Abstract

Automatic detection of prominence at the word and syllable-levels is critical for building computer-assisted language learning systems. It has been shown that prosody embeddings learned by the current state-of-the-art (SOTA) text-to-speech (TTS) systems could generate word- and syllable-level prominence in the synthesized speech as natural as in native speech. To understand the effectiveness of prosody embeddings from TTS for prominence detection under nonnative context, a comparative analysis is conducted on the embeddings extracted from native and non-native speech considering the prominence-related embeddings: duration, energy, and pitch from a SOTA TTS named FastSpeech2. These embeddings are extracted under two conditions considering: 1) only text, 2) both speech and text. For the first condition, the embeddings are extracted directly from the TTS inference mode, whereas for the second condition, we propose to extract from the TTS under training mode. Experiments are conducted on native speech corpus: Tatoeba, and non-native speech corpus: ISLE. For experimentation, word-level prominence locations are manually annotated for both corpora. The highest relative improvement on word & syllable-level prominence detection accuracies with the TTS embeddings are found to be 13.7% & 5.9% and 16.2% & 6.9% compared to those with the heuristic-based features and self-supervised Wav2Vec-2.0 representations, respectively.

Abstract

—Speech disfluencies in spontaneous communication can be categorized as either typical or atypical. Typical disfluencies, such as hesitations and repetitions, are natural occurrences in everyday speech, while atypical disfluencies are indicative of pathological disorders like stuttering. Distinguishing between these categories is crucial for improving voice assistants (VAs) for Persons Who Stutter (PWS), who often face premature cutoffs due to misidentification of speech termination. Accurate classification also aids in detecting stuttering early in children, preventing misdiagnosis as language development disfluency. This research introduces the IIITH-TISA dataset, the first Indian English stammer corpus, capturing atypical disfluencies. Additionally, we extend the IIITH-IED dataset with detailed annotations for typical disfluencies. We propose Perceptually Enhanced Zero-Time Windowed Cepstral Coefficients (PE-ZTWCC) combined with Shifted Delta Cepstra (SDC) as input features to a shallow Time Delay Neural Network (TDNN) classifier, capturing both local and wider temporal contexts. Our method achieves an average F1 score of 85.01% for disfluency classification, outperforming traditional features.

Abstract

Stuttering is recognized as a prevalent speech disorder that significantly affects individuals worldwide. Identifying and diagnosing in the early stages enhances the quality of life for individuals experiencing atypical speech patterns. Traditional methods for classifying stuttering primarily depend on subjective assessments and short-term acoustic analysis. Although helpful, these methods face limitations in accurately capturing all stutter types due to their inherent subjectivity and temporal constraints. This study uses the long-term average spectrum (LTAS) values for stutter classification derived from various filter banks such as Constant Q, Gamma-tone, and Single-frequency filter banks. It also compares these LTAS-based methods with cepstral coefficients, such as MFCC and ZTWCC. Classifiers such as SVM, LSTM, and Bi-LSTM deep networks were used to study the effectiveness of these representations in accurately discerning stuttered speech from fluent speech and reported the results.

Abstract

Automatic detection and severity assessment of dysarthria are crucial for delivering targeted therapeutic interventions to patients. While most existing research focuses primarily on speech modality, this study introduces a novel approach that leverages both speech and text modalities. By employing cross-attention mechanism, our method learns the acoustic and linguistic similarities between speech and text representations. This approach assesses specifically the pronunciation deviations across different severity levels, thereby enhancing the accuracy of dysarthric detection and severity assessment. All the experiments have been performed using UA-Speech dysarthric database. Improved accuracies of 99.53% and 93.20% in detection, and 98.12% and 51.97% for severity assessment have been achieved when speaker-dependent and speaker-independent, unseen and seen words settings are used. These findings suggest that by integrating text information, which provides a reference linguistic knowledge, a more robust framework has been developed for dysarthric detection and assessment, thereby potentially leading to more effective diagnoses.

Abstract

n personalizing interactions with smart devices, identifying keywords in an open-vocabulary context is crucial. Previous methods for open-vocabulary keyword spotting relied on a shared embedding space created by audio and text encoders. However, they suffered from heterogeneous modality representations, causing audio-text mismatch. To tackle this issue, our proposed framework utilizes knowledge from a pre-trained text-to-speech (TTS) system. This knowledge transfer incorporates awareness of audio projections into text representations derived from the text encoder. Consequently, this approach aids in preventing false triggers in scenarios such as closely related pronunciations of audio-text pairs. Additionally, our proposed approach benefits from the keyword embedding calculation only during keyword enrollment phase. The proposed system gives consistent performance across all the word lengths.

Abstract

Many real-world speech processing applications depend on the accurate estimation of epoch location. In a real-world scenario, speech may contain multiple speakers or multiple emotions in a single utterance. In the literature, the performance of epoch extraction algorithm was evaluated on speech utterance consisting of a single speaker or single emotion alone. In this study, speech utterance from different speakers, and emotions are combined to simulate multi-speaker and multi-emotion scenario, respectively. Five state-of-the-art epoch extraction methods were used to evaluate the performance for multi-speaker, multi-emotion, and voice disorder scenarios. We also analysed the performance of state-of-the-art epoch extraction methods after applying the region-wise approach on existing methods. CMU arctic, and Berlin emotional database is used in this study to simulate the multi-speaker, and multi-emotion scenarios, respectively for analysing the performance of epoch extraction algorithms. Saarbruecken voice disorder database is used to perform the experiments for voice disorder scenario. The result of this study indicates that performance of state-of-the art epoch extraction methods (which depend on pitch period) degrades in terms of false alarm rate (FAR) for these three scenarios. By applying the region-wise approach on these methods for extraction of epoch locations, FAR is reduced by 5% to 8% for the real-world scenario. It was also observed from this study, that the performance of dynamic programming based methods remains unchanged even after applying the region-wise approach on existing methods.

Abstract

The demand for high-quality parallel speech data has been increasing as deep learning-based Speech to Speech Machine Translation (SSMT) and automatic dubbing approaches gain popularity in speech applications. Traditional and well-established speech applications such as Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) heavily rely on large corpus of monolingual speech and the corresponding text. While there is a wealth of parallel text data available for both English and Indic languages, parallel speech data is available only for English and other European languages, yet it often lacks natural prosody and semantic alignment between the languages. For achieving cross-lingual prosody transfer, end-to-end SSMT models, and high-quality dubbing from English to Hindi, in this work, an English-Hindi parallel bilingual speech-text corpus named lIlT-Speech Twins 1.0 is created. This data contains twin-like English and Hindi speech-text pairs obtained from publicly available children's stories in both the languages, through manual and automatic processing. Starting with 8 stories in each language, totaling around 4 hours of audio, the final outcome was a 2-hour dataset. This was achieved through systematic segmentation, re-moval of non-speech background audio, and sentence-by-sentence alignment to ensure accurate meaning in both languages. In addition to ensuring proper alignment and transcription, this dataset offers a rich source of natural prosody, expressions, and emotions, due to the narrative diversity within the stories. The dataset also provides sig-nificant speaker variability, with different characters being voiced by various speakers, enhancing the richness of the lIlT-Speech Twins 1.0 corpus.

Abstract

Stuttering constitutes a pervasive speech disorder with a global impact on individuals. Detecting stuttering in speech would aid speech pathologists in monitoring fluency over time and enhance the quality of life for those with atypical speech patterns. Presently, the available public datasets tailored for stuttering identification are confined to the utterance level, with labels ascribed to individual utterances. However, in real-time situations, a considerable proportion of individuals grappling with stuttering manifest dysfluency at the syllable level. This limitation markedly constrains the effectiveness and applicability of dysfluency detection systems. In an effort to address this lacuna in knowledge, this study introduces a novel dataset comprising audio recordings directly obtained from native Kannada-speaking individuals afflicted with the stuttering disorder at the syllable level and also provides detailed information on the dataset's creation, including patient selection criteria and clip labelling. Additionally, benchmark models trained on this distinctive dataset are presented as part of this research endeavor.

Abstract

Identifying user-defined keywords is crucial for personalizing interactions with smart devices. Previous approaches of user-defined keyword spotting (UDKWS) have relied on short-term spectral features such as mel frequency cepstral coefficients (MFCC) to detect the spoken keyword. However, these features may face challenges in accurately identifying closely related pronunciation of audio-text pairs, due to their limited capability in capturing the temporal dynamics of the speech signal. To address this challenge, we propose to use shifted delta coefficients (SDC) which help in capturing pronunciation variability (transition between connecting phonemes) by incorporating long-term temporal information. The performance of the SDC feature is compared with various baseline features across four different datasets using a cross-attention based end-to-end system. Additionally, various configurations of SDC are explored to find the suitable temporal context for the UDKWS task. The experimental results reveal that the SDC feature outperforms the MFCC baseline feature, exhibiting an improvement of 8.32% in area under the curve (AUC) and 8.69% in terms of equal error rate (EER) on the challenging Libriphrase-hard dataset. Moreover, the proposed approach demonstrated superior performance when compared to state-of-the-art UDKWS techniques.

Abstract

Identifying keywords in an open-vocabulary context is crucial for personalizing interactions with smart devices. Previous approaches to open vocabulary keyword spotting dependon a shared embedding space created by audio and text encoders. However, these approaches suffer from heterogeneous modality representations (i.e., audio-text mismatch). To address this issue, our proposed framework leverages knowledge acquired from a pre-trained text-to-speech (TTS) system. This knowledge transfer allows for the incorporation of awareness of audio projections into the text representations derived from the text encoder. The performance of the proposed approach is compared with various baseline methods across four different datasets. The robustness of our proposed model is evaluated by assessing its performance across different word lengths and in an Out-of-Vocabulary (OOV) scenario. Additionally, the effectiveness of transfer learning from the TTS system is investigated by analyzing its different intermediate representations. The experimental results indicate that, in the challenging LibriPhrase Hard dataset, the proposed approach outperformed the cross-modality correspondence detector (CMCD) method by a significant improvement of 8.22% in area under the curve (AUC) and 12.56% in equal error rate (EER).