or had been altered to be 25, 35, or 45 cents sharp or flat. Second, the children discriminated the tones within the context of melody. They were asked whether two melodies were the same or different when the leading tone of the second melody was the same or had been altered to be 25, 35, or 45 cents sharp or flat. Long-term memory for melody was also investigated, as the children attempted to recall four different two-bar melodies after 1 week. The children with autism spectrum disorders demonstrated elevated pitch discrimination ability in the single-tone and melodic context as well as superior long-term memory for melody. Pitch memory correlated positively with scores on measures of nonverbal fluid reasoning ability. Superior short- and long-term pitch memory was found among children with autism spectrum disorders. The results indicate an aspect to cognitive functioning that may predict both enhanced nonverbal reasoning ability and atypical language development.

Criteria and tools for objectively analysing the vocal accuracy of a popular song

This study aims to validate our method for measuring accuracy in a melodic context. We analysed the popular song ‘Happy Birthday’ sung by 63 occasional and 14 professional singers thanks to AudioSculpt and OpenMusic (IRCAM, Paris, France). In terms of evaluation of the pitch interval deviation, we replicated the profile of occasional singers described in the literature (the slower the performance, the more accurate it is). Our results also confirm that the professional singers sing more accurately than occasional singers but not when a Western operatic singing technique is involved. These results support the relevance of our method for analysing vocal accuracy of occasional and professional singers and led us to discuss adaptations to be implemented for analysing the accuracy of operatic voices.

The Development of a String Sight-Reading Pitch Skill Hierarchy

This study was designed to determine a pitch skill hierarchy for string sight-reading, to determine the effects of key on string sight-reading achievement, and to determine the validity of a tonal pattern system as a measurement of melodic sight-reading skill for string players. High school string students ( n = 94) obtained a mean score of 27.28 out of 31 on a modified version of the Vocal Sight-Reading Inventory. Success rates ranging from .99 to .72 were established for 31 pitch skills, grouped into eight tonal categories. Significant differences were found between skills appearing in the keys of D and E, with 11 of 31 skills obtaining significantly differing results by key. A .95 correlation between note-by-note and skill-based scoring systems indicates that skill-based scoring is a valid measurement of string players’ sight-reading of tonal pitch skills within a melodic context. Researchers should explore whether these pitch skills hold their relative difficulty level with less accomplished players and between instrument types, establish a hierarchy for rhythm skills, further investigate the effect of key, and identify anxiety levels for sight-reading when the consequences of performance quality vary.

Tracking of pitch probabilities in congenital amusia

Auditory perception involves not only hearing a series of sounds but also making predictions about future ones. For typical listeners, these predictions are formed on the basis of long-term schematic knowledge, gained over a lifetime of exposure to the auditory environment. Individuals with a developmental disorder known as congenital amusia show marked difficulties with music perception and production. The current study investigated whether these difficulties can be explained, either by a failure to internalise the statistical regularities present in music, or by a failure to consciously access this information. Two versions of a melodic priming paradigm were used to probe participants’ abilities to form melodic pitch expectations, in an implicit and an explicit manner. In the implicit version (Experiment 1), participants made speeded, forced-choice discriminations concerning the timbre of a cued target note. In the explicit version (Experiment 2), participants used a 1–7 rating scale to indicate the degree to which the pitch of the cued target note was expected or unexpected. Target notes were chosen to have high or low probability in the context of the melody, based on the predictions of a computational model of melodic expectation. Analysis of the data from the implicit task revealed a melodic priming effect in both amusic and control participants whereby both groups showed faster responses to high probability than low probability notes rendered in the same timbre as the context. However, analysis of the data from the explicit task revealed that amusic participants were significantly worse than controls at using explicit ratings to differentiate between high and low probability events in a melodic context. Taken together, findings from the current study make an important contribution in demonstrating that amusic individuals track melodic pitch probabilities at an implicit level despite an impairment, relative to controls, when required to make explicit judgments in this regard. However the unexpected finding that amusics nevertheless are able to use explicit ratings to distinguish between high and low probability notes (albeit not as well as controls) makes a similarly important contribution in revealing a sensitivity to musical structure that has not previously been demonstrated in these individuals.

Detecting Scale Violations in Absence of Mismatch Requires Music-Syntactic Analysis: A Further Look at the Early Right Anterior Negativity (ERAN)

The purpose of this study was to determine whether infrequent scale violations in a sequence of in-key notes are detected when the deviants are matched for frequency of occurrence and preceding intervals with the control notes. We further investigated whether the detectability of scale violations is modulated by the presence of melodic context and by the level of musical training. Event related potentials were recorded from 14 musicians and 13 non-musicians. In non-musicians, the out-of-key notes elicited an early right anterior negativity (ERAN), which appeared prominently over right frontal sites only when presented within structured sequences; no effects were found when the out-of-key notes were presented within scrambled sequences. In musicians, the out-of-key notes elicited a similar bilateral ERAN in structured and scrambled sequences. Our findings suggest that scale information is processed at the level of music-syntactic analysis, and that the detection of deviants does not require activation of auditory sensory memory by mismatch effects. Scales are perceived as a broader context, not just as online interval relations. Additional melodic context information appears necessary to support the representation of scale deviants in non-musicians, but not in musically-trained individuals, likely as a consequence of stronger pre-existing representations.

Event-related potential (ERP) markers of melodic processing: The N2 component is modulated by structural complexity, not by melodic ‘meaningfulness’

Previous studies have demonstrated that the event-related potential (ERP) evoked by a note shows substantial differences depending on whether the note is part of a melodic context or presented in an unstructured repetition. In particular, the N2 component has been found to have considerably increased latency and a more frontal topography for notes presented in a melody. An open question is whether such effect is related to the ‘meaningfulness’ of a note sequence, that is due to the formation of abstract melodic entities, rather than more simply an indicator of cognitive load associated with processing a structurally-complex sequence as opposed to an unstructured repetition. In this study, we addressed this issue by recording ERPs from 10 healthy non-musicians listening to eight one-part unfamiliar tonal melodies and eight sequences of random notes. The two stimuli were matched for distribution of pitch, intervals and note duration as well as for entropy of the time-series of pitch and duration. While tonal melodies were rated more meaningful ( p < 0.001) and pleasant ( p < 0.001) by all participants, no effects were found for the N2 component amplitude ( p ≥ 0.8) and latency ( p = 0.2). Combined with previous findings, this indicates that the N2 evoked by each individual note responds to the structural complexity of the note sequence, i.e., to the presence of pitch and duration changes, but not to higher-level processing related to the formation of abstract melodic entities. In contrast, we found that the amplitude of the P2 component was marginally ( p = 0.04) elevated for random notes as compared to tonal melodies. This may be related to attentional modulation, or more specifically to associative components of auditory processing.

Effects of Context on Electrophysiological Response to Musical Accents

Listeners' aesthetic and emotional responses to music typically occur in the context of long musical passages that contain structures defined in terms of the events that precede them. We describe an electrophysiological study of listeners' brain responses to musical accents that coincided in longer musical sequences. Musically trained listeners performed a timbre-change detection task in which a single-tone timbre change was positioned within 4-bar melodies composed of 350-ms tones to coincide or not with melodic contour accents and temporal accents (induced with temporal gaps). Event-related potential responses to (task-relevant) attended timbre changes elicited an early negativity (MMN/N2b) around 200 ms and a late positive component around 350 ms (P300), reflecting updating of the timbre change in working memory. The amplitudes of both components changed systematically across the sequence, consistent with expectancy-based context effects. Furthermore, melodic contour changes modulated the MMN/N2b response (but not the P300) to timbre changes in later sequence positions. In contrast, task-irrelevant temporal gaps elicited an MMN that was not modulated by position within the context; absence of a P300 indicated that temporal-gap accents were not updated in working memory. Listeners' neural responses to musical structure changed systematically as sequential predictability and listeners' expectations changed across the melodic context.

Investigating cochlear implant place-pitch perception with the Modified Melodies test

There has been speculation that cochlear implant place-of-excitation cues could be more akin to the brightness attribute of timbre (the spectral profile) than to melodic pitch. As brightness can be ordered on a low-to-high scale, it would allow high scores on pitch-ranking tests. In contrast, the Modified Melodies test measures pitch perception in a melodic context. In each trial, a familiar melody was presented twice. In one presentation, randomly selected, the pitch was deliberately modified. The subject's task was to select the un-modified melody. Six Nucleus implant recipients were tested with melodies presented as pure tones in the frequency range C5–C6 (523–1046 Hz) through the ACE strategy on the Freedom processor. All subjects were able to identify incorrect melodic contours and three subjects were able to recognize errors in musical intervals. These results are consistent with the hypothesis that cochlear implant place cues alone are sufficient to convey a melody. Copyright © 2009 John Wiley & Sons, Ltd.

Range of Tuning for Tones With and Without Vibrato

PREVIOUS STUDIES ON VIBRATO PITCH HAVE attempted to determine the "principal pitch" of vibrato tones and not the range of tuning for such tones in a melodic context. This study investigates the range of acceptable tuning (RAT) for tones with and without vibrato, using repeating ascending and descending three-tone arpeggios. The second tone in each sequence was modulated or unmodulated with an initial pitch that was flat or sharp. With each repetition the pitch of the second tone increased or decreased by 3 cents depending upon the initial pitch condition. Participants indicated when they perceived the second tone to be in tune and out of tune. The RAT for vibrato tones was approximately 10 cents greater than for unmodulated tones. This is largely due to the lower RAT limit, indicating that the carrier frequency of vibrato tones can be 10 cents flatter than the fundamental frequency of an unmodulated tone.

Tonal expectations influence pitch perception

In this study, we investigated the influence of tonal relatedness on pitch perception in melodies. Tonal expectations for target tones were manipulated in melodic contexts while controlling sensory expectations, thus allowing us to assess specifically the influence oftonal expectations on pitch perception. Three experimentsprovided converging evidence that tonal relatedness modulates pitch perception in nonmusician listeners. Experiment 1 showed, with a rating task, the influence of the tonal relatedness of a target tone on listeners' judgments of tuning/mistuning. Experiment 2 showed, with a priming task, that pitch processing of in-tune tones was faster for tonally related targets than for less related targets. Experiment 3 showed, with a comparison task, that discrimination performance for small mistunings was better when the to-be-compared tones were tonally related to the melodic context. Findings are discussed in relation to psychoacoustic research on contextual pitch perception and to studies showing facilitation of early processing steps via knowledge- and attention-related processes.

Melody Extraction and Musical Onset Detection via Probabilistic Models of Framewise STFT Peak Data

We propose a probabilistic method for the joint segmentation and melody extraction for musical audio signals which arise from a monophonic score. The method operates on framewise short-time Fourier transform (STFT) peaks, enabling a computationally efficient inference of note onset, duration, and pitch attributes while retaining sufficient information for pitch determination and spectral change detection. The system explicitly models note events in terms of transient and steady-state regions as well as possible gaps between note events. In this way, the system readily distinguishes abrupt spectral changes associated with musical onsets from other abrupt change events. Additionally, the method may incorporate melodic context by modeling note-to-note dependences. The method is successfully applied to a variety of piano and violin recordings containing reverberation, effective polyphony due to legato playing style, expressive pitch variations, and background voices. While the method does not provide a sample-accurate segmentation, it facilitates the latter in subsequent processing by isolating musical onsets to frame neighborhoods and identifying possible pitch content before and after the true onset sample location