Autoregressive models are widely used for tasks such as image and audio generation. The sampling process of these models, however, does not allow interruptions and cannot adapt to real-time computational resources. This challenge impedes the deployment of powerful autoregressive models, which involve a slow sampling process that is sequential in nature and typically scales linearly with respect to the data dimension. To address this difficulty, we propose a new family of autoregressive models that enables anytime sampling. Inspired by Principal Component Analysis, we learn a structured representation space where dimensions are ordered based on their importance with respect to reconstruction. Using an autoregressive model in this latent space, we trade off sample quality for computational efficiency by truncating the generation process before decoding into the original data space. Experimentally, we demonstrate in several image and audio generation tasks that sample quality degrades gracefully as we reduce the computational budget for sampling. The approach suffers almost no loss in sample quality (measured by FID) using only 60\% to 80\% of all latent dimensions for image data. Code is available at https://github.com/Newbeeer/Anytime-Auto-Regressive-Model .
This paper describes an end-to-end adversarial singing voice conversion (EA-SVC) approach. It can directly generate arbitrary singing waveform by given phonetic posteriorgram (PPG) representing content, F0 representing pitch, and speaker embedding representing timbre, respectively. Proposed system is composed of three modules: generator $G$, the audio generation discriminator $D_{A}$, and the feature disentanglement discriminator $D_F$. The generator $G$ encodes the features in parallel and inversely transforms them into the target waveform. In order to make timbre conversion more stable and controllable, speaker embedding is further decomposed to the weighted sum of a group of trainable vectors representing different timbre clusters. Further, to realize more robust and accurate singing conversion, disentanglement discriminator $D_F$ is proposed to remove pitch and timbre related information that remains in the encoded PPG. Finally, a two-stage training is conducted to keep a stable and effective adversarial training process. Subjective evaluation results demonstrate the effectiveness of our proposed methods. Proposed system outperforms conventional cascade approach and the WaveNet based end-to-end approach in terms of both singing quality and singer similarity. Further objective analysis reveals that the model trained with the proposed two-stage training strategy can produce a smoother and sharper formant which leads to higher audio quality.
We present a novel system that gets as an input video frames of a musician playing the piano and generates the music for that video. Generation of music from visual cues is a challenging problem and it is not clear whether it is an attainable goal at all. Our main aim in this work is to explore the plausibility of such a transformation and to identify cues and components able to carry the association of sounds with visual events. To achieve the transformation we built a full pipeline named `\textit{Audeo}' containing three components. We first translate the video frames of the keyboard and the musician hand movements into raw mechanical musical symbolic representation Piano-Roll (Roll) for each video frame which represents the keys pressed at each time step. We then adapt the Roll to be amenable for audio synthesis by including temporal correlations. This step turns out to be critical for meaningful audio generation. As a last step, we implement Midi synthesizers to generate realistic music. \textit{Audeo} converts video to audio smoothly and clearly with only a few setup constraints. We evaluate \textit{Audeo} on `in the wild' piano performance videos and obtain that their generated music is of reasonable audio quality and can be successfully recognized with high precision by popular music identification software.
In a recent paper, we have presented a generative adversarial network (GAN)-based model for unconditional generation of the mel-spectrograms of singing voices. As the generator of the model is designed to take a variable-length sequence of noise vectors as input, it can generate mel-spectrograms of variable length. However, our previous listening test shows that the quality of the generated audio leaves room for improvement. The present paper extends and expands that previous work in the following aspects. First, we employ a hierarchical architecture in the generator to induce some structure in the temporal dimension. Second, we introduce a cycle regularization mechanism to the generator to avoid mode collapse. Third, we evaluate the performance of the new model not only for generating singing voices, but also for generating speech voices. Evaluation result shows that new model outperforms the prior one both objectively and subjectively. We also employ the model to unconditionally generate sequences of piano and violin music and find the result promising. Audio examples, as well as the code for implementing our model, will be publicly available online upon paper publication.
We introduce GACELA, a generative adversarial network (GAN) designed to restore missing musical audio data with a duration ranging between hundreds of milliseconds to a few seconds, i.e., to perform long-gap audio inpainting. While previous work either addressed shorter gaps or relied on exemplars by copying available information from other signal parts, GACELA addresses the inpainting of long gaps in two aspects. First, it considers various time scales of audio information by relying on five parallel discriminators with increasing resolution of receptive fields. Second, it is conditioned not only on the available information surrounding the gap, i.e., the context, but also on the latent variable of the conditional GAN. This addresses the inherent multi-modality of audio inpainting at such long gaps and provides the option of user-defined inpainting. GACELA was tested in listening tests on music signals of varying complexity and gap durations ranging from 375~ms to 1500~ms. While our subjects were often able to detect the inpaintings, the severity of the artifacts decreased from unacceptable to mildly disturbing. GACELA represents a framework capable to integrate future improvements such as processing of more auditory-related features or more explicit musical features.
Most generative models of audio directly generate samples in one of two domains: time or frequency. While sufficient to express any signal, these representations are inefficient, as they do not utilize existing knowledge of how sound is generated and perceived. A third approach (vocoders/synthesizers) successfully incorporates strong domain knowledge of signal processing and perception, but has been less actively researched due to limited expressivity and difficulty integrating with modern auto-differentiation-based machine learning methods. In this paper, we introduce the Differentiable Digital Signal Processing (DDSP) library, which enables direct integration of classic signal processing elements with deep learning methods. Focusing on audio synthesis, we achieve high-fidelity generation without the need for large autoregressive models or adversarial losses, demonstrating that DDSP enables utilizing strong inductive biases without losing the expressive power of neural networks. Further, we show that combining interpretable modules permits manipulation of each separate model component, with applications such as independent control of pitch and loudness, realistic extrapolation to pitches not seen during training, blind dereverberation of room acoustics, transfer of extracted room acoustics to new environments, and transformation of timbre between disparate sources. In short, DDSP enables an interpretable and modular approach to generative modeling, without sacrificing the benefits of deep learning. The library is publicly available at https://github.com/magenta/ddsp and we welcome further contributions from the community and domain experts.
Generative models for singing voice have been mostly concerned with the task of ``singing voice synthesis,'' i.e., to produce singing voice waveforms given musical scores and text lyrics. In this work, we explore a novel yet challenging alternative: singing voice generation without pre-assigned scores and lyrics, in both training and inference time. In particular, we outline three such generation schemes, and propose a pipeline to tackle these new tasks. Moreover, we implement such models using generative adversarial networks and evaluate them both objectively and subjectively.
Source separation for music is the task of isolating contributions, or stems, from different instruments recorded individually and arranged together to form a song. Such components include voice, bass, drums and any other accompaniments.Contrarily to many audio synthesis tasks where the best performances are achieved by models that directly generate the waveform, the state-of-the-art in source separation for music is to compute masks on the magnitude spectrum. In this paper, we compare two waveform domain architectures. We first adapt Conv-Tasnet, initially developed for speech source separation,to the task of music source separation. While Conv-Tasnet beats many existing spectrogram-domain methods, it suffersfrom significant artifacts, as shown by human evaluations. We propose instead Demucs, a novel waveform-to-waveform model,with a U-Net structure and bidirectional LSTM.Experiments on the MusDB dataset show that, with proper data augmentation, Demucs beats allexisting state-of-the-art architectures, including Conv-Tasnet, with 6.3 SDR on average, (and up to 6.8 with 150 extra training songs, even surpassing the IRM oracle for the bass source).Using recent development in model quantization, Demucs can be compressed down to 120MBwithout any loss of accuracy.We also provide human evaluations, showing that Demucs benefit from a large advantagein terms of the naturalness of the audio. However, it suffers from some bleeding,especially between the vocals and other source.
Convolutional neural networks (CNNs) with dilated filters such as the Wavenet or the Temporal Convolutional Network (TCN) have shown good results in a variety of sequence modelling tasks. However, efficiently modelling long-term dependencies in these sequences is still challenging. Although the receptive field of these models grows exponentially with the number of layers, computing the convolutions over very long sequences of features in each layer is time and memory-intensive, prohibiting the use of longer receptive fields in practice. To increase efficiency, we make use of the "slow feature" hypothesis stating that many features of interest are slowly varying over time. For this, we use a U-Net architecture that computes features at multiple time-scales and adapt it to our auto-regressive scenario by making convolutions causal. We apply our model ("Seq-U-Net") to a variety of tasks including language and audio generation. In comparison to TCN and Wavenet, our network consistently saves memory and computation time, with speed-ups for training and inference of over 4x in the audio generation experiment in particular, while achieving a comparable performance in all tasks.
End-to-end models for raw audio generation are a challenge, specially if they have to work with non-parallel data, which is a desirable setup in many situations. Voice conversion, in which a model has to impersonate a speaker in a recording, is one of those situations. In this paper, we propose Blow, a single-scale normalizing flow using hypernetwork conditioning to perform many-to-many voice conversion between raw audio. Blow is trained end-to-end, with non-parallel data, on a frame-by-frame basis using a single speaker identifier. We show that Blow compares favorably to existing flow-based architectures and other competitive baselines, obtaining equal or better performance in both objective and subjective evaluations. We further assess the impact of its main components with an ablation study, and quantify a number of properties such as the necessary amount of training data or the preference for source or target speakers.
Generative models have thrived in computer vision, enabling unprecedented image processes. Yet the results in audio remain less advanced. Our project targets real-time sound synthesis from a reduced set of high-level parameters, including semantic controls that can be adapted to different sound libraries and specific tags. These generative variables should allow expressive modulations of target musical qualities and continuously mix into new styles. To this extent we train AEs on an orchestral database of individual note samples, along with their intrinsic attributes: note class, timbre domain and extended playing techniques. We condition the decoder for control over the rendered note attributes and use latent adversarial training for learning expressive style parameters that can ultimately be mixed. We evaluate both generative performances and latent representation. Our ablation study demonstrates the effectiveness of the musical conditioning mechanisms. The proposed model generates notes as magnitude spectrograms from any probabilistic latent code samples, with expressive control of orchestral timbres and playing styles. Its training data subsets can directly be visualized in the 3D latent representation. Waveform rendering can be done offline with GLA. In order to allow real-time interactions, we fine-tune the decoder with a pretrained MCNN and embed the full waveform generation pipeline in a plugin. Moreover the encoder could be used to process new input samples, after manipulating their latent attribute representation, the decoder can generate sample variations as an audio effect would. Our solution remains rather fast to train, it can directly be applied to other sound domains, including an user's libraries with custom sound tags that could be mapped to specific generative controls. As a result, it fosters creativity and intuitive audio style experimentations.
Time-frequency (TF) representations provide powerful and intuitive features for the analysis of time series such as audio. But still, generative modeling of audio in the TF domain is a subtle matter. Consequently, neural audio synthesis widely relies on directly modeling the waveform and previous attempts at unconditionally synthesizing audio from neurally generated invertible TF features still struggle to produce audio at satisfying quality. In this article, focusing on the short-time Fourier transform, we discuss the challenges that arise in audio synthesis based on generated invertible TF features and how to overcome them. We demonstrate the potential of deliberate generative TF modeling by training a generative adversarial network (GAN) on short-time Fourier features. We show that by applying our guidelines, our TF-based network was able to outperform a state-of-the-art GAN generating waveforms directly, despite the similar architecture in the two networks.
Generative models are successfully used for image synthesis in the recent years. But when it comes to other modalities like audio, text etc little progress has been made. Recent works focus on generating audio from a generative model in an unsupervised setting. We explore the possibility of using generative models conditioned on class labels. Concatenation based conditioning and conditional scaling were explored in this work with various hyper-parameter tuning methods. In this paper we introduce Conditional WaveGANs (cWaveGAN). Find our implementation at https://github.com/acheketa/cwavegan
Dilated convolutions, also known as atrous convolutions, have been widely explored in deep convolutional neural networks (DCNNs) for various dense prediction tasks. However, dilated convolutions suffer from the gridding artifacts, which hampers the performance. In this work, we propose two simple yet effective degridding methods by studying a decomposition of dilated convolutions. Unlike existing models, which explore solutions by focusing on a block of cascaded dilated convolutional layers, our methods address the gridding artifacts by smoothing the dilated convolution itself. In addition, we point out that the two degridding approaches are intrinsically related and define separable and shared (SS) operations, which generalize the proposed methods. We further explore SS operations in view of operations on graphs and propose the SS output layer, which is able to smooth the entire DCNNs by only replacing the output layer. We evaluate our degridding methods and the SS output layer thoroughly, and visualize the smoothing effect through effective receptive field analysis. Results show that our methods degridding yield consistent improvements on the performance of dense prediction tasks, while adding negligible amounts of extra training parameters. And the SS output layer improves the performance significantly and is very efficient in terms of number of training parameters.
In this paper we propose a novel model for unconditional audio generation based on generating one audio sample at a time. We show that our model, which profits from combining memory-less modules, namely autoregressive multilayer perceptrons, and stateful recurrent neural networks in a hierarchical structure is able to capture underlying sources of variations in the temporal sequences over very long time spans, on three datasets of different nature. Human evaluation on the generated samples indicate that our model is preferred over competing models. We also show how each component of the model contributes to the exhibited performance.
Audio signals are sampled at high temporal resolutions, and learning to synthesize audio requires capturing structure across a range of timescales. Generative adversarial networks (GANs) have seen wide success at generating images that are both locally and globally coherent, but they have seen little application to audio generation. In this paper we introduce WaveGAN, a first attempt at applying GANs to unsupervised synthesis of raw-waveform audio. WaveGAN is capable of synthesizing one second slices of audio waveforms with global coherence, suitable for sound effect generation. Our experiments demonstrate that, without labels, WaveGAN learns to produce intelligible words when trained on a small-vocabulary speech dataset, and can also synthesize audio from other domains such as drums, bird vocalizations, and piano. We compare WaveGAN to a method which applies GANs designed for image generation on image-like audio feature representations, finding both approaches to be promising.