Previous research on late-reverberation modeling has mainly focused on exponentially decaying room impulse responses, whereas methods for accurately modeling non-exponential reverberation remain challenging. This paper extends the previously proposed basic dark-velvet-noise reverberation algorithm and proposes a parametrization scheme for modeling late reverberation with arbitrary temporal energy decay. Each pulse in the velvet-noise sequence is routed to a single dictionary filter that is selected from a set of filters based on weighted probabilities. The probabilities control the spectral evolution of the late-reverberation model and are optimized to fit a target impulse response via non-negative least-squares optimization. In this way, the frequency-dependent energy decay of a target late-reverberation impulse response can be fitted with mean and maximum T60 errors of 4% and 8%, respectively, requiring about 50% less coloration filters than a previously proposed filtered velvet-noise algorithm. Furthermore, the extended dark-velvet-noise reverberation algorithm allows the modeled impulse response to be gated, the frequency-dependent reverberation time to be modified, and the model's spectral evolution and broadband decay to be decoupled. The proposed method is suitable for the parametric late-reverberation synthesis of various acoustic environments, especially spaces that exhibit a non-exponential energy decay, motivating its use in musical audio and virtual reality.
This paper presents a novel approach to audio restoration, focusing on the enhancement of low-quality music recordings, and in particular historical ones. Building upon a previous algorithm called BABE, or Blind Audio Bandwidth Extension, we introduce BABE-2, which presents a series of significant improvements. This research broadens the concept of bandwidth extension to \emph{generative equalization}, a novel task that, to the best of our knowledge, has not been explicitly addressed in previous studies. BABE-2 is built around an optimization algorithm utilizing priors from diffusion models, which are trained or fine-tuned using a curated set of high-quality music tracks. The algorithm simultaneously performs two critical tasks: estimation of the filter degradation magnitude response and hallucination of the restored audio. The proposed method is objectively evaluated on historical piano recordings, showing a marked enhancement over the prior version. The method yields similarly impressive results in rejuvenating the works of renowned vocalists Enrico Caruso and Nellie Melba. This research represents an advancement in the practical restoration of historical music.
A common bane of artificial reverberation algorithms is spectral coloration, typically manifesting as metallic ringing, leading to a degradation in the perceived sound quality. This paper presents an optimization framework where a differentiable feedback delay network is used to learn a set of parameters to reduce coloration iteratively. The parameters under optimization include the feedback matrix, as well as the input and output gains. The optimization objective is twofold: to maximize spectral flatness through a spectral loss while maintaining temporal density by penalizing sparseness in the parameter values. A favorable narrower distribution of modal excitation is achieved while maintaining the desired impulse response density. In a subjective assessment, the new method proves effective in reducing perceptual coloration of late reverberation. The proposed method achieves computational savings compared to the baseline while preserving its performance. The effectiveness of this work is demonstrated through two application scenarios where natural-sounding synthetic impulse responses are obtained via the introduction of attenuation filters and an optimizable scattering feedback matrix.
With the development of audio playback devices and fast data transmission, the demand for high sound quality is rising, for both entertainment and communications. In this quest for better sound quality, challenges emerge from distortions and interferences originating at the recording side or caused by an imperfect transmission pipeline. To address this problem, audio restoration methods aim to recover clean sound signals from the corrupted input data. We present here audio restoration algorithms based on diffusion models, with a focus on speech enhancement and music restoration tasks. Traditional approaches, often grounded in handcrafted rules and statistical heuristics, have shaped our understanding of audio signals. In the past decades, there has been a notable shift towards data-driven methods that exploit the modeling capabilities of deep neural networks (DNNs). Deep generative models, and among them diffusion models, have emerged as powerful techniques for learning complex data distributions. However, relying solely on DNN-based learning approaches carries the risk of reducing interpretability, particularly when employing end-to-end models. Nonetheless, data-driven approaches allow more flexibility in comparison to statistical model-based frameworks whose performance depends on distributional and statistical assumptions that can be difficult to guarantee. Here, we aim to show that diffusion models can combine the best of both worlds and offer the opportunity to design audio restoration algorithms with a good degree of interpretability and a remarkable performance in terms of sound quality.
This letter introduces an innovative method to enhance the quality of audio time stretching by precisely decomposing a sound into sines, transients, and noise and by improving the processing of the latter component. While there are established methods for time-stretching sines and transients with high quality, the manipulation of noise or residual components has lacked robust solutions in prior research. The proposed method combines sound decomposition with previous techniques for audio spectral resynthesis. The time-stretched noise component is achieved by morphing its time-interpolated spectral magnitude with a white-noise excitation signal. This method stands out for its simplicity, efficiency, and audio quality. The results of a subjective experiment affirm the superiority of this approach over current state-of-the-art methods across all evaluated stretch factors. The proposed technique notably excels in extreme stretching scenarios, signifying a substantial elevation in performance. The proposed method holds promise for a wide range of applications in slow-motion media content, such as music or sports video production.
Several individualization methods have recently been proposed to estimate a subject's Head-Related Transfer Function (HRTF) using convenient input modalities such as anthropometric measurements or pinnae photographs. There exists a need for adaptively correcting the estimation error committed by such methods using a few data point samples from the subject's HRTF, acquired using acoustic measurements or perceptual feedback. To this end, we introduce a Convolutional Conditional Neural Process meta-learner specialized in HRTF error interpolation. In particular, the model includes a Spherical Convolutional Neural Network component to accommodate the spherical geometry of HRTF data. It also exploits potential symmetries between the HRTF's left and right channels about the median axis. In this work, we evaluate the proposed model's performance purely on time-aligned spectrum interpolation grounds under a simplified setup where a generic population-mean HRTF forms the initial estimates prior to corrections instead of individualized ones. The trained model achieves up to 3 dB relative error reduction compared to state-of-the-art interpolation methods despite being trained using only 85 subjects. This improvement translates up to nearly a halving of the data point count required to achieve comparable accuracy, in particular from 50 to 28 points to reach an average of -20 dB relative error per interpolated feature. Moreover, we show that the trained model provides well-calibrated uncertainty estimates. Accordingly, such estimates can inform the sequential decision problem of acquiring as few correcting HRTF data points as needed to meet a desired level of HRTF individualization accuracy.
Audio bandwidth extension involves the realistic reconstruction of high-frequency spectra from bandlimited observations. In cases where the lowpass degradation is unknown, such as in restoring historical audio recordings, this becomes a blind problem. This paper introduces a novel method called BABE (Blind Audio Bandwidth Extension) that addresses the blind problem in a zero-shot setting, leveraging the generative priors of a pre-trained unconditional diffusion model. During the inference process, BABE utilizes a generalized version of diffusion posterior sampling, where the degradation operator is unknown but parametrized and inferred iteratively. The performance of the proposed method is evaluated using objective and subjective metrics, and the results show that BABE surpasses state-of-the-art blind bandwidth extension baselines and achieves competitive performance compared to non-blind filter-informed methods when tested with synthetic data. Moreover, BABE exhibits robust generalization capabilities when enhancing real historical recordings, effectively reconstructing the missing high-frequency content while maintaining coherence with the original recording. Subjective preference tests confirm that BABE significantly improves the audio quality of historical music recordings. Examples of historical recordings restored with the proposed method are available on the companion webpage: (http://research.spa.aalto.fi/publications/papers/ieee-taslp-babe/)
The sound of magnetic recording media, such as open-reel and cassette tape recorders, is still sought after by today's sound practitioners due to the imperfections embedded in the physics of the magnetic recording process. This paper proposes a method for digitally emulating this character using neural networks. The signal chain of the proposed system consists of three main components: the hysteretic nonlinearity and filtering jointly produced by the magnetic recording process as well as the record and playback amplifiers, the fluctuating delay originating from the tape transport, and the combined additive noise component from various electromagnetic origins. In our approach, the hysteretic nonlinear block is modeled using a recurrent neural network, while the delay trajectories and the noise component are generated using separate diffusion models, which employ U-net deep convolutional neural networks. According to the conducted objective evaluation, the proposed architecture faithfully captures the character of the magnetic tape recorder. The results of this study can be used to construct virtual replicas of vintage sound recording devices with applications in music production and audio antiquing tasks.
Audio inpainting aims to reconstruct missing segments in corrupted recordings. Previous methods produce plausible reconstructions when the gap length is shorter than about 100\;ms, but the quality decreases for longer gaps. This paper explores recent advancements in deep learning and, particularly, diffusion models, for the task of audio inpainting. The proposed method uses an unconditionally trained generative model, which can be conditioned in a zero-shot fashion for audio inpainting, offering high flexibility to regenerate gaps of arbitrary length. An improved deep neural network architecture based on the constant-Q transform, which allows the model to exploit pitch-equivariant symmetries in audio, is also presented. The performance of the proposed algorithm is evaluated through objective and subjective metrics for the task of reconstructing short to mid-sized gaps. The results of a formal listening test show that the proposed method delivers a comparable performance against state-of-the-art for short gaps, while retaining a good audio quality and outperforming the baselines for the longest gap lengths tested, 150\;ms and 200\;ms. This work helps improve the restoration of sound recordings having fairly long local disturbances or dropouts, which must be reconstructed.
A deep neural network solution for time-scale modification (TSM) focused on large stretching factors is proposed, targeting environmental sounds. Traditional TSM artifacts such as transient smearing, loss of presence, and phasiness are heavily accentuated and cause poor audio quality when the TSM factor is four or larger. The weakness of established TSM methods, often based on a phase vocoder structure, lies in the poor description and scaling of the transient and noise components, or nuances, of a sound. Our novel solution combines a sines-transients-noise decomposition with an independent WaveNet synthesizer to provide a better description of the noise component and an improve sound quality for large stretching factors. Results of a subjective listening test against four other TSM algorithms are reported, showing the proposed method to be often superior. The proposed method is stereo compatible and has a wide range of applications related to the slow motion of media content.