Search

This paper considers speech and listening enhancement for signals captured in one noisy environment that must be played back to a listener in another noisy environment. In both far-end speech enhancement and near-end listening enhancement, overly prioritizing noise suppression or maximizing intelligibility can result in undue speech distortions and reduced quality, especially when intelligibility is already high in favorable noise conditions. To address this, the use of a minimum processing framework has been proposed with the aim of reducing noise or enhancing listening to a minimum degree while ensuring that a specified intelligibility level is maintained. Furthermore, results have shown that jointly considering both environments improves performance compared to blindly concatenating far- and near-end methods. In blind processing, near-end listening enhancement typically assumes that the far-end signal is devoid of noise, potentially leading to erroneously interpreting noise as speech. Additionally, if the transmitter and receiver are blind to each other’s presence, multiple instances of far- and near-end enhancement may occur and possibly work opposite directions, thus leading to degradations in the enhancement performance. In this paper, we perform a comprehensive exploration of our previously proposed joint far- and near-end minimum processing framework with systematic analysis and discussion. We derive a closed-form solution to the joint far- and near-end minimum processing optimization problem, with mean-square error processing penalty, a speech intelligibility constraint based on the approximated speech intelligibility index, and a noise power constraint. Performance was systematically studied using objective measures and listening tests for intelligibility, listening effort, and quality. We compared against relevant joint and blind methods with minimum and maximum processing. The results suggest that minimum processing achieves intelligibility comparable to maximum processing while preserving quality in higher signal-to-noise ratios, indicating its benefits in end-to-end communication. Joint processing provides advantages in objective estimated speech intelligibility for the minimum processing case, but not for maximum processing. However, no significant differences were observed in listening test results. This suggests that in certain speech and listening scenarios, it is feasible to optimize near- and far-end aspects separately, offering a more practical and convenient approach compared to joint optimization.

Supervised learning-based speech enhancement methods often work remarkably well in acoustic situations represented in the training corpus but generalize poorly to out-of-domain situations, i.e. situations not seen during training. This stands in the way of further improvement of these methods in realistic scenarios, as collecting paired noisy-clean recordings in the target application domain is typically not feasible. Recording noisy-only in-domain data is, though, much more practical. In this article, we tackle the problem of unsupervised domain adaptation in speech enhancement. Specifically, we propose a way to use in-domain noisy-only data in the training of a neural network to improve upon a model trained solely on out-of-domain paired data. For this, we make use of masked spectrogram prediction, a technique from self-supervised learning that aims to interpolate masked regions of a spectrogram. We hypothesize that masked spectrogram prediction encourages learning of features that represent well both speech and noise components of the noisy signals. These features can then be used to train a more robust speech enhancement system. We evaluate the proposed method on the VoiceBank-DEMAND and LibriFSD50k databases, with WSJ0-CHiME3 serving as the out-of-domain database. We show that the proposed method outperforms both the out-of-domain system and the baseline approaches, i.e. RemixIT and noisy-target training, and also combines well with the previously proposed RemixIT method.

Abstract Aims Heart failure (HF) is an increasing concern worldwide. A rising HF burden is expected due to the prospected future demographic changes with aging populations. Consequently, the long‐term follow‐up and treatment will be performed increasingly by primary care physicians in the future. Contemporary data on HF patients in primary care are needed to plan and ensure an effective and safe follow‐up of future patients. Methods and results The electronic patient journals of 148 primary care clinics in Denmark were searched in a standardized manner to identify patients with HF [code K77 of the International Classification of Primary Care, Second Edition]. Prespecified variables including demographic information, clinical variables, co‐morbidities, prescribed medications, and setting of follow‐up were recorded. In total, 1111 patients were included in the study. The mean timepoint for the HF diagnosis was August 2018. In 95% of cases, the diagnosis of HF was made in a specialized setting. The echocardiogram data used for phenotyping were available in 1042 (94%) of the 1111 patients. HF with reduced ejection fraction (HFrEF) was present in 43%, recovered HFrEF in 31%, and HF with mildly reduced (HFmrEF) or preserved ejection fraction (HFpEF) in 26%. In patients with HFrEF or recovered HFrEF, fundamental treatments were prescribed in 86% for angiotensin converting enzyme inhibitor (ACEI), angiotensin receptor blocker (ARB), or angiotensin receptor neprilysin inhibitor (ARNI), in 82% for beta‐blocker, in 38% for mineralocorticoid receptor antagonist (MRA), and in 12% for sodium‐glucose co‐transporter‐2 inhibitor (SGLT2i). Older patients were treated to a significantly lesser extent than young patients for all drug classes [odds ratio (OR) point estimates 0.50 to 0.69, all P‐values

Study objective: The objective was to assess the effect of ongoing angiotensin receptor-neprilysin inhibitor(ARNI) on the effect of the sodium glucose cotransporter 2 (SGLT2) inhibitor empagliflozin on left ventricular (LV) size and function in patients with heart failure and reduced ejection fraction(HFrEF). Design: Post hoc analysis of the Empire HF trial, an investigator-initiated, double-blind, randomized controlled trial. Participants: 190 patients with HFrEF with New York Heart association class I-III symptoms with an ejection fraction of 40 % or below. Patients were stratified according to ongoing ARNI treatment at baseline. Intervention: Empagliflozin 10 mg daily or placebo for 12 weeks. Echocardiography at baseline and follow-up. Main outcome measures: Left ventricular end-systolic volume index (LVESVI), end-diastolic volume index (LVEDVI), left atrial volume index (LAVI), left ventricular ejection fraction (LVEF). Results: A total of 58 patients (31 %) received ARNI at baseline. Compared to with placebo, empagliflozin reduced the LVESVI ([−6.2 (−14.1 to 1.6); p = 0.12] and [−3.3 (−8.2 to 1.6); p = 0.19], interaction P = 0.49), LVEDVI ([−11.2 (−21.2 to −1.2); p = 0.03] and [−2.9 (−8.7 to 2.9); p = 0.32], interaction P = 0.13), and LAVI ([−3.9 (−9.1 to 1.2); p = 0.14] and. [−1.8 (−4.4 to 0.7); p = 0.16], respectively, interaction P = 0.9) in patients treated with and without ARNI at baseline, respectively. No treatment-by-ARNI subgroup interaction were found. Unaffected by baseline ARNI treatment, empagliflozin did not improve LVEF. Conclusion: The effect of empagliflozin on cardiac structure and function compared to placebo was not affected by background treatment with ARNI.

Abstract Background Plasma growth differentiation factor-15 (GDF-15) biomarker levels increase in response to inflammation and tissue injury, and increased levels of GDF-15 are associated with increased risk of mortality in patients with heart failure with reduced ejection fraction (HFrEF). Sodium-glucose cotransporter-2 (SGLT2) inhibitors, which improve outcome in HFrEF, have been shown to increase plasma GDF-15 in diabetic patients. We aimed to investigate the effect of empagliflozin on GDF-15 in HFrEF patients. Methods This Empire HF Biomarker substudy was from the multicentre, randomized, double-blind, placebo-controlled Empire HF trial that included 190 patients from June 29, 2017, to September 10, 2019. Stable ambulatory HFrEF patients with ejection fraction of ≤ 40% were randomly assigned (1:1) to empagliflozin 10 mg once daily, or matching placebo for 12 weeks. Changes from baseline to 12 weeks in plasma levels of GDF-15, high-sensitive C-reactive protein (hsCRP), and high-sensitive troponin T (hsTNT) were assessed. Results A total of 187 patients who were included in this study, mean age was 64 ± 11 years; 85% male, 12% with type 2 diabetes, mean ejection fraction 29 ± 8, with no differences between the groups. Baseline median plasma GDF-15 was 1189 (918–1720) pg/mL with empagliflozin, and 1299 (952–1823) pg/mL for placebo. Empagliflozin increased plasma GDF-15 compared to placebo (adjusted between-groups treatment effect; ratio of change (1·09 [95% confidence interval (CI), 1.03–1.15]: p = 0.0040). The increase in plasma GDF15 was inversely associated with a decrease in left ventricular end-systolic (R = – 0.23, p = 0.031), and end-diastolic volume (R = – 0.29, p = 0.0066). There was no change in plasma hsCRP (1.09 [95%CI, 0.86–1.38]: p = 0.48) or plasma hsTNT (1.07 [95%CI, 0.97–1.19]: p = 0.18) compared to placebo. Patients with diabetes and treated with metformin demonstrated no increase in plasma GDF-15 with empagliflozin, p for interaction = 0·01. Conclusion Empagliflozin increased plasma levels of GDF-15 in patients with HFrEF, with no concomitant increase in hsTNT nor hsCRP. Trial registration: The Empire HF trial is registered with ClinicalTrials.gov, NCT03198585.

Prediction of Speech Intelligibility (SI) is a topic of interest for most speech processing applications, where intelligibility is of any importance, e.g., speech coding, transmission and enhancement. Traditionally, SI predictors have been based on signal processing methods and heuristics, but more recently, an increasing number of data-driven SI-predictors have been proposed. Data-driven prediction of SI requires large quantities of labelled data, ideally from many listening tests. Listening tests differ in factors such as vocabulary, talker, listener’s task, etc. collectively referred to as the paradigm. A naïve strategy of training SI-predictors directly on stimuli, pooled from different listening tests, is futile because the exact map from the stimulus to SI is determined, not only by the stimulus, but also by the paradigm. Data-driven SI-predictors trained in this way become specialized to the paradigms of the training data by erroneously attributing all paradigm influences on SI to the stimulus. The problem is fundamental and persists even in the idealized situation where training data is abundant. We propose a strategy for training data-driven SI-predictors that is independent of the paradigms, underlying the training data. The proposed strategy is to concatenate an SI-predictor and a layer of trainable dataset-specific mapping functions, each corresponding to a single paradigm in the training data. These mapping functions are trained jointly with the SI-predictor and serve to efficiently approximate the psychometric functions implied by each paradigm. The mapping functions prevent the predictor from specializing to these paradigms during training. We present an SI-predictor with a novel architecture that incorporates a convolutional network and an ESTOI back-end, train it with this strategy, compare it to naïve training and a range of existing non-data-driven predictors. The proposed training strategy and architecture results in higher performance overall and increased robustness to unseen paradigms.

Spoken keyword spotting (KWS) deals with the identification of keywords in audio streams and has become a fast-growing technology thanks to the paradigm shift introduced by deep learning a few years ago. This has allowed the rapid embedding of deep KWS in a myriad of small electronic devices with different purposes like the activation of voice assistants. Prospects suggest a sustained growth in terms of social use of this technology. Thus, it is not surprising that deep KWS has become a hot research topic among speech scientists, who constantly look for KWS performance improvement and computational complexity reduction. This context motivates this paper, in which we conduct a literature review into deep spoken KWS to assist practitioners and researchers who are interested in this technology. Specifically, this overview has a comprehensive nature by covering a thorough analysis of deep KWS systems (which includes speech features, acoustic modeling and posterior handling), robustness methods, applications, datasets, evaluation metrics, performance of deep KWS systems and audio-visual KWS. The analysis performed in this paper allows us to identify a number of directions for future research, including directions adopted from automatic speech recognition research and directions that are unique to the problem of spoken KWS.

In this paper, we propose a novel speech enhancement paradigm which can effectively solve the problem of retrieving a desired speech signal in a multi-talker environment. The proposed speech enhancement paradigm involves a three-step procedure consisting of separation, ranking, and enhancement. First, a speech separation system – which could be a conventional spatial filter bank or more advanced separation systems – separates mixtures of speech signals captured by microphones into speech signals from candidate speakers. Next, novel ranking algorithms – proposed in this paper – are applied to determine the talker-of-interest amongst the separated speech signals. Finally, the speech signal of the talker-of-interest is estimated as a linear combination of the separated signals, whose weights are determined by the ranking algorithms. We propose ranking algorithms, which exploit turn-taking patterns between conversational partners in order to determine the talker-of-interest amongst competing speakers. Unlike some existing solutions, our ranking algorithms do not require access to additional sensors, e.g., EEG electrodes, cameras, etc., but only rely on microphone signals. Specifically, the proposed algorithms rank the separated speech signals based on the probability of speech overlaps and gaps with the user’s own voice. The speech signal with highest ranking is the talker with minimum probability of speech overlap and gap with the user’s own voice. The proposed ranking algorithms are shown highly effective at determining the talker-of-interest, since conversational partners, i.e., the user and the talker-of-interest, behaviorally avoid speech overlaps and gaps. We evaluate the proposed speech enhancement paradigm in two practical hearing aid related applications, where the objective is to enhance a speech signal of a conversational partner in a multi-talker environment. The results of the evaluation demonstrate that the proposed speech enhancement systems in both applications significantly outperform conventional speech enhancement systems.

Abstract Background Mid-regional pro-atrial natriuretic peptide (MR-proANP) is a useful biomarker in outpatients with type 2 diabetes (T2D) to diagnose heart failure (HF). Elevated B-type natriuretic peptides are included in the definition of HF with preserved ejection fraction (HFpEF) but little is known about the prognostic value of including A-type natriuretic peptides (MR-proANP) in the evaluation of patients with T2D. Methods We prospectively evaluated the risk of incident cardiovascular (CV) events in outpatients with T2D (n = 806, mean ± standard deviation age 64 ± 10 years, 65% male, median [interquartile range] duration of diabetes 12 [6–17] years, 17.5% with symptomatic HFpEF) according to MR-proANP levels and stratified according to HF-status including further stratification according to a prespecified cut-off level of MR-proANP. Results A total of 126 CV events occurred (median follow-up 4.8 [4.1–5.3] years). An elevated MR-proANP, with a cut-off of 60 pmol/l or as a continuous variable, was associated with incident CV events (p

Abstract Background Data from recent cardiovascular outcome trials in patients with type 2 diabetes (T2D) suggest that sodium-glucose cotransporter 2 (SGLT2) inhibitors can prevent development of heart failure (HF) and prolong life in patients without HF. Ongoing event-driven trials are investigating whether the same effect is present in patients with well-defined HF. The mechanism behind the effect of SGLT2 inhibitors in patients with T2D and the potential effect in patients with overt HF is presently unknown. Methods This is a randomized, double-blinded, placebo-controlled, parallel group, clinical trial including HF patients with reduced left ventricular ejection fraction (HFrEF) with an ejection fraction ≤ 40% on optimal therapy recruited from specialized HF clinics in Denmark. The primary aim is to investigate the effect of the SGLT2 inhibitor empagliflozin on N-terminal pro-brain natriuretic peptide (NT-proBNP). Secondary endpoints include cardiac biomarkers, function and hemodynamics, metabolic and renal parameters, daily activity level, and quality of life. Patients are assigned 1:1 to 90 days treatment with empagliflozin 10 mg daily or placebo. Patients with T2D are required to be on recommended doses of anti-glycemic therapy with a hemoglobin A1c (HbA1c) of 6.5–10.0% (48–86 mmol/mol). To show a between-group difference in the change of NT-proBNP of 30%, a total of 189 patients will be included. Discussion The Empire HF trial will elucidate the effects and modes of action of empagliflozin in HFrEF patients with and without T2D and provide important mechanistic data which will complement ongoing event-driven trials. Trial registration Clinicaltrialsregister.eu, EudraCT Number 2017-001341-27. Registered on 29 May 2017. ClinicalTrials.gov, NCT03198585. Registered on 26 June 2017.

Although most noise reduction algorithms are critically dependent on the noise power spectral density (PSD), most procedures for noise PSD estimation fail to obtain good estimates in nonstationary noise conditions. Recently, a DFT-subspace-based method was proposed which improves noise PSD estimation under these conditions. However, this approach is based on eigenvalue decompositions per DFT bin, and might be too computationally demanding for low-complexity applications like hearing aids. In this paper we present a noise tracking method with low complexity, but approximately similar noise tracking performance as the DFT-subspace approach. The presented method uses a periodogram with resolution that is higher than the spectral resolution used in the noise reduction algorithm itself. This increased resolution enables estimation of the noise PSD even when speech energy is present at the time-frequency point under consideration. This holds in particular for voiced type of speech sounds which can be modelled using a small number of complex exponentials.