The growing sophistication of speech generated by Artificial Intelligence (AI) has introduced new challenges in audio deepfake detection. Text-to-speech (TTS) and voice conversion (VC) technologies can now produce convincing synthetic speech with high quality and intelligibility. This poses a serious threat to voice biometric security systems, such as automatic speaker recognition. It also increases the risks associated to the spread of spoken disinformation, where synthetic voices can be used to disseminate malicious content. In this study, we conduct an analysis of real and synthetic speech at phonetic and word levels. For that, a parallel dataset comprising real and synthetic speech signals were developed based on a subset of the LibriSpeech ASR corpus. Synthetic speech samples were generated using two TTS and one VC systems: Coqui TTS, VITS TTS, and StarGANv2 VC. We adopted HuBERT, a self-supervised speech model, to extract speech embeddings. The motivation for using this model stems from its ability to recognize sound units corresponding to the so-called pseudo phonemes. Our analysis is based on the KL divergence (KLD) between the distributions of synthetic and real phonemes, which allowed us to rank synthetic phonemes based on their alignment with their real counterpart. We also trained several classifiers per phoneme to distinguish between real and synthetic samples. We then compute the correlations between KLD and accuracies per phoneme. Besides showing a list of phonemes that are more discriminative, our findings suggest that vowels correlate better with the classifiers’ performance, suggesting that the KLD can be an indicator of the most distinguishable phonemes for deepfake detection.

The coronavirus disease 2019 (COVID-19) pandemic has drastically impacted life around the globe. As life returns to pre-pandemic routines, COVID-19 testing has become a key component, assuring that travellers and citizens are free from the disease. Conventional tests can be expensive, time-consuming (results can take up to 48h), and require laboratory testing. Rapid antigen testing, in turn, can generate results within 15-30 minutes and can be done at home, but research shows they achieve very poor sensitivity rates. In this paper, we propose an alternative test based on speech signals recorded at home with a portable device. It has been well-documented that the virus affects many of the speech production systems (e.g., lungs, larynx, and articulators). As such, we propose the use of new modulation spectral features and linear prediction analysis to characterize these changes and design a two-stage COVID-19 prediction system by fusing the proposed features. Experiments with three COVID-19 speech datasets (CSS, DiCOVA2, and Cambridge subset) show that the two-stage feature fusion system outperforms the benchmark systems of CSS and Cambridge datasets while maintaining lower complexity compared to DL-based systems. Furthermore, the two-stage system demonstrates higher generalizability to unseen conditions in a cross-dataset testing evaluation scheme. The generalizability and interpretability of our proposed system demonstrate the potential for accessible, low-cost, at-home COVID-19 testing.

Automatic speech emotion recognition (SER) has gained popularity over the last decade and numerous Challenges have emerged. While the latest Challenges have shown that deep neural networks achieve the best results, existing input features are still a bottleneck and cause severe performance degradation in realistic “in-the-wild” scenarios. In this paper, we propose two innovations to tackle this issue. First, we propose to combine the bag-of-audio-words methodology with modulation spectrum features for environmental robustness. Second, we take advantage of the inherent quality-awareness properties of modulation spectrum and propose the use of a quality feature as an additional feature to be used by the speech emotion recognizer. Experiments are conducted with three multi-lingual speech datasets used in recent SER Challenges degraded by different noise sources and levels, and room reverberation. Experimental results show the proposed features i) consistently outperforming benchmark systems, ii) providing complementary information to classical features, hence improving performance with feature fusion, and iii) showing robustness against environment and language mismatch. Moreover, we show that when the proposed system is provided with quality information, further improvements are obtained. Overall, the proposed bag of modulation spectrum features are shown to be a promising candidate for “in-the-wild” SER.