Search

While language is expressed in multiple modalities, including sign, writing, or whistles, speech is arguably the most common. The human vocal tract is capable of producing the bewildering diversity of the 7000 or so currently spoken languages, but relatively little is known about its genetic bases, especially in what concerns normal variation. Here, we capitalize on five cohorts totaling 632 Dutch twins with structural magnetic resonance imaging (MRI) data. Two raters placed clearly defined (semi)landmarks on each MRI scan, from which we derived 146 measures capturing the dimensions and shape of various vocal tract structures, but also aspects of the head and face. We used Genetic Covariance Structure Modeling to estimate the additive genetic, common environmental or non-additive genetic, and unique environmental components, while controlling for various confounds and for any systematic differences between the two raters. We found high heritability, h2, for aspects of the skull and face, the mandible, the anteroposterior (horizontal) dimension of the vocal tract, and the position of the hyoid bone. These findings extend the existing literature, and open new perspectives for understanding the complex interplay between genetics, environment, and culture that shape our vocal tracts, and which may help explain cross-linguistic differences in phonetics and phonology.

Purpose: If language production is impaired, will gestures compensate? Evidence in favor of this prediction has often been argued to come from aphasia, but it remains contested. Here, we tested whether thought content not present in speech due to language impairment is manifested in gestures, in 20 people with dysfluent (Broca’s) aphasia, 20 people with fluent (Wernicke’s) aphasia, and 20 matched neurotypical controls. Method: A new annotation scheme was created distinguishing types of gestures and whether they co-occurred with fluent or dysfluent/absent speech and were temporally aligned in content with coproduced speech. Results: Across both aphasia types, noncontent (beat) gestures, which by their nature cannot compensate for lost speech content, constituted the greatest proportion of all types of gestures produced. Content (i.e., descriptive, referential, and metaphorical) gestures were largely coproduced with fluent rather than dysfluent speech and tended to be aligned with the content conveyed in speech. They also did not differ in quantity depending on whether the dysfluencies were eventually resolved or not. Neither aphasia severity nor comprehension ability had an impact on the total amount of content gesture produced in people with aphasia, which was instead positively correlated with speech fluency. Conclusions: Together, these results suggest that gestures are unlikely to have a role in compensating for linguistic deficits and to serve as a representational system conveying thought content independent of language. Surprisingly, aphasia rather is a model of how gesture and language are inherently integrated and aligned: Even when language is impaired, it remains the essential provider of content.

Comunicació a congrès: 4th Phonetics and Phonology in Europe 2021 (PaPE 2021): Phonetics and Phonology: Real-world applications. 21 to 23 June 2021. Barcelona. Virtual.

Neural oscillations constitute an intrinsic property of functional brain organization that facilitates the tracking of linguistic units at multiple time scales through brain-to-stimulus alignment. This ubiquitous neural principle has been shown to facilitate speech segmentation and word learning based on statistical regularities. However, there is no common agreement yet on whether speech segmentation is mediated by a transition of neural synchronization from syllable to word rate, or whether the two time scales are concurrently tracked. Furthermore, it is currently unknown whether syllable transition probability contributes to speech segmentation when lexical stress cues can be directly used to extract word forms. Using inter-trial coherence (ITC) analyses in combinations with Event-Related Potentials (ERPs), we showed that speech segmentation based on both statistical regularities and lexical stress cues was accompanied by concurrent neural synchronization to syllables and words. In particular, ITC at the word rate was generally higher in structured compared to random sequences, and this effect was particularly pronounced in the flat condition. Furthermore, ITC at the syllable rate dynamically increased across the blocks of the flat condition, whereas a similar modulation was not observed in the stressed condition. Notably, in the flat condition ITC at both time scales correlated with each other, and changes in neural synchronization were accompanied by a rapid reconfiguration of the P200 and N400 components with a close relationship between ITC and ERPs. These results highlight distinct computational principles governing neural synchronization to pertinent linguistic units while segmenting speech under different listening conditions.

Highlights • Sixty 9-year-old children performed an audio-visual statistical learning task. • Children showed successful recognition of words and word-to-picture associations. • Enhanced neural tracking at word and syllable rate for consistent audio-visual streams. • Enhanced long-distance scalp-based phase synchronization suggested multi-sensory integration. • Multi-sensory cues may increase attention to enhance learning in children., When learning a new language, one must segment words from continuous speech and associate them with meanings. These complex processes can be boosted by attentional mechanisms triggered by multi-sensory information. Previous electrophysiological studies suggest that brain oscillations are sensitive to different hierarchical complexity levels of the input, making them a plausible neural substrate for speech parsing. Here, we investigated the functional role of brain oscillations during concurrent speech segmentation and meaning acquisition in sixty 9-year-old children. We collected EEG data during an audio-visual statistical learning task during which children were exposed to a learning condition with consistent word-picture associations and a random condition with inconsistent word-picture associations before being tested on their ability to recall words and word-picture associations. We capitalized on the brain dynamics to align neural activity to the same rate as an external rhythmic stimulus to explore modulations of neural synchronization and phase synchronization between electrodes during multi-sensory word learning. Results showed enhanced power at both word- and syllabic-rate and increased EEG phase synchronization between frontal and occipital regions in the learning compared to the random condition. These findings suggest that multi-sensory cueing and attentional mechanisms play an essential role in children's successful word learning.

Epidemiological studies suggest that bipolar disorder has a prevalence of about 1% in European countries, becoming one of the most disabling illnesses in working age adults, and often long-term and persistent with complex management and treatment. Therefore, the capacity of home monitoring for patients with this disorder is crucial for their quality of life. The current paper introduces the use of speech-based information as an easy-to-record, ubiquitous and non-intrusive health sensor suitable for home monitoring, and its application in the framework on the NYMPHA-MD project. Some preliminary results also show the potential of acoustic and prosodic features to detect and classify bipolar disorder, by predicting the values of the Hamilton Depression Rating Scale (HDRS) and the Young Mania Rating Scale (YMRS) from speech. The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is part of the NYMPHA-MD project, which has been funded by the European Union under Grant Agreement No. 610462. The first author has been funded by the Agencia Estatal de Investigación (AEI), Ministerio de Ciencia, Innovación y Universidades and the Fondo Social Europeo (FSE) under grant RYC-2015-17239 (AEI/FSE, UE).

The use of private layers is explored for a multi-task learning approach to speech emotion recognition. The methods used include LSTMs, multi-task learning, high level representation, and "in the wild" testing.

In electroencephalography (EEG) measurements, processing of periodic sounds in the ascending auditory pathway generates the frequency-following response (FFR) phase-locked to the fundamental frequency (F0) and its harmonics of a sound. We measured FFRs to the steady-state (vowel) part of syllables /ba/ and /aw/ occurring in binaural rapid streams of speech sounds as frequently repeating standard syllables or as infrequent (p = 0.2) deviant syllables among standard /wa/ syllables. Our aim was to study whether concurrent active phonological processing affects early processing of irrelevant speech sounds reflected by FFRs to these sounds. To this end, during syllable delivery, our healthy adult participants performed tasks involving written letters delivered on a computer screen in a rapid stream. The stream consisted of vowel letters written in red, infrequently occurring consonant letters written in the same color, and infrequently occurring vowel letters written in blue. In the phonological task, the participants were instructed to press a response key to the consonant letters differing phonologically but not in color from the frequently occurring red vowels, whereas in the non-phonological task, they were instructed to respond to the vowel letters written in blue differing only in color from the frequently occurring red vowels. We observed that the phonological task enhanced responses to deviant /ba/ syllables but not responses to deviant /aw/ syllables. This suggests that active phonological task performance may enhance processing of such small changes in irrelevant speech sounds as the 30-ms difference in the initial formant-transition time between the otherwise identical syllables /ba/ and /wa/ used in the present study.

We introduce a deceptively simple behavioral task that robustly identifies two qualitatively different groups within the general population. When presented with an isochronous train of random syllables, some listeners are compelled to align their own concurrent syllable production with the perceived rate, whereas others remain impervious to the external rhythm. Using both neurophysiological and structural imaging approaches, we show group differences with clear consequences for speech processing and language learning. When listening passively to speech, high synchronizers show increased brain-to-stimulus synchronization over frontal areas, and this localized pattern correlates with precise microstructural differences in the white matter pathways connecting frontal to auditory regions. Finally, the data expose a mechanism that underpins performance on an ecologically relevant word-learning task. We suggest that this task will help to better understand and characterize individual performance in speech processing and language learning. A simple behavioral task identifies two qualitatively different groups within the general population, according to their speech-to-speech synchronization abilities. Group pertinence predicts brain function and anatomy, as well as word-learning performance.

Speaking fluently requires three main processes to run smoothly: conceptualization, formulation, and articulation. This study investigates to what extent fluency in spontaneous speech in both first (L1) and second (L2) languages can be explained by individual differences in articulatory skills. A group of L2 English learners (n = 51) performed three semispontaneous speaking tasks in their L1 Spanish and in their L2 English. In addition, participants performed articulatory skill tasks that measured the speed at which their articulatory speech plans could be initiated (delayed picture naming) and the rate and accuracy at which their articulatory gestures could be executed (diadochokinetic production). The results showed that fluency in spontaneous L2 speech can be predicted by L1 fluency, replicating earlier studies and showing that L2 fluency measures are, to a large degree, measures of personal speaking style. Articulatory skills were found to contribute modestly to explaining variance in both L1 and L2 fluency.

Contains fulltext : 190594.pdf (Publisher’s version ) (Open Access) The human brain can quickly adapt to changes in the environment. One example is phonetic recalibration: a speech sound is interpreted differently depending on the visual speech and this interpretation persists in the absence of visual information. Here, we examined the mechanisms of phonetic recalibration. Participants categorized the auditory syllables /aba/ and /ada/, which were sometimes preceded by the so-called McGurk stimuli (in which an /aba/ sound, due to visual /aga/ input, is often perceived as 'ada'). We found that only one trial of exposure to the McGurk illusion was sufficient to induce a recalibration effect, i.e. an auditory /aba/ stimulus was subsequently more often perceived as 'ada'. Furthermore, phonetic recalibration took place only when auditory and visual inputs were integrated to 'ada' (McGurk illusion). Moreover, this recalibration depended on the sensory similarity between the preceding and current auditory stimulus. Finally, signal detection theoretical analysis showed that McGurk-induced phonetic recalibration resulted in both a criterion shift towards /ada/ and a reduced sensitivity to distinguish between /aba/ and /ada/ sounds. The current study shows that phonetic recalibration is dependent on the perceptual integration of audiovisual information and leads to a perceptual shift in phoneme categorization. 13 p.

Highlights • Amount of attention to rules during artificial language learning was manipulated. • Indirect measures showed incidental rule learning irrespective of attention. • Explicit knowledge after learning was affected by the amount of attention. • The amount of attention at encoding did not affect consolidation after sleep., Incidental learning plays a crucial role in the initial phases of language acquisition. However the knowledge derived from implicit learning, which is based on prediction-based mechanisms, may become explicit. The role that attention plays in the formation of implicit and explicit knowledge of the learned material is unclear. In the present study, we investigated the role that attention plays in the acquisition of non-adjacent rule learning from speech. In addition, we also tested whether the amount of attention during learning changes the representation of the learned material after a 24 h delay containing sleep. For that, we developed an experiment run on two consecutive days consisting on the exposure to an artificial language that contained non-adjacent dependencies (rules) between words whereas different conditions were established to manipulate the amount of attention given to the rules (target and non-target conditions). Furthermore, we used both indirect and direct measures of learning that are more sensitive to implicit and explicit knowledge, respectively. Whereas the indirect measures indicated that learning of the rules occurred regardless of attention, more explicit judgments after learning showed differences in the type of learning reached under the two attention conditions. 24 hours later, indirect measures showed no further improvements during additional language exposure and explicit judgments indicated that only the information more robustly learned in the previous day, was consolidated.

©2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Speech enhancement deep learning systems usually require large amounts of training data to operate in broad conditions or real applications. This makes the adaptability of those systems into new, low resource environments an important topic. In this work, we present the results of adapting a speech enhancement generative adversarial network by fine-tuning the generator with small amounts of data. We investigate the minimum requirements to obtain a stable behavior in terms of several objective metrics in two very different languages: Catalan and Korean. We also study the variability of test performance to unseen noise as a function of the amount of different types of noise available for training. Results show that adapting a pre-trained English model with 10 min of data already achieves a comparable performance to having two orders of magnitude more data. They also demonstrate the relative stability in test performance with respect to the number of training noise types.

[eng] Vocal learning, one of the subcomponents of language, is put at center stage in this dissertation. The overall hypothesis is that vocal learning lays the foundation for both language evolution (phylogeny) and development (ontogeny), and also high-level cognition. The computational ability found in vocal learning is seen as so enhanced in humans as to yield the kind of recursion that supports language. Empirical evidence on vocal learning in nonhuman animals and humans from behavioral, neuroanatomical, neurophysiological, genetic, and evolutionary fields is suggestive that vocal learning interacts with other cognitive domains at multiple levels. The positive correlation between the hippocampal volume and open-ended vocal production in avian vocal learning species suggests the possible involvement of the hippocampus in vocal learning. The empirical studies of foxp2 in nonhuman animals and humans suggest that foxp2 plays a role in multimodal communication and general cognition. Phylogenetically, Sapiens’ vocal learning abilities are unique among primates. Compared with nonhuman primates, our species possesses stronger and more enhanced connections between the superior temporal cortex and premotor cortex as well as the striatum. In Sapiens, meaning aside, vocal learning as such can explain many features found in speech and its ontogeny such as the specialized auditory mechanism for speech, the preferential attention to speech in newborns, the primacy of vocal imitation among multimodal (visual and auditory) imitative skills and the stages seen in learning to speak. All these characteristics seem to be different and abnormal, albeit to different degrees, in autism. A 25-30% of the autistic population is non/minimally verbal but even the high functioning end of the autistic spectrum presents with abnormalities, such as difficulties in processing speed and an impaired imitative capacity that could be satisfactorily explained if language entered again the definition (and diagnosis) of what autism is, with an special emphasis on vocal learning., [cat] El vocal learning, un dels subcomponents del llenguatge, ocupa un espai central en aquesta tesi. La hipòtesi general és que el vocal learning constitueix el fonament de l’evolució (filogènia) i del desenvolupament (ontogènia) lingüístics, i també de la cognició. L'habilitat computacional que es dóna en el vocal learning es veu en els humans tan potenciada com per ser la base del tipus de recursió en què es basa el llenguatge. Proves empíriques sobre el vocal learning en animals no humans i en humans, des de camps que inclouen des del comportament, la neuroanatomia, la neurofisiologia, la genètica i la teoria de l’evolució, suggereixen que el vocal learning interactua amb altres dominis cognitius a molts i diferents nivells. Filogenèticamet, les habilitats de vocal learning en el Sapiens són úniques entre els primats. Comparada amb els primats no humans, la nostra espècie posseeix unes connexions més denses i potents entre el còrtex temporal superior i el còrtex premotor així com l’estriat. En el Sapiens, deixant de banda el significat, el vocal learning tot sol pot explicar molts trets de la parla i la seva ontogènia com ara l’especialització auditiva per a la parla, l’atenció preferent a la parla en els nadons, la primacia de la imitació vocal entre les habilitats imitatives multimodals (de base visual i auditiva), i els estadis que s’observen en l’adquisició de la parla. Totes aquestes característiques sembla que són diferents i anòmales, tot i que en diferent graus, en l’autisme. Un 25-30% de la població autista és no verbal o mínimament però fins i tot a la banda de l’espectre autista que es considera d’alt funcionament s’hi donen anomalies, tal com ara un cert dèficit en velocitat de processament i una capacitat deficient d’imitació, que podrien explicar-se més satisfactòriament si un dèficit de llenguatge entrés altra vegada a la definició (i diagnòstic) del que és l’autisme, amb un èmfasi especial en el vocal learning.

This work introduces new results on the modeling of early-vocal development using artificial intelligent cognitive architectures and a simulated vocal tract. The problem is addressed using intrinsically-motivated learning algorithms for autonomous sensorimotor exploration, a kind of algorithm belonging to the active learning architectures family. The artificial agent is able to autonomously select goals to explore its own sensorimotor system in regions where its competence to execute intended goals is improved. We propose to include a somatosensory system to provide a proprioceptive feedback signal to reinforce learning through the autonomous discovery of motor constraints. Constraints are represented by a somatosensory model which is unknown beforehand to the learner. Both the sensorimotor and somatosensory system are modeled using Gaussian mixture models. We argue that using an architecture which includes a somatosensory model would reduce redundancy in the sensorimotor model and drive the learning process more efficiently than algorithms taking into account only auditory feedback. The role of this proposed system is to predict whether an undesired collision within the vocal tract under a certain motor configuration is likely to occur. Thus, compromised motor configurations are rejected, guaranteeing that the agent is less prone to violate its own constraints.

The present paper deals with the description and characterization of the melodic patterns of absolute interrogative utterances in Northern German in spontaneous speech from an intonation and semantic pragmatic point of view. This research has been carried out based on 246 absolute questions from spontaneous speech settings by multiple speakers of different gender, age and education by applying the Melodic Analysis of Speech (MAS) method developed by Cantero (2002). As a result, we found five intonation patterns for absolute questions in German: Falling Final Inflection, Rising Final Inflection, Rising -falling Final Inflection, High nucleus Final Inflection and Rising body and Final Inflection. The first three have been previously defined by researchers using the ToBI method, whilst the fourth and fifth patterns had not, to date, been described in this context. In addition to defining the melodic features of each one in spontaneous speech, we have also contributed to providing the different pragmatic meanings discovered in the various contexts in which each pattern appears.

This study aims to demonstrate the potential for understanding first millennium glass beads not as individual representatives of types, but as collections of objects brought together and curated by owners. It uses the author’s experience as a skilled bead maker to investigate processes of bead production and mechanics of bead collection current in Scandinavia and Anglo-Saxon England in the period of 6th to 9th century AD. In the study the bead collections of seven graves are examined from the perspective of their production techniques, materials, and damage from wear and cremation. The results point to beads being acquired in different numbers and often worn for long periods of time before being buried.

There is still no categorical answer for why humans, and no other species, have speech, or why speech is the way it is. Several purely anatomical arguments have been put forward, but they have been shown to be false, biologically implausible, or of limited scope. This perspective paper supports the idea that evolutionary theories of speech could benefit from a focus on the cognitive mechanisms that make speech possible, for which antecedents in evolutionary history and brain correlates can be found. This type of approach is part of a very recent, but rapidly growing tradition, which has provided crucial insights on the nature of human speech by focusing on the biological bases of vocal learning. Here, we call attention to what might be an important ingredient for speech. We contend that a general mechanism of attention, which manifests itself not only in visual but also auditory (and possibly other) modalities, might be one of the key pieces of human speech, in addition to the mechanisms underlying vocal learning, and the pairing of facial gestures with vocalic units.

This paper presents the perceptual experiments that were carried out in order to validate the methodology of transforming expressive speech styles using voice quality (VoQ) parameters modelling, along with the well-known prosody (F0, duration, and energy), from a neutral style into a number of expressive ones. The main goal was to validate the usefulness of VoQ in the enhancement of expressive synthetic speech in terms of speech quality and style identification. A harmonic plus noise model (HNM) was used to modify VoQ and prosodic parameters that were extracted from an expressive speech corpus. Perception test results indicated the improvement of obtained expressive speech styles using VoQ modelling along with prosodic characteristics.

In this demo we will briefly outline the scope of the european EUNISON project, which aims at a unified numerical simulation of the physics of voice by resorting to supercomputer fa-cilities, and present some of its preliminary results obtained to date.

Language acquisition is a complex process that requires the synergic involvement of different cognitive functions, which include extracting and storing the words of the language and their embedded rules for progressive acquisition of grammatical information. As has been shown in other fields that study learning processes, synchronization mechanisms between neuronal assemblies might have a key role during language learning. In particular, studying these dynamics may help uncover whether different oscillatory patterns sustain more item-based learning of words and rule-based learning from speech input. Therefore, we tracked the modulation of oscillatory neural activity during the initial exposure to an artificial language, which contained embedded rules. We analyzed both spectral power variations, as a measure of local neuronal ensemble synchronization, as well as phase coherence patterns, as an index of the long-range coordination of these local groups of neurons. Synchronized activity in the gamma band (20–40 Hz), previously reported to be related to the engagement of selective attention, showed a clear dissociation of local power and phase coherence between distant regions. In this frequency range, local synchrony characterized the subjects who were focused on word identification and was accompanied by increased coherence in the theta band (4–8 Hz). Only those subjects who were able to learn the embedded rules showed increased gamma band phase coherence between frontal, temporal, and parietal regions.

This paper describes a multi-domain text-to-speech (MD-TTS)synthesis strategy for generating speech among different domainsand so increasing the flexibility of high quality TTS systems. Tothat effect, the MD-TTS introduces a flexible TTS architecture thatincludes an automatic domain classification module, which allowsMD-TTS systems to be implemented by different synthesis strate-gies and speech corpus typologies. In this work, the performanceof a corpus-based MD-TTS system is subjectively validated bymeans of several perceptual tests

In this paper, a generic pitch marks filtering algorithm (PMFA) is introduced in order to achieve reliable and smooth pitch marks from any input pitch tracking or marking algorithm. The proposed PMFA is a simple yet effective filtering process based on restricted dynamic programming, but very helpful for minimizing human intervention when creating large speech corpora. Moreover, this work introduces a novel pitch marking evaluation measure for directly comparing pitch marking algorithms with different location criteria. The experiments demonstrate that the proposed PFMA improves the results of the input state-of-the-art pitch tracking and marking algorithms dramatically.

This work is a step further in our research towards developing a new strategy for high quality text-to-speech (TTS) synthesis among different domains. In this context, it is necessary to select the most appropriate domain for synthesizing the text input to the TTS system, task that can be solved including a text classifier (TC) in the classic TTS architecture. Since speech speaking style and prosody depend on the sequentiality and text structure of the message, the TC should consider not only thematic but also stylistic aspects of text. To this end, we introduce a new text modelling scheme based on an associative relational network, which represents texts as a weighted word-based graph. The conducted experiments validate the proposal in terms of both objective (text classification efficiency) and subjective (perceived synthetic speech quality) evaluation criteria.

This paper presents a new approach for designing aconcatenative text-to-speech (TTS) system based onmulti-domain unit selection. The method achievesgood synthetic quality with reasonable computationalcost for a general-purpose TTS system. The architec-ture of the multi-domain database and the text classi-fication algorithm for domain assignment are the basisof the method. The performance of the adjusted textclassification algorithm for the multi-domain TTS aimis analyzed in several encouraging experiments.

The combination of Vector Quantization (VQ) and predictive techniques -named AVPC systems- has been shown as interesting systems for speech waveform coding at medium-high rates (ref. 1 to 3). In this work we present a such system including a previous four subbbands splitting. This allows good quality of speech at low-medium rates (1-1.25 bits/sample). A comparative study show that the AVPC-SBC outperforms the simple AVPC coder.

This paper presents an algorithm that provides a syllabic segmentation of speech following the syllabification rules of Spanish language. The implemented algorithm is divided into two parts. First, an initial segmentation is made based on energy contour, sonority and duration. Second, a fine adjustement of syllable boundaries and final segmentation is made by applying syllabic rules.

This communication reports the use of demisyllables for continuous speech recognition in a specific application: the recognition of Spanish numbers. After a brief outline of the recognition system, a description of demisyllable syntactic constraints and one-speaker reference generation is provided. Finally, the recognition performance is assessed by means of two experiments: the recognition of integer numbers from zero to one thousand and telephone numbers uttered in a Spanish way (strings of integers from zero to ninety nine), in both applications the results that the system yielded were excellent.

The use of an orthogonal expansion for feature selection and data compression in isolated word recognition is presented. Assuming that the spectral evolution, given by a LPC analysis, is a noisy measure in the sense that it is a linear combination of a set of real features, the objective of the orthogonal expansion is to find these real features. The new feature vectors are used to perform the recognition process. The recognition results as well as the meaning of the orthogonal expansion are discussed

Envelope and instantaneous phase of a real signal are two concepts mainly associated with modulation systems. In this paper is shown how envelope and lnstantaneous phase can be applied in coding systems. In particular, the envelope of t he residual provides informatton ahnut pitch ilnd energy of the signal and can be user to properly codify it. A parametric version of the envelope is taken from the residual of a s ignal and this informat ion ls used ln two different coders: an APC corleranda stochastlcally excited coder. In APC coder the residual envelope is used to quantize the prediction error providing step size and dynamlc bU assignment information. In stochastically excitec coders, the input to the synthesls filter is composecl by the product of the parametric envelope of the residual multiplied by an instantaneous phase taken from a codebook of random phases .