PyTorchKR Papers

최신 논문

총 124편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2212.06359공식 코드 · Papers with Code

Score-based Generative Modeling Secretly Minimizes the Wasserstein Distance

Score-based generative models are shown to achieve remarkable empirical performances in various applications such as image generation and audio synthesis. However, a theoretical understanding of score-based diffusion models is still incomplete. Recently, Song et al. showed that the training objective of score-based generative models is equivalent to minimizing the Kullback-Leibler divergence of the generated distribution from the data distribution. In this work, we show that score-based models also minimize the Wasserstein distance between them under suitable assumptions on the model. Specifically, we prove that the Wasserstein distance is upper bounded by the square root of the objective function up to multiplicative constants and a fixed constant offset. Our proof is based on a novel application of the theory of optimal transport, which can be of independent interest to the society. Our numerical experiments support our findings. By analyzing our upper bounds, we provide a few techniques to obtain tighter upper bounds.

arXiv:2211.08715

Conditional variational autoencoder to improve neural audio synthesis for polyphonic music sound

Deep generative models for audio synthesis have recently been significantly improved. However, the task of modeling raw-waveforms remains a difficult problem, especially for audio waveforms and music signals. Recently, the realtime audio variational autoencoder (RAVE) method was developed for high-quality audio waveform synthesis. The RAVE method is based on the variational autoencoder and utilizes the two-stage training strategy. Unfortunately, the RAVE model is limited in reproducing wide-pitch polyphonic music sound. Therefore, to enhance the reconstruction performance, we adopt the pitch activation data as an auxiliary information to the RAVE model. To handle the auxiliary information, we propose an enhanced RAVE model with a conditional variational autoencoder structure and an additional fully-connected layer. To evaluate the proposed structure, we conducted a listening experiment based on multiple stimulus tests with hidden references and an anchor (MUSHRA) with the MAESTRO. The obtained results indicate that the proposed model exhibits a more significant performance and stability improvement than the conventional RAVE model.

arXiv:2210.14661

Full-band General Audio Synthesis with Score-based Diffusion

Recent works have shown the capability of deep generative models to tackle general audio synthesis from a single label, producing a variety of impulsive, tonal, and environmental sounds. Such models operate on band-limited signals and, as a result of an autoregressive approach, they are typically conformed by pre-trained latent encoders and/or several cascaded modules. In this work, we propose a diffusion-based generative model for general audio synthesis, named DAG, which deals with full-band signals end-to-end in the waveform domain. Results show the superiority of DAG over existing label-conditioned generators in terms of both quality and diversity. More specifically, when compared to the state of the art, the band-limited and full-band versions of DAG achieve relative improvements that go up to 40 and 65%, respectively. We believe DAG is flexible enough to accommodate different conditioning schemas while providing good quality synthesis.

arXiv:2210.14874

Anisotropic multiresolution analyses for deepfake detection

Generative Adversarial Networks (GANs) have paved the path towards entirely new media generation capabilities at the forefront of image, video, and audio synthesis. However, they can also be misused and abused to fabricate elaborate lies, capable of stirring up the public debate. The threat posed by GANs has sparked the need to discern between genuine content and fabricated one. Previous studies have tackled this task by using classical machine learning techniques, such as k-nearest neighbours and eigenfaces, which unfortunately did not prove very effective. Subsequent methods have focused on leveraging on frequency decompositions, i.e., discrete cosine transform, wavelets, and wavelet packets, to preprocess the input features for classifiers. However, existing approaches only rely on isotropic transformations. We argue that, since GANs primarily utilize isotropic convolutions to generate their output, they leave clear traces, their fingerprint, in the coefficient distribution on sub-bands extracted by anisotropic transformations. We employ the fully separable wavelet transform and multiwavelets to obtain the anisotropic features to feed to standard CNN classifiers. Lastly, we find the fully separable transform capable of improving the state-of-the-art.

arXiv:2209.00130

Evaluating generative audio systems and their metrics

Recent years have seen considerable advances in audio synthesis with deep generative models. However, the state-of-the-art is very difficult to quantify; different studies often use different evaluation methodologies and different metrics when reporting results, making a direct comparison to other systems difficult if not impossible. Furthermore, the perceptual relevance and meaning of the reported metrics in most cases unknown, prohibiting any conclusive insights with respect to practical usability and audio quality. This paper presents a study that investigates state-of-the-art approaches side-by-side with (i) a set of previously proposed objective metrics for audio reconstruction, and with (ii) a listening study. The results indicate that currently used objective metrics are insufficient to describe the perceptual quality of current systems.

arXiv:2208.05314

Convergence of denoising diffusion models under the manifold hypothesis

Denoising diffusion models are a recent class of generative models exhibiting state-of-the-art performance in image and audio synthesis. Such models approximate the time-reversal of a forward noising process from a target distribution to a reference density, which is usually Gaussian. Despite their strong empirical results, the theoretical analysis of such models remains limited. In particular, all current approaches crucially assume that the target density admits a density w.r.t. the Lebesgue measure. This does not cover settings where the target distribution is supported on a lower-dimensional manifold or is given by some empirical distribution. In this paper, we bridge this gap by providing the first convergence results for diffusion models in this more general setting. In particular, we provide quantitative bounds on the Wasserstein distance of order one between the target data distribution and the generative distribution of the diffusion model.

arXiv:2206.06811

Adversarial Audio Synthesis with Complex-valued Polynomial Networks

Time-frequency (TF) representations in audio synthesis have been increasingly modeled with real-valued networks. However, overlooking the complex-valued nature of TF representations can result in suboptimal performance and require additional modules (e.g., for modeling the phase). To this end, we introduce complex-valued polynomial networks, called APOLLO, that integrate such complex-valued representations in a natural way. Concretely, APOLLO captures high-order correlations of the input elements using high-order tensors as scaling parameters. By leveraging standard tensor decompositions, we derive different architectures and enable modeling richer correlations. We outline such architectures and showcase their performance in audio generation across four benchmarks. As a highlight, APOLLO results in $17.5\%$ improvement over adversarial methods and $8.2\%$ over the state-of-the-art diffusion models on SC09 dataset in audio generation. Our models can encourage the systematic design of other efficient architectures on the complex field.

arXiv:2206.06259공식 코드 · Papers with Code

Realistic Gramophone Noise Synthesis using a Diffusion Model

This paper introduces a novel data-driven strategy for synthesizing gramophone noise audio textures. A diffusion probabilistic model is applied to generate highly realistic quasiperiodic noises. The proposed model is designed to generate samples of length equal to one disk revolution, but a method to generate plausible periodic variations between revolutions is also proposed. A guided approach is also applied as a conditioning method, where an audio signal generated with manually-tuned signal processing is refined via reverse diffusion to improve realism. The method has been evaluated in a subjective listening test, in which the participants were often unable to recognize the synthesized signals from the real ones. The synthetic noises produced with the best proposed unconditional method are statistically indistinguishable from real noise recordings. This work shows the potential of diffusion models for highly realistic audio synthesis tasks.

arXiv:2206.04658공식 코드 · Papers with Code

BigVGAN: A Universal Neural Vocoder with Large-Scale Training

Despite recent progress in generative adversarial network (GAN)-based vocoders, where the model generates raw waveform conditioned on acoustic features, it is challenging to synthesize high-fidelity audio for numerous speakers across various recording environments. In this work, we present BigVGAN, a universal vocoder that generalizes well for various out-of-distribution scenarios without fine-tuning. We introduce periodic activation function and anti-aliased representation into the GAN generator, which brings the desired inductive bias for audio synthesis and significantly improves audio quality. In addition, we train our GAN vocoder at the largest scale up to 112M parameters, which is unprecedented in the literature. We identify and address the failure modes in large-scale GAN training for audio, while maintaining high-fidelity output without over-regularization. Our BigVGAN, trained only on clean speech (LibriTTS), achieves the state-of-the-art performance for various zero-shot (out-of-distribution) conditions, including unseen speakers, languages, recording environments, singing voices, music, and instrumental audio. We release our code and model at: https://github.com/NVIDIA/BigVGAN

arXiv:2206.02284

Tagged-MRI Sequence to Audio Synthesis via Self Residual Attention Guided Heterogeneous Translator

Understanding the underlying relationship between tongue and oropharyngeal muscle deformation seen in tagged-MRI and intelligible speech plays an important role in advancing speech motor control theories and treatment of speech related-disorders. Because of their heterogeneous representations, however, direct mapping between the two modalities -- i.e., two-dimensional (mid-sagittal slice) plus time tagged-MRI sequence and its corresponding one-dimensional waveform -- is not straightforward. Instead, we resort to two-dimensional spectrograms as an intermediate representation, which contains both pitch and resonance, from which to develop an end-to-end deep learning framework to translate from a sequence of tagged-MRI to its corresponding audio waveform with limited dataset size.~Our framework is based on a novel fully convolutional asymmetry translator with guidance of a self residual attention strategy to specifically exploit the moving muscular structures during speech.~In addition, we leverage a pairwise correlation of the samples with the same utterances with a latent space representation disentanglement strategy.~Furthermore, we incorporate an adversarial training approach with generative adversarial networks to offer improved realism on our generated spectrograms.~Our experimental results, carried out with a total of 63 tagged-MRI sequences alongside speech acoustics, showed that our framework enabled the generation of clear audio waveforms from a sequence of tagged-MRI, surpassing competing methods. Thus, our framework provides the great potential to help better understand the relationship between the two modalities.

arXiv:2205.14807공식 코드 · Papers with Code

BinauralGrad: A Two-Stage Conditional Diffusion Probabilistic Model for Binaural Audio Synthesis

Binaural audio plays a significant role in constructing immersive augmented and virtual realities. As it is expensive to record binaural audio from the real world, synthesizing them from mono audio has attracted increasing attention. This synthesis process involves not only the basic physical warping of the mono audio, but also room reverberations and head/ear related filtrations, which, however, are difficult to accurately simulate in traditional digital signal processing. In this paper, we formulate the synthesis process from a different perspective by decomposing the binaural audio into a common part that shared by the left and right channels as well as a specific part that differs in each channel. Accordingly, we propose BinauralGrad, a novel two-stage framework equipped with diffusion models to synthesize them respectively. Specifically, in the first stage, the common information of the binaural audio is generated with a single-channel diffusion model conditioned on the mono audio, based on which the binaural audio is generated by a two-channel diffusion model in the second stage. Combining this novel perspective of two-stage synthesis with advanced generative models (i.e., the diffusion models),the proposed BinauralGrad is able to generate accurate and high-fidelity binaural audio samples. Experiment results show that on a benchmark dataset, BinauralGrad outperforms the existing baselines by a large margin in terms of both object and subject evaluation metrics (Wave L2: 0.128 vs. 0.157, MOS: 3.80 vs. 3.61). The generated audio samples (https://speechresearch.github.io/binauralgrad) and code (https://github.com/microsoft/NeuralSpeech/tree/master/BinauralGrad) are available online.

arXiv:2204.07064

Streamable Neural Audio Synthesis With Non-Causal Convolutions

Deep learning models are mostly used in an offline inference fashion. However, this strongly limits the use of these models inside audio generation setups, as most creative workflows are based on real-time digital signal processing. Although approaches based on recurrent networks can be naturally adapted to this buffer-based computation, the use of convolutions still poses some serious challenges. To tackle this issue, the use of causal streaming convolutions have been proposed. However, this requires specific complexified training and can impact the resulting audio quality. In this paper, we introduce a new method allowing to produce non-causal streaming models. This allows to make any convolutional model compatible with real-time buffer-based processing. As our method is based on a post-training reconfiguration of the model, we show that it is able to transform models trained without causal constraints into a streaming model. We show how our method can be adapted to fit complex architectures with parallel branches. To evaluate our method, we apply it on the recent RAVE model, which provides high-quality real-time audio synthesis. We test our approach on multiple music and speech datasets and show that it is faster than overlap-add methods, while having no impact on the generation quality. Finally, we introduce two open-source implementation of our work as Max/MSP and PureData externals, and as a VST audio plugin. This allows to endow traditional digital audio workstation with real-time neural audio synthesis on a laptop CPU.

arXiv:2203.01573

The Vicomtech Audio Deepfake Detection System based on Wav2Vec2 for the 2022 ADD Challenge

This paper describes our submitted systems to the 2022 ADD challenge withing the tracks 1 and 2. Our approach is based on the combination of a pre-trained wav2vec2 feature extractor and a downstream classifier to detect spoofed audio. This method exploits the contextualized speech representations at the different transformer layers to fully capture discriminative information. Furthermore, the classification model is adapted to the application scenario using different data augmentation techniques. We evaluate our system for audio synthesis detection in both the ASVspoof 2021 and the 2022 ADD challenges, showing its robustness and good performance in realistic challenging environments such as telephonic and audio codec systems, noisy audio, and partial deepfakes.

arXiv:2112.09312공식 코드 · Papers with Code

MIDI-DDSP: Detailed Control of Musical Performance via Hierarchical Modeling

Musical expression requires control of both what notes are played, and how they are performed. Conventional audio synthesizers provide detailed expressive controls, but at the cost of realism. Black-box neural audio synthesis and concatenative samplers can produce realistic audio, but have few mechanisms for control. In this work, we introduce MIDI-DDSP a hierarchical model of musical instruments that enables both realistic neural audio synthesis and detailed user control. Starting from interpretable Differentiable Digital Signal Processing (DDSP) synthesis parameters, we infer musical notes and high-level properties of their expressive performance (such as timbre, vibrato, dynamics, and articulation). This creates a 3-level hierarchy (notes, performance, synthesis) that affords individuals the option to intervene at each level, or utilize trained priors (performance given notes, synthesis given performance) for creative assistance. Through quantitative experiments and listening tests, we demonstrate that this hierarchy can reconstruct high-fidelity audio, accurately predict performance attributes for a note sequence, independently manipulate the attributes of a given performance, and as a complete system, generate realistic audio from a novel note sequence. By utilizing an interpretable hierarchy, with multiple levels of granularity, MIDI-DDSP opens the door to assistive tools to empower individuals across a diverse range of musical experience.

arXiv:2111.10003

Differentiable Wavetable Synthesis

Differentiable Wavetable Synthesis (DWTS) is a technique for neural audio synthesis which learns a dictionary of one-period waveforms i.e. wavetables, through end-to-end training. We achieve high-fidelity audio synthesis with as little as 10 to 20 wavetables and demonstrate how a data-driven dictionary of waveforms opens up unprecedented one-shot learning paradigms on short audio clips. Notably, we show audio manipulations, such as high quality pitch-shifting, using only a few seconds of input audio. Lastly, we investigate performance gains from using learned wavetables for realtime and interactive audio synthesis.

arXiv:2111.08462

Towards Lightweight Controllable Audio Synthesis with Conditional Implicit Neural Representations

The high temporal resolution of audio and our perceptual sensitivity to small irregularities in waveforms make synthesizing at high sampling rates a complex and computationally intensive task, prohibiting real-time, controllable synthesis within many approaches. In this work we aim to shed light on the potential of Conditional Implicit Neural Representations (CINRs) as lightweight backbones in generative frameworks for audio synthesis. Our experiments show that small Periodic Conditional INRs (PCINRs) learn faster and generally produce quantitatively better audio reconstructions than Transposed Convolutional Neural Networks with equal parameter counts. However, their performance is very sensitive to activation scaling hyperparameters. When learning to represent more uniform sets, PCINRs tend to introduce artificial high-frequency components in reconstructions. We validate this noise can be minimized by applying standard weight regularization during training or decreasing the compositional depth of PCINRs, and suggest directions for future research.

arXiv:2111.05174공식 코드 · Papers with Code

CAESynth: Real-Time Timbre Interpolation and Pitch Control with Conditional Autoencoders

In this paper, we present a novel audio synthesizer, CAESynth, based on a conditional autoencoder. CAESynth synthesizes timbre in real-time by interpolating the reference sounds in their shared latent feature space, while controlling a pitch independently. We show that training a conditional autoencoder based on accuracy in timbre classification together with adversarial regularization of pitch content allows timbre distribution in latent space to be more effective and stable for timbre interpolation and pitch conditioning. The proposed method is applicable not only to creation of musical cues but also to exploration of audio affordance in mixed reality based on novel timbre mixtures with environmental sounds. We demonstrate by experiments that CAESynth achieves smooth and high-fidelity audio synthesis in real-time through timbre interpolation and independent yet accurate pitch control for musical cues as well as for audio affordance with environmental sound. A Python implementation along with some generated samples are shared online.

arXiv:2111.05011공식 코드 · Papers with Code

RAVE: A variational autoencoder for fast and high-quality neural audio synthesis

Deep generative models applied to audio have improved by a large margin the state-of-the-art in many speech and music related tasks. However, as raw waveform modelling remains an inherently difficult task, audio generative models are either computationally intensive, rely on low sampling rates, are complicated to control or restrict the nature of possible signals. Among those models, Variational AutoEncoders (VAE) give control over the generation by exposing latent variables, although they usually suffer from low synthesis quality. In this paper, we introduce a Realtime Audio Variational autoEncoder (RAVE) allowing both fast and high-quality audio waveform synthesis. We introduce a novel two-stage training procedure, namely representation learning and adversarial fine-tuning. We show that using a post-training analysis of the latent space allows a direct control between the reconstruction fidelity and the representation compactness. By leveraging a multi-band decomposition of the raw waveform, we show that our model is the first able to generate 48kHz audio signals, while simultaneously running 20 times faster than real-time on a standard laptop CPU. We evaluate synthesis quality using both quantitative and qualitative subjective experiments and show the superiority of our approach compared to existing models. Finally, we present applications of our model for timbre transfer and signal compression. All of our source code and audio examples are publicly available.

arXiv:2111.04726

Estimating High Order Gradients of the Data Distribution by Denoising

The first order derivative of a data density can be estimated efficiently by denoising score matching, and has become an important component in many applications, such as image generation and audio synthesis. Higher order derivatives provide additional local information about the data distribution and enable new applications. Although they can be estimated via automatic differentiation of a learned density model, this can amplify estimation errors and is expensive in high dimensional settings. To overcome these limitations, we propose a method to directly estimate high order derivatives (scores) of a data density from samples. We first show that denoising score matching can be interpreted as a particular case of Tweedie's formula. By leveraging Tweedie's formula on higher order moments, we generalize denoising score matching to estimate higher order derivatives. We demonstrate empirically that models trained with the proposed method can approximate second order derivatives more efficiently and accurately than via automatic differentiation. We show that our models can be used to quantify uncertainty in denoising and to improve the mixing speed of Langevin dynamics via Ozaki discretization for sampling synthetic data and natural images.

arXiv:2108.00443

A Survey on Audio Synthesis and Audio-Visual Multimodal Processing

With the development of deep learning and artificial intelligence, audio synthesis has a pivotal role in the area of machine learning and shows strong applicability in the industry. Meanwhile, significant efforts have been dedicated by researchers to handle multimodal tasks at present such as audio-visual multimodal processing. In this paper, we conduct a survey on audio synthesis and audio-visual multimodal processing, which helps understand current research and future trends. This review focuses on text to speech(TTS), music generation and some tasks that combine visual and acoustic information. The corresponding technical methods are comprehensively classified and introduced, and their future development trends are prospected. This survey can provide some guidance for researchers who are interested in the areas like audio synthesis and audio-visual multimodal processing.