MOS-Bench: Benchmarking Generalization Abilities of Subjective Speech Quality Assessment Models

Subjective speech quality assessment (SSQA) is critical for evaluating speech samples as perceived by human listeners. While model-based SSQA has enjoyed great success thanks to the development of deep neural networks (DNNs), generalization remains a key challenge, especially for unseen, out-of-domain data. To benchmark the generalization abilities of SSQA models, we present MOS-Bench, a diverse collection of datasets. In addition, we also introduce SHEET, an open-source toolkit containing complete recipes to conduct SSQA experiments. We provided benchmark results for MOS-Bench, and we also explored multi-dataset training to enhance generalization. Additionally, we proposed a new performance metric, best score difference/ratio, and used latent space visualizations to explain model behavior, offering valuable insights for future research.

Two-stage Framework for Robust Speech Emotion Recognition Using Target Speaker Extraction in Human Speech Noise Conditions

Developing a robust speech emotion recognition (SER) system in noisy conditions faces challenges posed by different noise properties. Most previous studies have not considered the impact of human speech noise, thus limiting the application scope of SER. In this paper, we propose a novel two-stage framework for the problem by cascading target speaker extraction (TSE) method and SER. We first train a TSE model to extract the speech of target speaker from a mixture. Then, in the second stage, we utilize the extracted speech for SER training. Additionally, we explore a joint training of TSE and SER models in the second stage. Our developed system achieves a 14.33% improvement in unweighted accuracy (UA) compared to a baseline without using TSE method, demonstrating the effectiveness of our framework in mitigating the impact of human speech noise. Moreover, we conduct experiments considering speaker gender, showing that our framework performs particularly well in different-gender mixture.

Improved Architecture for High-resolution Piano Transcription to Efficiently Capture Acoustic Characteristics of Music Signals

Automatic music transcription (AMT), aiming to convert musical signals into musical notation, is one of the important tasks in music information retrieval. Recently, previous works have applied high-resolution labels, i.e., the continuous onset and offset times of piano notes, as training targets, achieving substantial improvements in transcription performance. However, there still remain some issues to be addressed, e.g., the harmonics of notes are sometimes recognized as false positive notes, and the size of AMT model tends to be larger to improve the transcription performance. To address these issues, we propose an improved high-resolution piano transcription model to well capture specific acoustic characteristics of music signals. First, we employ the Constant-Q Transform as the input representation to better adapt to musical signals. Moreover, we have designed two architectures: the first is based on a convolutional recurrent neural network (CRNN) with dilated convolution, and the second is an encoder-decoder architecture that combines CRNN with a non-autoregressive Transformer decoder. We conduct systematic experiments for our models. Compared to the high-resolution AMT system used as a baseline, our models effectively achieve 1) consistent improvement in note-level metrics, and 2) the significant smaller model size, which shed lights on future work.

Improvements of Discriminative Feature Space Training for Anomalous Sound Detection in Unlabeled Conditions

In anomalous sound detection, the discriminative method has demonstrated superior performance. This approach constructs a discriminative feature space through the classification of the meta-information labels for normal sounds. This feature space reflects the differences in machine sounds and effectively captures anomalous sounds. However, its performance significantly degrades when the meta-information labels are missing. In this paper, we improve the performance of a discriminative method under unlabeled conditions by two approaches. First, we enhance the feature extractor to perform better under unlabeled conditions. Our enhanced feature extractor utilizes multi-resolution spectrograms with a new training strategy. Second, we propose various pseudo-labeling methods to effectively train the feature extractor. The experimental evaluations show that the proposed feature extractor and pseudo-labeling methods significantly improve performance under unlabeled conditions.

The VoiceMOS Challenge 2024: Beyond Speech Quality Prediction

We present the third edition of the VoiceMOS Challenge, a scientific initiative designed to advance research into automatic prediction of human speech ratings. There were three tracks. The first track was on predicting the quality of ``zoomed-in'' high-quality samples from speech synthesis systems. The second track was to predict ratings of samples from singing voice synthesis and voice conversion with a large variety of systems, listeners, and languages. The third track was semi-supervised quality prediction for noisy, clean, and enhanced speech, where a very small amount of labeled training data was provided. Among the eight teams from both academia and industry, we found that many were able to outperform the baseline systems. Successful techniques included retrieval-based methods and the use of non-self-supervised representations like spectrograms and pitch histograms. These results showed that the challenge has advanced the field of subjective speech rating prediction.

SVDD 2024: The Inaugural Singing Voice Deepfake Detection Challenge

With the advancements in singing voice generation and the growing presence of AI singers on media platforms, the inaugural Singing Voice Deepfake Detection (SVDD) Challenge aims to advance research in identifying AI-generated singing voices from authentic singers. This challenge features two tracks: a controlled setting track (CtrSVDD) and an in-the-wild scenario track (WildSVDD). The CtrSVDD track utilizes publicly available singing vocal data to generate deepfakes using state-of-the-art singing voice synthesis and conversion systems. Meanwhile, the WildSVDD track expands upon the existing SingFake dataset, which includes data sourced from popular user-generated content websites. For the CtrSVDD track, we received submissions from 47 teams, with 37 surpassing our baselines and the top team achieving a 1.65% equal error rate. For the WildSVDD track, we benchmarked the baselines. This paper reviews these results, discusses key findings, and outlines future directions for SVDD research.

Quantifying the effect of speech pathology on automatic and human speaker verification

This study investigates how surgical intervention for speech pathology (specifically, as a result of oral cancer surgery) impacts the performance of an automatic speaker verification (ASV) system. Using two recently collected Dutch datasets with parallel pre and post-surgery audio from the same speaker, NKI-OC-VC and SPOKE, we assess the extent to which speech pathology influences ASV performance, and whether objective/subjective measures of speech severity are correlated with the performance. Finally, we carry out a perceptual study to compare judgements of ASV and human listeners. Our findings reveal that pathological speech negatively affects ASV performance, and the severity of the speech is negatively correlated with the performance. There is a moderate agreement in perceptual and objective scores of speaker similarity and severity, however, we could not clearly establish in the perceptual study, whether the same phenomenon also exists in human perception.

2DP-2MRC: 2-Dimensional Pointer-based Machine Reading Comprehension Method for Multimodal Moment Retrieval

Moment retrieval aims to locate the most relevant moment in an untrimmed video based on a given natural language query. Existing solutions can be roughly categorized into moment-based and clip-based methods. The former often involves heavy computations, while the latter, due to overlooking coarse-grained information, typically underperforms compared to moment-based models. Hence, this paper proposes a novel 2-Dimensional Pointer-based Machine Reading Comprehension for Moment Retrieval Choice (2DP-2MRC) model to address the issue of imprecise localization in clip-based methods while maintaining lower computational complexity than moment-based methods. Specifically, we introduce an AV-Encoder to capture coarse-grained information at moment and video levels. Additionally, a 2D pointer encoder module is introduced to further enhance boundary detection for target moment. Extensive experiments on the HiREST dataset demonstrate that 2DP-2MRC significantly outperforms existing baseline models.

CtrSVDD: A Benchmark Dataset and Baseline Analysis for Controlled Singing Voice Deepfake Detection

Recent singing voice synthesis and conversion advancements necessitate robust singing voice deepfake detection (SVDD) models. Current SVDD datasets face challenges due to limited controllability, diversity in deepfake methods, and licensing restrictions. Addressing these gaps, we introduce CtrSVDD, a large-scale, diverse collection of bonafide and deepfake singing vocals. These vocals are synthesized using state-of-the-art methods from publicly accessible singing voice datasets. CtrSVDD includes 47.64 hours of bonafide and 260.34 hours of deepfake singing vocals, spanning 14 deepfake methods and involving 164 singer identities. We also present a baseline system with flexible front-end features, evaluated against a structured train/dev/eval split. The experiments show the importance of feature selection and highlight a need for generalization towards deepfake methods that deviate further from training distribution. The CtrSVDD dataset and baselines are publicly accessible.

Multi-speaker Text-to-speech Training with Speaker Anonymized Data

The trend of scaling up speech generation models poses a threat of biometric information leakage of the identities of the voices in the training data, raising privacy and security concerns. In this paper, we investigate training multi-speaker text-to-speech (TTS) models using data that underwent speaker anonymization (SA), a process that tends to hide the speaker identity of the input speech while maintaining other attributes. Two signal processing-based and three deep neural network-based SA methods were used to anonymize VCTK, a multi-speaker TTS dataset, which is further used to train an end-to-end TTS model, VITS, to perform unseen speaker TTS during the testing phase. We conducted extensive objective and subjective experiments to evaluate the anonymized training data, as well as the performance of the downstream TTS model trained using those data. Importantly, we found that UTMOS, a data-driven subjective rating predictor model, and GVD, a metric that measures the gain of voice distinctiveness, are good indicators of the downstream TTS performance. We summarize insights in the hope of helping future researchers determine the goodness of the SA system for multi-speaker TTS training.