Overview

Speech enhancement seeks to recover clear, intelligible speech from recordings degraded by background noise, reverberation, echo, or other speakers. Methods range from spectral masking and multichannel beamforming, such as MVDR, to neural time-domain models and newer diffusion, flow-matching, and codec-based generators; visual cues and speaker references can help identify the target. Systems are commonly tested on benchmarks such as VoiceBank+DEMAND, DNS, and CHiME, while real-world meetings, hearing aids, and mobile devices bring challenges in domain shift, latency, and power use. A central tension is that better perceptual scores do not always mean more accurate speech recognition, and generative methods must improve quality without changing the speaker’s words or introducing artifacts.

Latest papers

134 papers11 papers in the last four weeks, down 54% on the four weeks before. 0.1% of all new papers.

Nov 9, 2026

GigaEdge Spench: An SDK for Practical Streaming Speech Enhancement

We present GigaEdge Spench, an open-source SDK for streaming speech enhancement on CPU servers, mobile devices, and in browsers. Applications can build enhancement channels from neural denoising, normalization and dynamics blocks, sample-rate conversion, and acoustic echo cancellation where supported. These stages can be combined, while normalization can also run without a neural model. The release includes an 8 kHz INT8 denoiser and discriminative and hybrid 48 kHz FP32 models. A shared C++ core supports ONNX Runtime and TensorFlow Lite through native interfaces, Android and iOS SDKs, and WebAssembly integration. Independent stream contexts share model resources, and consecutive frames from one stream can be processed in a single inference call. We evaluate the released models on EARS-WHAM v2 and compare their quality and processing costs with RNNoise and DPDFNet2. All tested native and Web configurations process files faster than real time, although larger chunks require more time to collect input. We also measure how many streams a shared model can process on a CPU. Our contribution is a configurable deployment system and an experimental study of how model choice, chunk size, and execution environment affect its use. All reported SDK timing uses ONNX Runtime. Index Terms: speech enhancement, streaming inference, cross-platform deployment, temporal frame grouping, computational efficiency
Nov 9, 2026

Textraction: Unified Text-Guided Speaker Extraction and Enhancement for Single- and Multi-Channel Audio

We present Textraction, a text-guided speaker extraction framework that enables the selective extraction of desired speakers or noise sources from an audio mixture using only natural-language descriptions. The proposed method operates within a single unified model that supports both single-channel and multi-channel inputs and it allows conditioning on either semantic textual descriptions of the target source or on its spatial description. Dedicated single- and multi-channel encoders feed a shared stack of time--frequency transformer blocks, which is conditioned on the prompt and on the recording configuration in every block through cross-attention and estimates the target directly from the mixture in one forward pass, without a separation stage or a source-selection stage. The system can extract an arbitrary subset of entities from a mixture, assuming up to two simultaneously active speakers and two noise sources, and interprets prompts specifying which entities to keep, which to remove, or both. In addition, we show that the framework naturally applies to a speech-enhancement setting, in which a single target speaker is present alongside background noise. To circumvent the application of computationally costly large language models used in prior work, we introduce a lightweight, task-adapted text encoder obtained by contrastively fine-tuning a compact text-retrieval model, which learns to be sensitive to instruction polarity (e.g., keep vs. remove) while remaining robust to paraphrasing. Experiments across attribute-based extraction, spatially guided extraction, and speech enhancement show that the proposed approach effectively isolates the desired entities under various acoustic and spatial conditions, and matches or surpasses dedicated spatial-extraction and enhancement baselines while offering substantially greater flexibility.
Nov 5, 2026

Design of a Speech Enhancement-Based Deep Semantic Autoencoder for Wireless Communications

Speech communication plays a crucial role in the Internet of Things and autonomous networks. However, traditional communication systems suffer from insufficient semantic extraction and limited noise suppression capability. To address these issues, we propose a Speech enhancement-based Deep Semantic Autoencoder (S-DSA) communication system for wireless communications. It consists of a dual-branch semantic encoder, a channel encoder, a channel decoder, and two semantic decoders, enabling joint source-channel coding for robust transmission. A key novelty lies in an adaptive semantic noise decoupling mechanism that disentangles target speech semantics from noise representations through an interaction module, thus maintaining semantic consistency under low SNR conditions. Furthermore, a joint training strategy optimizes speech reconstruction and noise estimation, while a lightweight testing strategy reduces complexity by disabling the noise-decoding path. The proposed S-DSA enables more effective separation between speech and noise, and yields higher robustness as well as better reconstruction quality. These findings highlight the potential of the proposed S-DSA as a practical solution for next-generation intelligent wireless communications.
Nov 1, 2026

Q-BWE: Quantum Chaos-Informed Speech Bandwidth Extension

High-frequency reconstruction in bandwidth extension (BWE) depends on joined interactions among nonlinear spectral, temporal, and perceptual features; assigning a separate discriminator to each such feature may be redundant and may fail to model their joint interactions. We introduce quantum-informed generative model Q-BWE, where we design a five-qubit Quantum Chaos Discriminator (QCD) instead of using separate discriminator for every features, for modeling nonlinear joined-feature interactions efficiently. Our QCD explores a compact five-qubit quantum-inspired inductive bias based on data re-uploading circuit, trainable controlled NOT (CNOT) rotations, quantum entangling gates, and statevector measurement map. Five features for five qubits are selected from a 26-dimensional speech-dynamics bank and are encoded by a two-layer statevector circuit, producing 31 measurements map of local behavior, correlations, mixedness, and Born entropy. QCD guides a Lattice-ConformerNeXt generator with FiLM conditioning and a cosine-masked residual pathway, while delayed-reward Thompson sampling adaptively controls its contribution. The circuit is simulated on classical hardware. Overall, Q-BWE achieves competitive performance on VCTK English, and MLS French datasets in multilingual and multi speaker setups under clean and noisy conditions across a broad set of evaluation metrics with approximately 1.72x fewer parameters (33.5M vs. 19.4M) compared to recent CIS-BWE baseline published in ACL.
Nov 1, 2026

High Frequency Safeguard and Weighted Objective Scoring Bandwidth Extension

Speech Bandwidth Extension (BWE) reconstructs missing high frequencies from a band-limited signal. Please note that perceptually important information is often concentrated in short high-frequency (HF) events. Therefore, an effective BWE requires both accurate HF generation and explicit control over where, how often, and in what spectral form HF energy should be produced. To address the above HF supervision, we present HFSWO, a compact complex-spectral Swin-based dual-stream adversarial framework with two newly proposed specialized discriminators: 1. High-Frequency Safeguard Discriminator (HFSD) converts excessive framewise HF-energy ratios into a differentiable safeguard-constraint penalty optimized through a projected primal--dual augmented-Lagrangian update; 2. Weighted Objective Scoring Discriminator (WOSD) learns an interpretable convex objective scoring over spectral flatness, centroid, and rolloff. Our rigorous experiments on VCTK English and MLS French datasets across multiple frequency ranges, including noisy conditions, show significantly improved perceptual scores with 74.6% fewer parameters than the existing baseline AP-BWE, establishing a new baseline in the BWE task in multilingual and multispeaker setups.
Oct 27, 2026

Perceptual Differences and Preferences in Speech Enhancement

Speech enhancement (SE) systems are commonly evaluated with objective metrics, with and without references, and with categorical subjective listening tests, all measuring speech quality and intelligibility. So far, it is unclear from such evaluations whether the processed speech is perceptually different or is preferred by everyday listeners. We investigate discrimination and preference as complementary dimensions of SE evaluation through crowdsourced triangle and pairwise preference tests across four SE systems, with comparisons to clean references. Results show that both listeners' ability to distinguish and whether they exhibit systematic preference vary across systems and substantially across acoustic conditions. Pairwise evaluations of objective metrics showed variable correlation with both discriminability and preference, motivating further studies into these novel aspects in speech enhancement evaluation.
Oct 26, 2026

The Voice on the Line

A transcriber listening through one microphone also hears everyone near it. We train causal voice isolators: masking transformers of 140M, 243M and 374M parameters, 20 ms of latency, enrolling the caller from the call itself. Short evaluation windows hide two failures that real calls expose: forgetting the caller when another voice talks alone for a while, and damaging a caller who is alone. Long calls with turns, memory beyond the attention window, real headset crosstalk and drifting caller levels address both, but each trades one against the other. WavLM-Large's own weights, made causal and trained with the model, move the trade: M with them (555M parameters) lowers the error on our development sets from 11.44% to 10.55%; a student trained from scratch to imitate WavLM does not, and w2v-BERT 2.0 and Whisper's encoder, pre-trained on far more audio, help less or not at all. A lone caller loses most at the start of a recording: no training call opened with someone already talking. Training on windows cut from longer calls lowers the error on clean development calls from 3.92% to 2.96%; this is our release. Live, the release makes 16.13% word errors on 60 s AMI windows, against 23.66% for VoxLock, the strongest open isolator, and 10.19% on its own AMI windows, against 11.71%; it is level on Krisp's benchmark (5.74% against 5.63%), and a caller who is alone still loses more words than with VoxLock (5.58% against 2.92% on clean test calls). Distilled into 22.1M parameters without WavLM, it runs live on one CPU thread inside a LiveKit agent, at 33% of a core per call in 20 ms frames, with 10.66% error on our development sets (the release 10.49%). We open-source the release and this CPU model, with a LiveKit plugin, at https://huggingface.co/collections/Scicom-intl/voice-isolator.
Oct 26, 2026

Realizing Low-Precision Storage and Sparse Execution in a Folded FPGA Speech Enhancer

Hardware-aware training often optimizes FPGA resources, latency, or bit operations, but folding changes what compression saves: shared arithmetic decouples precision, weight storage, and scheduled work. We derive training objectives from the measured costs of a folded streaming speech-enhancement accelerator. Mixed-precision weights reduce on-chip storage; structured channel sparsity removes entire operations from the engine schedule. We deploy a bit-exact implementation on the Kria KV260. The best sparse model eliminates 47.3% of convolution MACs, cutting frame time by 45% (1.86 to 1.03 ms) and dynamic energy by 60% (83 to 33 μμJ/frame) against a storage-priced model measured in the same session, with higher speech quality. Relative to the most efficient dense build measured separately, frame time falls by 42% and energy by 48%. Reducing weight width from 8 to 6 bits lowers dynamic power by 27% but changes frame time by less than 5%. Plain channel group-lasso removes more work than cycle-cost-weighted lasso. Thus, folded-accelerator training should price storage and removable work separately, rather than rely on a single bit-operation metric.
Oct 23, 2026

Noisy Target Training Meets Drifting for Weakly-Supervised Speech Enhancement

In speech enhancement, standard supervised training requires pairs of noisy and clean speech. Such pairs are typically created by artificially degrading clean speech, leading to a domain mismatch between synthetic training data and real recordings. Noisy target training (NyTT) mitigates this limitation by training an enhancement model to predict a noisy recording from a noisier version generated by adding extra noise. However, since the training target is noisy, NyTT might learn to leave residual noise in the enhanced output. In this paper, we address this issue in NyTT by incorporating a Sinkhorn drifting loss that matches the distribution of the enhanced speech to that of unpaired clean speech. The Sinkhorn drifting loss attracts enhanced speech toward high-density regions of the clean-speech distribution while repelling it from the current enhanced-speech distribution in latent space. This loss uses only unpaired clean speech rather than the clean counterpart of each degraded recording, preserving the weakly-supervised nature of NyTT. Experimental results on VoiceBank-DEMAND and CHiME-3 real recordings demonstrate the efficacy of the proposed combination.
Oct 22, 2026

When Do Self-Supervised Representations Help Speech Restoration?

Self-supervised learning (SSL) representations are widely used in speech restoration (SR), and the prevailing trend is SSL feature cleansing: fine-tuning the SSL to recover clean features from degraded speech, then conditioning the restoration model on them. However, the representational capability of SSL when encoding degraded speech has been underexplored, and we ask whether cleansing always benefits SR training. Across SSL encoders and perceptual quality proxies, we discover perceptual correlation collapse---under severe degradation, cleansing degraded SSL features toward clean ones misaligns with the SR objective. We hypothesize the restoration backbone can supply the SSL with clues to resolve this collapse, since its intermediate layers already produce partially restored features. We therefore propose Early Restoration Injection (ERI), which routes intermediate features from the restoration backbone into the SSL through a lightweight conditioning network. ERI excels over the unconditioned baseline and, on the VoiceFixer benchmark, achieves competitive performance among multiple baselines.
Oct 21, 2026

XingChen-UniFE: A Unified Generative Speech Front-End via Complementary Acoustic-Semantic Modeling

This paper presents XingChen-UniFE, a unified large audio language model (LALM) for general speech restoration (GSR) and target speaker extraction (TSE). The proposed model employs complementary semantic and acoustic representations through a dual-stream encoder, where a self-supervised learning model provides hierarchical semantic features and a learnable acoustic branch captures fine-grained structures from complex spectra. A bidirectional cross-attention module further facilitates information exchange between the two representations. The fused representation conditions the LALM for discrete speech token generation with task and speaker information. Experiments demonstrate strong performance on both GSR and TSE and robust generalization to unseen degradation conditions.
Oct 15, 2026

CF-AVSS: A Coarse-to-Fine Audio-Visual Speech Separation Framework with Target-Speaker Refinement and Post-Fusion for Real-World AVSE Challenge

Audio-visual speech enhancement (AVSE) leverages visual cues to recover target speech from mixtures corrupted by reverberation, noise, and overlapping speakers. We propose a coarse-to-fine two-stage AVSE framework. In the first stage, a progressive SPAV-TFGridNet is extended to two synchronized visual streams to produce visually ordered coarse estimates by progressively suppressing interfering speech while preserving target reverberation and shared background noise. In the second stage, a TFGridNet-based target speaker refinement network takes the original mixture and both coarse estimates as input to further recover target speech and improve signal quality. The coarse and refined estimates are finally combined using development-set-calibrated fusion. On the ISCSLP 2026 Real-World AVSE Challenge, our system tied for first place on Track 1 and ranked second on Track 2 in OVRL, while achieving the lowest CER on both tracks at 12.31% and 14.79%, respectively.
Oct 14, 2026

SNOAR: Self-training with Noise filtering, Observation adding, and Remixing for Speech Denoising Without Clean References

Unsupervised speech denoising methods such as MixIT and noisy-target training can be trained directly on noisy recordings, avoiding the sim-to-real gap of synthetic training data, but remain well below supervised performance. Teacher--student self-training, as in RemixIT, narrows this gap by remixing a teacher's estimates into new mixtures that supervise a student; however, since the student is trained only toward the teacher's estimates, teacher errors, in particular over-suppressed speech, are never recovered and propagate as the teacher is updated from the student. We propose SNOAR, a self-training framework that corrects the teacher's estimates before use: the noise leaked into its speech estimate is removed by differential noise filtering (DNF), and part of the noisy observation is added back in to restore suppressed speech and attenuate artifacts. We show that the two corrections depend on each other, as without DNF the loop collapses to the identity mapping. Compared to RemixIT, SNOAR is more stable along training and outperforms it in signal-based, perceptual, and word error rate metrics, reaching 97% of the SI-SDR improvement of a fully-supervised topline on WHAM!; on real CHiME-3 recordings, it improves every metric of three different teachers, while RemixIT degrades their perceptual quality.
Oct 13, 2026

Multimodal Contrastive Representation Learning Based Voice Conversion for Electrolaryngeal Speech Enhancement

Electrolaryngeal (EL) speech suffers from severe acoustic mismatches with normal speech, resulting in degraded naturalness and intelligibility. These mismatches also make sequence-to-sequence voice conversion struggle to learn robust intermediate representations, hindering EL-speech-to-normal-speech conversion (EL2SP) performance. To address this, we propose a novel multimodal contrastive representation learning framework to 1) achieve richer multimodal-representation learning, and 2) facilitate the transferability of multimodal representations to speech-only EL2SP. First, EL speech and text representations are fused to learn multimodal representations for target speech generation, during which cross-modal contrastive learning is employed to encourage a semantically consistent speech--text representation space. Next, the fused representations are treated as a fixed target to guide speech-only representation learning through cross-stage contrastive learning, enabling multimodal knowledge to be transferred to the EL2SP task while preserving architectural and inference simplicity. Experimental results demonstrate that the proposed method consistently outperforms both the speech-only and multimodal-based baselines.
Oct 8, 2026

Efficient Audio-Visual Speech Enhancement via Reinforcement Learning-Based Neural Architecture Search

Audio-visual speech enhancement (AVSE) improves speech intelligibility in noisy environments but often relies on computationally expensive models unsuitable for resource-constrained edge devices. This paper introduces a reinforcement learning-driven neural architecture search (RL-NAS) framework to discover compact, efficient AVSE architectures. The search space preserves temporal alignment between audio and visual streams, while a novel reward function jointly optimises speech enhancement performance, measured by SI-SNR improvement, and model compactness. Using proximal policy optimisation with constraint-aware action masking, the framework evaluates only 35 candidate architectures and identifies a model with 2.9 million parameters, compared with 22 million in the baseline. The proposed architecture achieves a 14.5 dB SI-SNR improvement, only 0.7 dB below the baseline, while reducing model size by approximately nine-fold. Evaluation on commercial mobile CPUs demonstrates a 4.3-fold reduction in real-time factor, supporting practical on-device deployment. Compared with random search, RL-NAS achieves 92% higher search efficiency. Architectural analysis reveals that shallower but wider temporal convolutional blocks and aggressive visual pathway pruning can maintain enhancement performance while substantially reducing computational requirements. These findings demonstrate the potential of reinforcement learning-guided architecture search for developing lightweight, efficient and deployable multimodal speech enhancement systems.
Oct 8, 2026

SEER: Source-Conditioned Emotion Enhancement via Retrieval for Cochlear-Implant Speech

Cochlear implants (CIs) restore speech access but weaken cues needed for vocal emotion recognition. Prior CI-oriented enhancement requires parallel normal/strong recordings and intensity labels. We propose SEER, a retrieval-based framework that learns which same-emotion reference helps each source remain recognizable after CI processing. A source-conditioned retriever learns CI-aware utility from sampled emotional voice conversion outcomes, while uncertainty-guided exploration avoids exhaustive pair evaluation; neither parallel recordings nor intensity labels are required. SEER improves Source macro-F1 at N8 by 7.30 points on RAVDESS and 11.66 points on ESD, with significant ESD gains across N4/N8/N16. Sixteen-listener RAVDESS gains are significant across all conditions. Exhaustive analysis finds an aggregate benefit from stronger references but little effect from matching gender or content.
Oct 6, 2026

SwinDS-BWE: A Parameter-Efficient Swin-1D Lattice Generator with Decision-Science Discriminators for Speech Bandwidth Extension

We propose SwinDS-BWE, a decision-science-inspired bandwidth extension (BWE) model with two coupled contributions: (1) Swin Transformers with lattice-style cross-stream interaction yield locality-aware modeling with linear-in-sequence per-window attention cost and reduce the prior AP-BWE generator size 0.5x from 33M to 17M. (2) Three novel lightweight decision-science-inspired discriminators augment AP-BWE's performance: a CVaR discriminator tail-pools activations to emphasize worst high-frequency (HF) segments, a Chance-Constraint HF discriminator penalizes excessive HF energy via a differentiable barrier, and a Multi-Criteria Utility discriminator learns convex style weights over spectral criteria. SwinDS-BWE surpasses prior AP-BWE with a 30x smaller discriminator (42.3M vs. 1.36M) and higher fidelity on English and French datasets. This work shows that decision-science-inspired critics can supervise BWE to reduce discriminator size.
Oct 5, 2026

AuraSE: Low-Hallucination Generative Speech Enhancement via Multimodal Flow Matching and Inference Policy Optimization

Generative speech enhancement models can produce cleaner and more natural-sounding speech than conventional discriminative approaches, but may hallucinate by changing speech content or speaker identity, even with transcript conditioning. We present AuraSE, a flow-matching framework that addresses hallucination through complementary modality and inference designs. First, a double-stream-to-single-stream multimodal Diffusion Transformer (MMDiT) allows transcript and acoustic representations to interact while preserving a dedicated pathway for the degraded input. Second, we find that the best decoder configuration, governed by guidance scale, sampling temperature, and step count, varies substantially across utterances. This observation motivates Inference Policy Optimization (IPO), an online, on-policy preference optimization method. IPO generates multiple candidates from the current model under different inference configurations, ranks them with a multi-objective reward, and learns from their relative preferences. AuraSE-IPO ranks first on 11 of 12 metrics across the synthetic test sets and obtains the highest DNSMOS and blind-listening scores among the evaluated systems on the real DNS blind test set. At deployment, it uses a fixed 1010-step ODE decoder without classifier-free guidance (CFG) or per-utterance configuration search.
Sep 29, 2026

Long-Term Memory-Guided Enhancement for Target Perception in Audio-Language Models

Audio large language models (ALLMs) can reason about the content of audio recordings to perform complex tasks. However, these capabilities usually collapse in real-world environments when background noise and competing sources mix the target sound. Inspired by long-term memory in human listening, we propose Long-Term Memory-Guided Audio Enhancement (LTM-AE) to improve selective target perception by refining the audio representations of ALLMs without training. LTM-AE extracts representations in hidden states from separate clean reference recordings as long-term memory for each category, guiding enhancement toward a user-specified listening target. We reconstruct incoming audio tokens in the selected category long-term memory and interpolate the reconstructions with the original tokens before language backbone decoding. This interpolation controls the influence of stored auditory experience while keeping all ALLM parameters fixed. Diagnostic readouts across twenty sound categories and three ALLMs show that LTM-AE strengthens responses to a specified target amid three interfering sources. Averaged over constrained and free-form classification, accuracy gains over raw mixtures range from 29.53 to 46.15 percentage points across multiple open source models. For speech content recovery, LTM-AE with an additional learned token-level gate reduces Qwen2-Audio's word error rate from 23.07% to 14.77%. This work takes an initial step toward using principles of human long-term memory to enhance ALLMs for real-world listening. Our code is available at https://github.com/aynlp/ltm-audio-code
Sep 28, 2026

Perceptual Quality Loss or Loss of Perceptual Quality?

Contemporary deep speech enhancement (SE) models are often trained with specific auxiliary terms in the loss function as a way to improve their performance in terms of perceptual metrics. Nevertheless, a higher score on a perceptual metric does not necessarily correlate with an improved listening experience. Through objective and subjective experiments, we assess the performance of SE models trained with two different types of auxiliary PESQ loss terms. The numerical evaluation on a suite of standard metrics suggests that, while models optimized for PESQ naturally obtain higher PESQ scores in the test set, for most other metrics the scores do not significantly change. In some cases, the PESQ loss even results in worse PESQ scores on mismatched data. A formal listening experiment reveals that the models without a PESQ loss were generally preferred over models that include it, across all settings. Finally, we analyze the relative importance of PESQ in the composite metrics CSIG, CBAK and COVL, and find that PESQ dominates all of them. Our study highlights the perils of over-reliance on PESQ and stresses the importance of a complete evaluation procedure for SE.
Sep 28, 2026

Domain-Incremental Learning for Generative Speech Enhancement

We propose a domain-incremental learning framework for generative speech enhancement (SE) that learns from a sequence of datasets or domains recorded under diverse acoustic conditions. Fine-tuning a pretrained model on continuously evolving domains leads to catastrophic forgetting of previously acquired knowledge, while zero-shot generalization often fails to adequately adapt to unseen domains. To address these challenges, we first develop a novel language model-based generative SE model that we then use as a pretrained backbone and incrementally adapt it to acoustically mismatched domains using lightweight domain-specific Low-Rank Adaptation. The proposed framework enables the model to acquire enhancement capabilities for new domains while preserving performance on previously learned domains. Evaluated on four heterogeneous speech datasets, our approach effectively adapts to new domains without forgetting previously learned domains.
Sep 28, 2026

Unsupervised Speech Enhancement via Drifting

This paper addresses unsupervised speech enhancement in the unpaired setting using drifting methods, where training relies on separate collections of degraded and clean audio without corresponding pairs. While recent drifting approaches enable unpaired training, they do so at a heavy cost: because the objective optimizes only a marginal prior over clean speech, the enhancer gradually loses the input's linguistic content and speaker identity. To fix this, we introduce input-conditioned drifting. We preserve the pull of the clean corpus while re-tethering the output to the degraded input via two mechanisms: an anchor encoder supplies the missing likelihood by pulling toward the input's features, and a key encoder conditions the prior by re-weighting retrieved frames. Neither requires labels or paired data. Using a training-free encoder selection criterion, Word Error Rate on VoiceBank-DEMAND falls to 10.1% (unprocessed: 11.7%), speaker similarity recovers from 0.490 to 0.879, and the recipe transfers in part to dereverberation on WSJ0-REVERB: content improves, rendering quality does not.
Sep 28, 2026

Prox-Friendly Log-Magnitude Prior on Complex-Valued Signal

The logarithmic transform is essential in audio signal processing since human auditory perception is approximately logarithmic with respect to magnitude. However, directly incorporating prior knowledge about signals (e.g., harmonic structure) in the log-magnitude domain into optimization problems solved by standard proximal splitting algorithms remains challenging. To address this issue, this paper proposes a novel regularizer termed EPILOG (Exponential Penalty for Imposing priors on LOG-magnitude). EPILOG indirectly imposes prior knowledge on the log-magnitude of a complex-valued signal through regularization of an auxiliary variable that is shown to be linked with the log-magnitude. Furthermore, we derive its variable-wise proximity operators and develop a proximal splitting algorithm using these operators. Experiments on speech dereverberation demonstrate the effectiveness of the proposed regularizer, particularly in promoting cepstral-domain sparsity.
Sep 27, 2026

Transformer-based Neural Beamforming for Real-Time Speech Enhancement on Smart Low-Power Hearable Devices

Accurate, efficient, and low-latency spatial beamforming is a key component in emerging smart hearable devices, enhancing speech while suppressing noise and interference. However, handling multiple input sources under strict real-time constraints poses significant challenges for the low-power, resource-constrained microcontroller units (MCUs) used in hearables. We present an optimized methodology for the real-time execution of a neural-network-based minimum variance distortionless response (MVDR) beamformer on MCUs. Using six microphones and a three-stage mixed-precision scheme (float32 MVDR, int8 CNN, float16 Transformer), the pipeline pairs a CNN that estimates the MVDR weights with a lightweight Transformer that applies a per-frame correction. By time-slicing weight estimation with beamforming, it achieves a 15ms per-frame latency while refreshing a complete set of CNN-derived weights every 564ms. The deployed mixed-precision pipeline attains a short-time objective intelligibility (STOI) of 97.65%, a scale-invariant signal-to-noise ratio (SI-SNR) of 20.26dB, and a wideband PESQ of 3.676 at an average power of 45.9mW. A speech activity detection (SAD) module (98.5% accuracy, 0.62mJ per inference) bypasses the pipeline during silence; under realistic deployment conditions, the system exceeds the 16h all-day target on a 100mAh battery, with an estimated lifetime of up to ∼\sim20h. To our knowledge, this is the first real-time multi-channel Transformer-based neural beamforming pipeline deployed on an MCU-class device.
Sep 26, 2026

WhisperVC-AV: Audio-Visual Content Restoration for Noise-Robust Whisper-to-Normal Voice Conversion

Environmental noise obscures the content cues needed for whisper-to-normal conversion. We propose WhisperVC-AV, which restores acoustic content features while leaving the original WhisperVC conversion module unchanged. Its context-guided restoration module combines temporal acoustic context with synchronized lip features through attention and gated residual correction. Experiments on AISHELL6-Whisper show lower character error rates (CERs) than WhisperVC across three ASR systems, on clean speech and under all six signal-to-noise ratio (SNR) conditions with MUSAN noise. The largest gains occur at 0 dB SNR, where Qwen3-ASR CER falls from 36.29% to 28.88%. WhisperVC-AV also improves predicted speech quality while maintaining speaker similarity. The CER gains extend to unseen background noise without retraining, while visual controls support the use of utterance-specific lip cues. Audio examples are available on our demo page.
Sep 25, 2026

Room Impulse Response Embeddings for Speech Enhancement in Noisy and Reverberant Environments

We propose a self-supervised approach for learning room impulse response (RIR) representations from single-channel noisy-reverberant speech. It consists of first training on reverberant data, then on noisy-reverberant data, and finally with a teacher-student approach, where the student learns to replicate the teacher's embeddings when given a noisy version of the reverberant input. We assess their representational capabilities by estimating acoustic room parameters from them. Conditioning a discriminative speech enhancement model on the derived embeddings yields consistent gains across all evaluated metrics, including downstream word error rate, for both reverberant and noisy-reverberant speech.
Sep 24, 2026

Adapting Personalized Speech Enhancement for Low-Latency Audio-Visual Target-Speaker Extraction

Online audio-visual target-speaker extraction aims to remove competing voices while preserving speech quality and bounding lookahead. Existing extractors are built and evaluated for separation on synthetic mixtures, leaving listening quality and meeting behavior largely untested. We introduce Audio-Visual Personalized Voice Quality Enhancement (AV-PVQE), which approaches these requirements from the other direction. We start from a personalized speech enhancement model that reconstructs a requested voice at high quality but confuses the target in 46% of two-speaker mixtures despite clean enrollment. Adding mouth features at its speaker-conditioning input and jointly fine-tuning the visual and reconstruction networks reduces this rate to 1.6%, with no future frames and 20 ms of algorithmic delay. Compared with an online autoregressive audio-visual extractor, AV-PVQE yields separation gains on two synthetic benchmarks and larger gains on recorded meetings, and keeps its advantage on excerpts with more speakers than the fine-tuning mixtures. In personalized P.835 listening tests on two meeting corpora, it improves overall quality over this extractor by 0.57 and 0.63 MOS, with similar mean rating relative to the starting model. Preservation and rejection tests show that it keeps the target intact when no competing voice is present and suppresses competing speech when the target is absent.
Sep 24, 2026

Does per-frame early exit pay? A compute-matched study of dynamic depth for on-device speech enhancement

Deep learning-based speech enhancement is increasingly deployed on-device in hearing aids, headsets, and earbuds. Most of these devices, however, can only accelerate static int8 graphs, so a depth-varying network must be implemented as several graphs, orchestrated by a policy. In this paper, we supervise every intermediate depth of one causal model, then we fine-tune its output heads to guarantee that deeper outputs are never worse than shallower ones. Using this training protocol, we can derive a family of static models that are more Pareto-efficient than their equivalently-sized counterparts trained from scratch on the same budget. Specifically, we achieve up to 0.11 higher PESQ for equivalent compute, and match the best PESQ at 30% less compute. We then quantize the models to int8 and measure the latency-quality frontier on an STM32N6 microcontroller. On VoiceBank-DEMAND, the dynamic enhancer lies on the same frontier as the static models, rather than trading quality for dynamic execution. Running the policy on the companion Cortex-M55 takes only 26 μμs per frame, while splitting the enhancer into separate NPU graphs adds 2.2% latency overhead. The cost of dynamic execution is therefore small.
Sep 24, 2026

Beyond Model Size: Redesigning LiSenNet for embedded speech enhancement

Deploying real-time speech enhancement on resource-constrained devices requires meeting strict latency, memory, and energy constraints. Microcontroller NPUs can accelerate neural inference under these constraints, but only through a restricted set of operators in static, integer-quantized graphs. Recent speech-enhancement networks have reduced parameter counts and MACs to levels nominally suitable for microcontrollers, but their operators and execution patterns often remain incompatible with restricted NPUs. We address this gap by redesigning LiSenNet, a 37k parameter sub-band dual-path model, for the STM32N6570-DK Neural-ART accelerator. We replace its recurrent bottleneck with convolutional frequency and temporal mixers, reformulate unsupported operations as static int8-compatible primitives, and use bounded decoder activations to preserve quality after quantization. On VoiceBank-DEMAND, the final NPU-compatible model matches or exceeds the recurrent LiSenNet baseline, reaching PESQ 3.08 versus 3.01 in FP32 and 3.01 versus 2.93 in int8. Deployed on a microcontroller, it processes each 16 ms input hop in 4.83 ms, corresponding to a real-time factor of 0.30. Stateless receptive-field recomputation is an order of magnitude slower at the same frame rate despite higher accelerator utilization. These results show that parameter count and operator compatibility, quantization range, and persistent streaming state must be co-designed to achieve efficient real-time speech enhancement on restricted NPUs.
Sep 24, 2026

Transcript-Supervised Post-Training of Generative Speech Enhancement on Real Recordings via Reinforce Adjoint Matching

We adapt Reinforce Adjoint Matching (RAM), a reward-based post-training method, to generative speech enhancement (SE). Starting from a pretrained SE model, RAM tilts the model's conditional distribution toward outputs with higher reward. During training, the current model generates enhanced speech on-policy, evaluates each generated endpoint with a potentially non-differentiable reward, and analytically re-noises the endpoint to construct inputs for a reward-guided regression objective. This enables post-training directly on real recordings using weak supervision, such as text transcripts, without requiring paired clean speech targets or reward gradients. We investigate word error rate (WER)-based post-training and whether recognition performance can be improved without compromising perceptual speech quality. Experiments on real CHiME-4 recordings reduce WER by 5.08 percentage points relative to pretrained FlowSE without reducing any of the reported non-intrusive speech quality metrics. A subjective listening test at the default reward scale finds no statistically significant preference between the post-trained and pretrained models.