In the last years, the task of Query-by-Example Spoken Term Detection (QbE-STD), which aims to find occurrences of a spoken query in a set of audio documents, has gained the interest of the research community for its versatility in settings where untranscribed, multilingual and acoustically unconstrained spoken resources, or spoken resources in low-resource languages, must be searched. This paper describes and reports experimental results for a QbE-STD system that achieved the best performance in the recent Spoken Web Search (SWS) evaluation, held as part of MediaEval 2013. Though not optimized for speed, the system operates faster than real-time. The system exploits high-performance phone decoders to extract frame-level phone posteriors (a common representation in QbE-STD tasks). Then, given a query and a audio document, a distance matrix is computed between their phone posterior representations, followed by a newly introduced distance normalization technique and an iterative Dynamic Time Warping (DTW) matching procedure with some heuristic prunings. Results show that remarkable performance improvements can be achieved by using multiple examples per query and, specially, through the late (score-level) fusion of different subsystems, each based on a different set of phone posteriors.

In this paper, we present the task and describe the main findings of the 2014 “Query-by-Example Speech Search Task” (QUESST) evaluation. The purpose of QUESST was to perform language independent search of spoken queries on spoken documents, while targeting languages or acoustic conditions for which very few speech resources are available. This evaluation investigated for the first time the performance of query-by-example search against morphological and morpho-syntactic variability, requiring participants to match variants of a spoken query in several languages of different morphological complexity. Another novelty is the use of the normalized cross entropy cost (Cnxe) as the primary performance metric, keeping Term-Weighted Value (TWV) as a secondary metric for comparison with previous evaluations. After analyzing the most competitive submissions (by five teams), we find that, although low-level “pattern matching” approaches provide the best performance for “exact” matches, “symbolic” approaches working on higher-level representations seem to perform better in more complex settings, such as matching morphological variants. Finally, optimizing the output scores for Cnxe seems to generate systems that are more robust to differences in the operating point and that also perform well in terms of TWV, whereas the opposite might not be always true.

This paper presents an alternative feature set to the traditional MFCC-SDC used in acoustic approaches to Spoken Language Recognition: the log-likelihood ratios of phone posterior probabilities, hereafter Phone Log-Likelihood Ratios (PLLR), produced by a phone recognizer. In this work, an iVector system trained on this set of features (plus dynamic coefficients) is evaluated and compared to (1) an acoustic iVector system (trained on the MFCC-SDC feature set) and (2) a phonotactic (Phone-lattice-SVM) system, using two different benchmarks: the NIST 2007 and 2009 LRE datasets. iVector systems trained on PLLR features proved to be competitive, reaching or even outperforming the MFCC-SDC-based iVector and the phonotactic systems. The fusion of the proposed approach with the acoustic and phonotactic systems provided even more significant improvements, outperforming state-of-the-art systems on both benchmarks.

This paper evaluates the performance of the twelve primary systems submitted to the evaluation on speaker verification in the context of a mobile environment using the MOBIO database. The mobile environment provides a challenging and realistic test-bed for current state-of-the-art speaker verification techniques. Results in terms of equal error rate (EER), half total error rate (HTER) and detection error trade-off (DET) confirm that the best performing systems are based on total variability modeling, and are the fusion of several sub-systems. Nevertheless, the good old UBM-GMM based systems are still competitive. The results also show that the use of additional data for training as well as gender-dependent features can be helpful.

The Mel-frequency cepstral coefficients (MFCC) and their derivatives are commonly used as acoustic features for speaker recognition. The issue arises of whether some of those features are redundant or dependent on other features. Probably, not all of them are equally relevant for speaker recognition. Reduced feature sets allow more robust estimates of the model parameters. Also, less computational resources are required, which is crucial for real-time speaker recognition applications using low-resource devices. In this paper, we use feature weighting as an intermediate step towards feature selection. Genetic algorithms are used to find the optimal set of weights for a 38-dimensional feature set, consisting of 12 MFCC, their first and second derivatives, energy and its first derivative. To evaluate each set of weights, speaker recognition errors are counted over a validation dataset. Speaker models are based on empirical distributions of acoustic labels, obtained through vector quantization. On average, weighting acoustic features yields between 15% and 25% error reduction in speaker recognition tests. Finally, features are sorted according to their weights, and the K features with greatest average ranks are retained and evaluated. We conclude that combining feature weighting and feature selection allows to reduce costs without degrading performance

Previous works in English have revealed that disfluencies follow regular patterns and that incorporating them into the language model of a speech recognizer leads to lower perplexities and sometimes to a better performance. Although work on disfluency modeling has been applied outside the English community (e.g., in Japanese), as far as we know there is no specific work dealing with disfluencies in Spanish. In this paper, we follow a data driven approach in exploring the potential benefit of modeling disfluencies in a speech recognizer in Spanish. Two databases of human-computer and human-human dialogs are considered, which allow the absolute and relative frequencies of disfluencies in the two situations to be compared. The rate of disfluencies in human-human dialogs is found to be very close to that found for similar databases in English. Due to setup factors, the rate of disfluencies found in human-computer dialogs was remarkably higher than that reported for similar databases in English. In any case, from the point of view of speech recognition, the high frequencies of disfluencies and the distinct features of the acoustic events related to them support the need for explicit acoustic models. The regularities observed in the distribution of filled pauses and speech repairs reveal that including them in the language model of the speech recognizer may be also helpful. The extent to which the number of events depends on utterance length and on the speaker is also explored. Statistics are shown that follow previous studies for English, and a sizeable space is devoted to comparing our results with them. Finally, various possible cues for the automatic detection of speech repairs—a key issue from the point of view of speech understanding—are explored: silent pauses, filled pauses, lengthenings, cut off words and discourse markers. As previously observed for English, none of them was found to be reliable by itself. More information, especially at the acoustic-prosodic level, is no doubt needed to reliably detect speech repairs.