Eduardo Fonseca

Joseph Turian, Jordie Shier, Humair Raj Khan et al. · cmu

What audio embedding approach generalizes best to a wide range of downstream tasks across a variety of everyday domains without fine-tuning? The aim of the HEAR benchmark is to develop a general-purpose audio representation that provides a strong basis for learning in a wide variety of tasks and scenarios. HEAR evaluates audio representations using a benchmark suite across a variety of domains, including speech, environmental sound, and music. HEAR was launched as a NeurIPS 2021 shared challenge. In the spirit of shared exchange, each participant submitted an audio embedding model following a common API that is general-purpose, open-source, and freely available to use. Twenty-nine models by thirteen external teams were evaluated on nineteen diverse downstream tasks derived from sixteen datasets. Open evaluation code, submitted models and datasets are key contributions, enabling comprehensive and reproducible evaluation, as well as previously impossible longitudinal studies. It still remains an open question whether one single general-purpose audio representation can perform as holistically as the human ear.

Mariana-Iuliana Georgescu, Eduardo Fonseca, Radu Tudor Ionescu et al.

Can we leverage the audiovisual information already present in video to improve self-supervised representation learning? To answer this question, we study various pretraining architectures and objectives within the masked autoencoding framework, motivated by the success of similar methods in natural language and image understanding. We show that we can achieve significant improvements on audiovisual downstream classification tasks, surpassing the state-of-the-art on VGGSound and AudioSet. Furthermore, we can leverage our audiovisual pretraining scheme for multiple unimodal downstream tasks using a single audiovisual pretrained model. We additionally demonstrate the transferability of our representations, achieving state-of-the-art audiovisual results on Epic Kitchens without pretraining specifically for this dataset.

R. Channing Moore, Daniel P. W. Ellis, Eduardo Fonseca et al.

Machine learning from training data with a skewed distribution of examples per class can lead to models that favor performance on common classes at the expense of performance on rare ones. AudioSet has a very wide range of priors over its 527 sound event classes. Classification performance on AudioSet is usually evaluated by a simple average over per-class metrics, meaning that performance on rare classes is equal in importance to the performance on common ones. Several recent papers have used dataset balancing techniques to improve performance on AudioSet. We find, however, that while balancing improves performance on the public AudioSet evaluation data it simultaneously hurts performance on an unpublished evaluation set collected under the same conditions. By varying the degree of balancing, we show that its benefits are fragile and depend on the evaluation set. We also do not find evidence indicating that balancing improves rare class performance relative to common classes. We therefore caution against blind application of balancing, as well as against paying too much attention to small improvements on a public evaluation set.

Scott Wisdom, Hakan Erdogan, Daniel Ellis et al.

We introduce the Free Universal Sound Separation (FUSS) dataset, a new corpus for experiments in separating mixtures of an unknown number of sounds from an open domain of sound types. The dataset consists of 23 hours of single-source audio data drawn from 357 classes, which are used to create mixtures of one to four sources. To simulate reverberation, an acoustic room simulator is used to generate impulse responses of box shaped rooms with frequency-dependent reflective walls. Additional open-source data augmentation tools are also provided to produce new mixtures with different combinations of sources and room simulations. Finally, we introduce an open-source baseline separation model, based on an improved time-domain convolutional network (TDCN++), that can separate a variable number of sources in a mixture. This model achieves 9.8 dB of scale-invariant signal-to-noise ratio improvement (SI-SNRi) on mixtures with two to four sources, while reconstructing single-source inputs with 35.5 dB absolute SI-SNR. We hope this dataset will lower the barrier to new research and allow for fast iteration and application of novel techniques from other machine learning domains to the sound separation challenge.

Daniel P. W. Ellis, Eduardo Fonseca, Ron J. Weiss et al. · deepmind

Editing complex real-world sound scenes is difficult because individual sound sources overlap in time. Generative models can fill-in missing or corrupted details based on their strong prior understanding of the data domain. We present a system for editing individual sound events within complex scenes able to delete, insert, and enhance individual sound events based on textual edit descriptions (e.g., ``enhance Door'') and a graphical representation of the event timing derived from an ``event roll'' transcription. We present an encoder-decoder transformer working on SoundStream representations, trained on synthetic (input, desired output) audio example pairs formed by adding isolated sound events to dense, real-world backgrounds. Evaluation reveals the importance of each part of the edit descriptions -- action, class, timing. Our work demonstrates ``recomposition'' is an important and practical application.

Eduardo Fonseca, Andres Ferraro, Xavier Serra

Recent studies have put into question the commonly assumed shift invariance property of convolutional networks, showing that small shifts in the input can affect the output predictions substantially. In this paper, we analyze the benefits of addressing lack of shift invariance in CNN-based sound event classification. Specifically, we evaluate two pooling methods to improve shift invariance in CNNs, based on low-pass filtering and adaptive sampling of incoming feature maps. These methods are implemented via small architectural modifications inserted into the pooling layers of CNNs. We evaluate the effect of these architectural changes on the FSD50K dataset using models of different capacity and in presence of strong regularization. We show that these modifications consistently improve sound event classification in all cases considered. We also demonstrate empirically that the proposed pooling methods increase shift invariance in the network, making it more robust against time/frequency shifts in input spectrograms. This is achieved by adding a negligible amount of trainable parameters, which makes these methods an appealing alternative to conventional pooling layers. The outcome is a new state-of-the-art mAP of 0.541 on the FSD50K classification benchmark.

Shawn Hershey, Daniel P W Ellis, Eduardo Fonseca et al.

To reveal the importance of temporal precision in ground truth audio event labels, we collected precise (~0.1 sec resolution) "strong" labels for a portion of the AudioSet dataset. We devised a temporally strong evaluation set (including explicit negatives of varying difficulty) and a small strong-labeled training subset of 67k clips (compared to the original dataset's 1.8M clips labeled at 10 sec resolution). We show that fine-tuning with a mix of weak and strongly labeled data can substantially improve classifier performance, even when evaluated using only the original weak labels. For a ResNet50 architecture, d' on the strong evaluation data including explicit negatives improves from 1.13 to 1.41. The new labels are available as an update to AudioSet.

Eduardo Fonseca, Aren Jansen, Daniel P. W. Ellis et al.

Real-world sound scenes consist of time-varying collections of sound sources, each generating characteristic sound events that are mixed together in audio recordings. The association of these constituent sound events with their mixture and each other is semantically constrained: the sound scene contains the union of source classes and not all classes naturally co-occur. With this motivation, this paper explores the use of unsupervised automatic sound separation to decompose unlabeled sound scenes into multiple semantically-linked views for use in self-supervised contrastive learning. We find that learning to associate input mixtures with their automatically separated outputs yields stronger representations than past approaches that use the mixtures alone. Further, we discover that optimal source separation is not required for successful contrastive learning by demonstrating that a range of separation system convergence states all lead to useful and often complementary example transformations. Our best system incorporates these unsupervised separation models into a single augmentation front-end and jointly optimizes similarity maximization and coincidence prediction objectives across the views. The result is an unsupervised audio representation that rivals state-of-the-art alternatives on the established shallow AudioSet classification benchmark.

Eduardo Fonseca, Diego Ortego, Kevin McGuinness et al.

Self-supervised representation learning can mitigate the limitations in recognition tasks with few manually labeled data but abundant unlabeled data---a common scenario in sound event research. In this work, we explore unsupervised contrastive learning as a way to learn sound event representations. To this end, we propose to use the pretext task of contrasting differently augmented views of sound events. The views are computed primarily via mixing of training examples with unrelated backgrounds, followed by other data augmentations. We analyze the main components of our method via ablation experiments. We evaluate the learned representations using linear evaluation, and in two in-domain downstream sound event classification tasks, namely, using limited manually labeled data, and using noisy labeled data. Our results suggest that unsupervised contrastive pre-training can mitigate the impact of data scarcity and increase robustness against noisy labels, outperforming supervised baselines.

Nicolas Turpault, Romain Serizel, Scott Wisdom et al.

We propose a benchmark of state-of-the-art sound event detection systems (SED). We designed synthetic evaluation sets to focus on specific sound event detection challenges. We analyze the performance of the submissions to DCASE 2021 task 4 depending on time related modifications (time position of an event and length of clips) and we study the impact of non-target sound events and reverberation. We show that the localization in time of sound events is still a problem for SED systems. We also show that reverberation and non-target sound events are severely degrading the performance of the SED systems. In the latter case, sound separation seems like a promising solution.

Eduardo Fonseca, Xavier Favory, Jordi Pons et al.

Most existing datasets for sound event recognition (SER) are relatively small and/or domain-specific, with the exception of AudioSet, based on over 2M tracks from YouTube videos and encompassing over 500 sound classes. However, AudioSet is not an open dataset as its official release consists of pre-computed audio features. Downloading the original audio tracks can be problematic due to YouTube videos gradually disappearing and usage rights issues. To provide an alternative benchmark dataset and thus foster SER research, we introduce FSD50K, an open dataset containing over 51k audio clips totalling over 100h of audio manually labeled using 200 classes drawn from the AudioSet Ontology. The audio clips are licensed under Creative Commons licenses, making the dataset freely distributable (including waveforms). We provide a detailed description of the FSD50K creation process, tailored to the particularities of Freesound data, including challenges encountered and solutions adopted. We include a comprehensive dataset characterization along with discussion of limitations and key factors to allow its audio-informed usage. Finally, we conduct sound event classification experiments to provide baseline systems as well as insight on the main factors to consider when splitting Freesound audio data for SER. Our goal is to develop a dataset to be widely adopted by the community as a new open benchmark for SER research.

Nicolas Turpault, Scott Wisdom, Hakan Erdogan et al.

Performing sound event detection on real-world recordings often implies dealing with overlapping target sound events and non-target sounds, also referred to as interference or noise. Until now these problems were mainly tackled at the classifier level. We propose to use sound separation as a pre-processing for sound event detection. In this paper we start from a sound separation model trained on the Free Universal Sound Separation dataset and the DCASE 2020 task 4 sound event detection baseline. We explore different methods to combine separated sound sources and the original mixture within the sound event detection. Furthermore, we investigate the impact of adapting the sound separation model to the sound event detection data on both the sound separation and the sound event detection.

Eduardo Fonseca, Shawn Hershey, Manoj Plakal et al.

The study of label noise in sound event recognition has recently gained attention with the advent of larger and noisier datasets. This work addresses the problem of missing labels, one of the big weaknesses of large audio datasets, and one of the most conspicuous issues for AudioSet. We propose a simple and model-agnostic method based on a teacher-student framework with loss masking to first identify the most critical missing label candidates, and then ignore their contribution during the learning process. We find that a simple optimisation of the training label set improves recognition performance without additional computation. We discover that most of the improvement comes from ignoring a critical tiny portion of the missing labels. We also show that the damage done by missing labels is larger as the training set gets smaller, yet it can still be observed even when training with massive amounts of audio. We believe these insights can generalize to other large-scale datasets.

Eduardo Fonseca, Frederic Font, Xavier Serra

Label noise is emerging as a pressing issue in sound event classification. This arises as we move towards larger datasets that are difficult to annotate manually, but it is even more severe if datasets are collected automatically from online repositories, where labels are inferred through automated heuristics applied to the audio content or metadata. While learning from noisy labels has been an active area of research in computer vision, it has received little attention in sound event classification. Most recent computer vision approaches against label noise are relatively complex, requiring complex networks or extra data resources. In this work, we evaluate simple and efficient model-agnostic approaches to handling noisy labels when training sound event classifiers, namely label smoothing regularization, mixup and noise-robust loss functions. The main advantage of these methods is that they can be easily incorporated to existing deep learning pipelines without need for network modifications or extra resources. We report results from experiments conducted with the FSDnoisy18k dataset. We show that these simple methods can be effective in mitigating the effect of label noise, providing up to 2.5\% of accuracy boost when incorporated to two different CNNs, while requiring minimal intervention and computational overhead.

Andres Perez-Lopez, Eduardo Fonseca, Xavier Serra

This work describes and discusses an algorithm submitted to the Sound Event Localization and Detection Task of DCASE2019 Challenge. The proposed methodology relies on parametric spatial audio analysis for source localization and detection, combined with a deep learning-based monophonic event classifier. The evaluation of the proposed algorithm yields overall results comparable to the baseline system. The main highlight is a reduction of the localization error on the evaluation dataset by a factor of 2.6, compared with the baseline performance.

Eduardo Fonseca, Manoj Plakal, Frederic Font et al.

This paper introduces Task 2 of the DCASE2019 Challenge, titled "Audio tagging with noisy labels and minimal supervision". This task was hosted on the Kaggle platform as "Freesound Audio Tagging 2019". The task evaluates systems for multi-label audio tagging using a large set of noisy-labeled data, and a much smaller set of manually-labeled data, under a large vocabulary setting of 80 everyday sound classes. In addition, the proposed dataset poses an acoustic mismatch problem between the noisy train set and the test set due to the fact that they come from different web audio sources. This can correspond to a realistic scenario given by the difficulty in gathering large amounts of manually labeled data. We present the task setup, the FSDKaggle2019 dataset prepared for this scientific evaluation, and a baseline system consisting of a convolutional neural network. All these resources are freely available.

Eduardo Fonseca, Manoj Plakal, Daniel P. W. Ellis et al.

As sound event classification moves towards larger datasets, issues of label noise become inevitable. Web sites can supply large volumes of user-contributed audio and metadata, but inferring labels from this metadata introduces errors due to unreliable inputs, and limitations in the mapping. There is, however, little research into the impact of these errors. To foster the investigation of label noise in sound event classification we present FSDnoisy18k, a dataset containing 42.5 hours of audio across 20 sound classes, including a small amount of manually-labeled data and a larger quantity of real-world noisy data. We characterize the label noise empirically, and provide a CNN baseline system. Experiments suggest that training with large amounts of noisy data can outperform training with smaller amounts of carefully-labeled data. We also show that noise-robust loss functions can be effective in improving performance in presence of corrupted labels.

Xavier Favory, Eduardo Fonseca, Frederic Font et al.

Properly annotated multimedia content is crucial for supporting advances in many Information Retrieval applications. It enables, for instance, the development of automatic tools for the annotation of large and diverse multimedia collections. In the context of everyday sounds and online collections, the content to describe is very diverse and involves many different types of concepts, often organised in large hierarchical structures called taxonomies. This makes the task of manually annotating content arduous. In this paper, we present our user-centered development of two tools for the manual annotation of audio content from a wide range of types. We conducted a preliminary evaluation of functional prototypes involving real users. The goal is to evaluate them in a real context, engage in discussions with users, and inspire new ideas. A qualitative analysis was carried out including usability questionnaires and semi-structured interviews. This revealed interesting aspects to consider when developing tools for the manual annotation of audio content with labels drawn from large hierarchical taxonomies.

Eduardo Fonseca, Manoj Plakal, Frederic Font et al.

This paper describes Task 2 of the DCASE 2018 Challenge, titled "General-purpose audio tagging of Freesound content with AudioSet labels". This task was hosted on the Kaggle platform as "Freesound General-Purpose Audio Tagging Challenge". The goal of the task is to build an audio tagging system that can recognize the category of an audio clip from a subset of 41 diverse categories drawn from the AudioSet Ontology. We present the task, the dataset prepared for the competition, and a baseline system.

Eduardo Fonseca, Rong Gong, Xavier Serra

In the past, Acoustic Scene Classification systems have been based on hand crafting audio features that are input to a classifier. Nowadays, the common trend is to adopt data driven techniques, e.g., deep learning, where audio representations are learned from data. In this paper, we propose a system that consists of a simple fusion of two methods of the aforementioned types: a deep learning approach where log-scaled mel-spectrograms are input to a convolutional neural network, and a feature engineering approach, where a collection of hand-crafted features is input to a gradient boosting machine. We first show that both methods provide complementary information to some extent. Then, we use a simple late fusion strategy to combine both methods. We report classification accuracy of each method individually and the combined system on the TUT Acoustic Scenes 2017 dataset. The proposed fused system outperforms each of the individual methods and attains a classification accuracy of 72.8% on the evaluation set, improving the baseline system by 11.8%.