Multi-channel speech separation in dynamic environments is challenging as time-varying spatial and spectral features evolve at different temporal scales. Existing methods typically employ sequential architectures, forcing a single network stream to simultaneously model both feature types, creating an inherent modeling conflict. In this paper, we propose a dual-branch parallel spectral-spatial (PS2) architecture that separately processes spectral and spatial features through parallel streams. The spectral branch uses a bi-directional long short-term memory (BLSTM)-based frequency module, a Mamba-based temporal module, and a self-attention module to model spectral features. The spatial branch employs bi-directional gated recurrent unit (BGRU) networks to process spatial features that encode the evolving geometric relationships between sources and microphones. Features from both branches are integrated through a cross-attention fusion mechanism that adaptively weights their contributions. Experimental results demonstrate that the PS2 outperforms existing state-of-the-art (SOTA) methods by 1.6-2.2 dB in scale-invariant signal-to-distortion ratio (SI-SDR) for moving speaker scenarios, with robust separation quality under different reverberation times (RT60), noise levels, and source movement speeds. Even with fast source movements, the proposed model maintains SI-SDR improvements of over 13 dB. These improvements are consistently observed across multiple datasets, including WHAMR! and our generated WSJ0-Demand-6ch-Move dataset.

Audio context determines which sound components and sources are relevant and which can be perceived as irrelevant (noise) by listeners. For example, traffic noise is informative in urban surveillance but noise for a phone call at the same location. Most current audio denoising systems apply fixed target-noise definitions, often removing useful components in one context while failing to suppress irrelevant components. To address this, we introduce the concept automatic contextual audio denoising (ACAD) which defines target and noise based on the inferred context. In this work, we restrict context to be associated with an acoustic scene class. We label sound events outside the event distribution of a scene class (noise) as out-of-context (OC) and events typical for that scene as in-context (IC). We implement a deep learning method that automatically infers the context of the audio signal and removes OC components, and benchmark it against variants: without context inference, with oracle context, and with separately provided uninformative context. On paired clean/noisy data across diverse contexts, where OC components in one context may be IC in another, our proposed method outperforms other approaches across standard objective metrics, indicating that the model can infer context and context-dependent processing can enhance denoising.

We present a deep neural network approach for encoding microphone array signals into Ambisonics that generalizes to arbitrary microphone array configurations with fixed microphone count but varying locations and frequency-dependent directional characteristics. Unlike previous methods that rely only on array geometry as metadata, our approach uses directional array transfer functions, enabling accurate characterization of real-world arrays. The proposed architecture employs separate encoders for audio and directional responses, combining them through cross-attention mechanisms to generate array-independent spatial audio representations. We evaluate the method on simulated data in two settings: a mobile phone with complex body scattering, and a free-field condition, both with varying numbers of sound sources in reverberant environments. Evaluations demonstrate that our approach outperforms both conventional digital signal processing-based methods and existing deep neural network solutions. Furthermore, using array transfer functions instead of geometry as metadata input improves accuracy on realistic arrays.

Generative adversarial networks (GANs) and diffusion models have recently achieved state-of-the-art performance in audio super-resolution (ADSR), producing perceptually convincing wideband audio from narrowband inputs. However, existing evaluations primarily rely on signal-level or perceptual metrics, leaving open the question of how closely the distributions of synthetic super-resolved and real wideband audio match. Here we address this problem by analyzing the separability of real and super-resolved audio in various embedding spaces. We consider both middle-band (4→16~kHz) and full-band (16→48~kHz) upsampling tasks for speech and music, training linear classifiers to distinguish real from synthetic samples based on multiple types of audio embeddings. Comparisons with objective metrics and subjective listening tests reveal that embedding-based classifiers achieve near-perfect separation, even when the generated audio attains high perceptual quality and state-of-the-art metric scores. This behavior is consistent across datasets and models, including recent diffusion-based approaches, highlighting a persistent gap between perceptual quality and true distributional fidelity in ADSR models.

A system for training a microphone geometry assisted encoder model and then utilizing the trained model to generate spatial audio signals that have been captured by a plurality of microphones. In a method for generating spatial audio signals, the method includes receiving geometry data related to a plurality of microphones of an audio capturing device and audio signal data captured by the plurality of microphones. The method also includes generating a spatial audio signal based on an output of a trained microphone geometry assisted encoder model. The trained microphone geometry assisted encoder model includes a geometry encoder configured to encode the geometry data and a signal encoder configured to encode the audio signal data. The trained microphone geometry assisted encoder model further includes a signal decoder having a plurality of layers and configured to generate the output upon which the spatial audio signal is based.

Examples of the disclosure relate to a model that can be used for speech enhancement. The model comprises an encoder part comprising a sequence of encoding layers and caused to receive input data. The input data is based on a current frame of a noisy speech signal and one or more past frames of the noisy speech signal. The sequence of encoding layers is caused to process the input data so that output data of the encoder part comprises a reduced number of the multiple frequency positions and a single temporal position. The model also comprises a decoder part comprising a sequence of decoding layers caused to receive data from a prior decoding layer. The output data of the decoder part comprises multiple frequency positions and a single temporal position. The output data of the decoder part is for post processing to provide an output signal for speech enhancement.

The present disclosure relates to an apparatus, that obtains a far-end signal and a near-end microphone signal, determines, based on at least the far-end signal, a far-end speech signal estimate and a far-end noise signal estimate, determines, based on at least the near-end microphone signal, a near-end microphone speech signal estimate and a near-end microphone noise signal estimate, determines, based on at least the far-end speech signal estimate and the near-end microphone speech signal estimate, a predicted near-end speech signal, determines, based on at least the far-end noise signal estimate and the near-end microphone noise signal estimate, a predicted near-end noise signal and outputs at least the predicted near-end speech signal and predicted near-end noise signal.