{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T12:07:29Z","timestamp":1789387649416,"version":"build-2803163510"},"reference-count":61,"publisher":"Acoustical Society of America (ASA)","issue":"3","content-domain":{"domain":["pubs.aip.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2026,9,1]]},"abstract":"<jats:p>Full-band speech restoration aims to recover high-fidelity speech from signals degraded by noise, reverberation, bandwidth limitation, or their combinations. Existing methods are typically designed for a specific type of degradation, which limits their flexibility when multiple distortions coexist. To address this problem, this paper proposes a dual-stream versatile speech enhancement network (DSVN)\u2014a unified task-conditioned system for full-band speech restoration. The model introduces a dual-stream spectral block to jointly model frequency-wise spectral structures and temporal dynamics. A band-refill decoder is further designed to recover missing or degraded high-frequency components. Task-conditioned modulation is introduced to enable the same network to adapt to different objectives. Evaluation results demonstrate that DSVN improves restoration quality under both single- and mixed-degradation scenarios. The model shows clear advantages in non-intrusive quality assessment, spectral distance, and automatic speech recognition evaluation, indicating that the enhanced speech is not only cleaner but also more useful for downstream systems. Additional analysis of internal responses and ablation studies provides evidence that the proposed modules contribute to different aspects of restoration.<\/jats:p>","DOI":"10.1121\/10.0046397","type":"journal-article","created":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T11:57:15Z","timestamp":1789387035000},"page":"2272-2287","update-policy":"https:\/\/doi.org\/10.1063\/aip-crossmark-policy-page","source":"Crossref","is-referenced-by-count":0,"title":["A dual-stream spectral network for versatile full-band speech enhancement"],"prefix":"10.1121","volume":"160","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-1758-3746","authenticated-orcid":false,"given":"Junkang","family":"Yang","sequence":"first","affiliation":[{"name":"Integrated Graduate School of Medicine, Engineering, and Agricultural Sciences, University of Yamanashi 1 , Kofu,","place":["Japan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7717-8312","authenticated-orcid":false,"given":"Hiromitsu","family":"Nishizaki","sequence":"additional","affiliation":[{"name":"Integrated Graduate School of Medicine, Engineering, and Agricultural Sciences, University of Yamanashi 1 , Kofu,","place":["Japan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-1382-8962","authenticated-orcid":false,"given":"Chee Siang","family":"Leow","sequence":"additional","affiliation":[{"name":"Integrated Graduate School of Medicine, Engineering, and Agricultural Sciences, University of Yamanashi 1 , Kofu,","place":["Japan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2069-0390","authenticated-orcid":false,"given":"Hongqing","family":"Liu","sequence":"additional","affiliation":[{"name":"Chongqing Key Lab of Mobile Communications Technology, Chongqing University of Posts and Telecommunications 2 , Chongqing,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1056-7660","authenticated-orcid":false,"given":"Lu","family":"Gan","sequence":"additional","affiliation":[{"name":"College of Engineering, Design and Physical Science, Brunel University 3 , London,","place":["United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"231","published-online":{"date-parts":[[2026,9,14]]},"reference":[{"key":"2026091407571437200_c1","first-page":"1","article-title":"HIFI++: A unified framework for bandwidth extension and speech enhancement","year":"2023"},{"key":"2026091407571437200_c2","first-page":"1","article-title":"CPT-boosted wav2vec2.0: Towards noise robust speech recognition for classroom environments","year":"2025"},{"issue":"5","key":"2026091407571437200_c3","doi-asserted-by":"publisher","first-page":"808","DOI":"10.1080\/01969722.2025.2577925","article-title":"Deep learning-based chatbox with attention bidirectional recurrent network-based puffer fish search for improved accessibility and cost reduction","volume":"57","year":"2026","journal-title":"Cybern. Syst."},{"key":"2026091407571437200_c4","article-title":"Description on IEEE ICME 2024 grand challenge: Semi-supervised acoustic scene classification under domain shift","year":"2024"},{"issue":"12","key":"2026091407571437200_c5","doi-asserted-by":"publisher","first-page":"1036","DOI":"10.1109\/LSP.2007.906208","article-title":"Cepstral smoothing of spectral filter gains for speech enhancement without musical noise","volume":"14","year":"2007","journal-title":"IEEE Signal Process. Lett."},{"key":"2026091407571437200_c6","first-page":"888","article-title":"Universal speech enhancement with regression and generative mamba","year":"2025"},{"issue":"6","key":"2026091407571437200_c7","doi-asserted-by":"publisher","first-page":"1505","DOI":"10.1109\/JSTSP.2022.3188113","article-title":"WavLM: Large-scale self-supervised pre-training for full stack speech processing","volume":"16","year":"2022","journal-title":"IEEE J. Sel. Top. Signal Process."},{"key":"2026091407571437200_c8","first-page":"833","article-title":"A deformable convolution GAN approach for speech dereverberation in cochlear implant users","year":"2025"},{"key":"2026091407571437200_c50","doi-asserted-by":"publisher","DOI":"10.1080\/01969722.2025.2521708","article-title":"Enhanced transfer learning-based CNN for abnormal human activity detection in video surveillance using spatial-temporal features","year":"2025","journal-title":"Cybern. Syst."},{"key":"2026091407571437200_c9","doi-asserted-by":"publisher","first-page":"725","DOI":"10.1109\/OJSP.2024.3378602","article-title":"ICASSP 2023 deep noise suppression challenge","volume":"5","year":"2024","journal-title":"IEEE Open J. Signal Process."},{"key":"2026091407571437200_c10","doi-asserted-by":"publisher","first-page":"1725","DOI":"10.1109\/TASLPRO.2026.3675815","article-title":"A harmonic-coupled generative adversarial network for speech super-resolution in low bandwidth scenarios","volume":"34","year":"2026","journal-title":"IEEE Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c11","unstructured":"github.com (2026a). See https:\/\/github.com\/BiolabHHU\/HCGAN for HCGAN."},{"key":"2026091407571437200_c12","unstructured":"github.com (2026b). See https:\/\/github.com\/haoheliu\/voicefixer for VoiceFixer."},{"key":"2026091407571437200_c13","unstructured":"github.com (2026c). See https:\/\/github.com\/JaeBinCHA7\/DEMUCS-for-Speech-Enhancement for DEMUCS-for-Speech-Enhancement."},{"key":"2026091407571437200_c14","unstructured":"github.com (2026d). See https:\/\/github.com\/slp-rl\/aero for AERO."},{"key":"2026091407571437200_c15","unstructured":"github.com (2026e). See https:\/\/github.com\/sp-uhh\/sgmse for Speech Enhancement and Dereverberation with Diffusion-based Generative Models."},{"key":"2026091407571437200_c16","unstructured":"github.com (2026f). See https:\/\/github.com\/urgent-challenge\/urgent2026_challenge_track1 for the official baseline, dataset, and evaluation scripts for the ICASSP 2026 URGENT challenge."},{"issue":"1","key":"2026091407571437200_c17","doi-asserted-by":"publisher","first-page":"034013","DOI":"10.1155\/2007\/34013","article-title":"Inverse filtering for speech dereverberation less sensitive to noise and room transfer function fluctuations","volume":"2007","year":"2007","journal-title":"EURASIP J. Adv. Signal Process."},{"key":"2026091407571437200_c18","unstructured":"Ho, J., Kalchbrenner, N., Weissenborn, D., and Salimans, T. (2019). \u201cAxial attention in multidimensional transformers,\u201d arXiv:1912.12180."},{"key":"2026091407571437200_c19","first-page":"1501","article-title":"Speech bandwidth extension by improved codebook mapping towards increased phonetic classification","year":"2005"},{"issue":"1","key":"2026091407571437200_c20","doi-asserted-by":"publisher","first-page":"229","DOI":"10.1109\/TASL.2007.911054","article-title":"Evaluation of objective quality measures for speech enhancement","volume":"16","year":"2008","journal-title":"IEEE Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c21","first-page":"3829","article-title":"HD-DEMUCS: General speech restoration with heterogeneous decoders","year":"2023"},{"key":"2026091407571437200_c22","unstructured":"Kingma, D. P., and Ba, J. (2017). \u201cAdam: A method for stochastic optimization,\u201d arXiv:1412.6980."},{"issue":"1","key":"2026091407571437200_c23","doi-asserted-by":"publisher","first-page":"7","DOI":"10.1186\/s13634-016-0306-6","article-title":"A summary of the reverb challenge: State-of-the-art and remaining challenges in reverberant speech processing research","volume":"2016","year":"2016","journal-title":"EURASIP J. Adv. Signal Process."},{"issue":"4","key":"2026091407571437200_c24","doi-asserted-by":"publisher","first-page":"534","DOI":"10.1109\/TASL.2008.2009015","article-title":"Suppression of late reverberation effect on speech signal using long-term multiple-step linear prediction","volume":"17","year":"2009","journal-title":"IEEE Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c25","first-page":"1","article-title":"Generative speech foundation model pretraining for high-quality speech extraction and restoration","year":"2025"},{"key":"2026091407571437200_c26","article-title":"MelGAN: Generative adversarial networks for conditional waveform synthesis","volume-title":"Proceedings of the 33rd International Conference on Neural Information Processing Systems","year":"2019"},{"key":"2026091407571437200_c27","doi-asserted-by":"publisher","first-page":"1200","DOI":"10.1109\/TASLPRO.2026.3662487","article-title":"Ednet: A versatile speech enhancement framework with gating mamba mechanism and phase shift-invariant training","volume":"34","year":"2026","journal-title":"IEEE Trans. Audio. Speech Lang. Process."},{"issue":"1","key":"2026091407571437200_c28","doi-asserted-by":"publisher","first-page":"656","DOI":"10.1109\/TAFFC.2025.3629970","article-title":"Describe where you are: Improving noise-robustness for speech emotion recognition with text description of the environment","volume":"17","year":"2026","journal-title":"IEEE Trans. Affect. Comput."},{"key":"2026091407571437200_c29","doi-asserted-by":"publisher","first-page":"2724","DOI":"10.1109\/TASLP.2023.3294692","article-title":"Storm: A diffusion-based stochastic regeneration model for speech enhancement and dereverberation","volume":"31","year":"2023","journal-title":"IEEE\/ACM Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c30","first-page":"21919","article-title":"ICASSP 2026 urgent speech enhancement challenge","year":"2026"},{"key":"2026091407571437200_c31","first-page":"1","article-title":"Less is more: Data curation matters in scaling speech enhancement","year":"2025"},{"key":"2026091407571437200_c32","first-page":"4232","article-title":"VoiceFixer: A Unified Framework for High-Fidelity Speech Restoration","year":"2022"},{"key":"2026091407571437200_c33","doi-asserted-by":"publisher","first-page":"236","DOI":"10.1109\/TASLP.2024.3519881","article-title":"Towards high-quality and efficient speech bandwidth extension with parallel amplitude and phase prediction","volume":"33","year":"2025","journal-title":"IEEE Trans. Audio. Speech Lang. Process."},{"issue":"2","key":"2026091407571437200_c34","doi-asserted-by":"publisher","first-page":"225","DOI":"10.1080\/01969722.2025.2521703","article-title":"Optimizing Alzheimer's disease detection with dual attention CNN and gooseneck barnacle search","volume":"57","year":"2026","journal-title":"Cybern. Syst."},{"key":"2026091407571437200_c35","first-page":"1","article-title":"AERO: Audio super resolution in the spectral domain","year":"2023"},{"key":"2026091407571437200_c36","first-page":"2127","article-title":"NISQA: A deep CNN-self-attention model for multidimensional speech quality prediction with crowdsourced datasets","year":"2021"},{"key":"2026091407571437200_c37","doi-asserted-by":"publisher","first-page":"5092","DOI":"10.1109\/TASLP.2024.3507566","article-title":"Blind audio bandwidth extension: A diffusion-based zero-shot approach","volume":"32","year":"2024","journal-title":"IEEE\/ACM Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c38","unstructured":"Nagel, F., Disch, S., and Rettelbach, N. (2009). \u201cA phase vocoder driven bandwidth extension method with novel transient handling for audio codecs,\u201d https:\/\/publica.fraunhofer.de\/handle\/publica\/364102 (Last viewed August 1, 2026)."},{"key":"2026091407571437200_c39","doi-asserted-by":"crossref","unstructured":"Perez, E., Strub, F., de Vries, H., Dumoulin, V., and Courville, A. (2017). \u201cFilm: Visual reasoning with a general conditioning layer,\u201d arXiv:1709.07871.","DOI":"10.1609\/aaai.v32i1.11671"},{"key":"2026091407571437200_c40","first-page":"28492","article-title":"Robust speech recognition via large-scale weak supervision","year":"2023"},{"key":"2026091407571437200_c41","first-page":"2492","article-title":"The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, subjective testing framework, and challenge results","year":"2020"},{"key":"2026091407571437200_c42","doi-asserted-by":"publisher","first-page":"2351","DOI":"10.1109\/TASLP.2023.3285241","article-title":"Speech enhancement and dereverberation with diffusion-based generative models","volume":"31","year":"2023","journal-title":"IEEE\/ACM Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c43","first-page":"4873","article-title":"EARS: An anechoic fullband speech dataset benchmarked for speech enhancement and dereverberation","year":"2024"},{"key":"2026091407571437200_c44","first-page":"749","article-title":"Perceptual evaluation of speech quality (PESQ)\u2014A new method for speech quality assessment of telephone networks and codecs","year":"2001"},{"key":"2026091407571437200_c45","first-page":"863","article-title":"TS-URGENet: A three-stage universal robust and generalizable speech enhancement network","year":"2025"},{"key":"2026091407571437200_c46","first-page":"858","article-title":"Interspeech 2025 URGENT Speech Enhancement Challenge","year":"2025"},{"key":"2026091407571437200_c47","doi-asserted-by":"publisher","first-page":"104408","DOI":"10.1016\/j.dsp.2024.104408","article-title":"Time domain speech enhancement with CNN and time-attention transformer","volume":"147","year":"2024","journal-title":"Digit. Signal Process."},{"key":"2026091407571437200_c48","first-page":"629","article-title":"Speech enhancement based on a priori signal to noise estimation","year":"1996"},{"issue":"7","key":"2026091407571437200_c49","doi-asserted-by":"publisher","first-page":"2125","DOI":"10.1109\/TASL.2011.2114881","article-title":"An algorithm for intelligibility prediction of time\u2013frequency weighted noisy speech","volume":"19","year":"2011","journal-title":"IEEE Trans. Audio. Speech Lang. Process."},{"key":"2026091407571437200_c51","first-page":"146","article-title":"Investigating RNN-based speech enhancement methods for noise-robust Text-to-Speech","year":"2016"},{"key":"2026091407571437200_c52","first-page":"1368","article-title":"WHAM!: Extending speech separation to noisy environments","year":"2019"},{"key":"2026091407571437200_c53","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1016\/j.specom.2023.02.006","article-title":"CASE-Net: Integrating local and non-local attention operations for speech enhancement","volume":"148","year":"2023","journal-title":"Speech Commun."},{"key":"2026091407571437200_c54","unstructured":"Yamagishi, J., Veaux, C., and MacDonald, K. (2019). \u201cCSTR VCTK Corpus: English multi-speaker corpus for CSTR voice cloning toolkit (version 0.92),\u201d The Rainbow Passage, International Dialects of English Archive, http:\/\/web.ku.edu\/\u223cidea\/readings\/rainbow.htm."},{"key":"2026091407571437200_c55","doi-asserted-by":"publisher","first-page":"101959","DOI":"10.1016\/j.csl.2026.101959","article-title":"A dual-branch parallel network for speech enhancement and restoration","volume":"100","year":"2026","journal-title":"Comput. Speech Lang."},{"issue":"5","key":"2026091407571437200_c56","doi-asserted-by":"publisher","first-page":"3143","DOI":"10.1121\/10.0034364","article-title":"Spectral network based on lattice convolution and adversarial training for noise-robust speech super-resolution","volume":"156","year":"2024","journal-title":"J. Acoust. Soc. Am."},{"key":"2026091407571437200_c57","first-page":"1171","article-title":"Speech super resolution and noise suppression system using a two-stage neural network","year":"2024"},{"key":"2026091407571437200_c58","first-page":"2483","article-title":"High fidelity speech enhancement with band-split RNN","year":"2023"},{"key":"2026091407571437200_c59","first-page":"1526","article-title":"LibriTTS: A corpus derived from LibriSpeech for text-to-speech","year":"2019"},{"key":"2026091407571437200_c60","first-page":"9122","article-title":"Multi-scale temporal frequency convolutional network with axial attention for speech enhancement","year":"2022"},{"key":"2026091407571437200_c61","doi-asserted-by":"publisher","first-page":"109574","DOI":"10.1016\/j.apacoust.2023.109574","article-title":"Dual branch deep interactive UNet for monaural noisy-reverberant speech enhancement","volume":"212","year":"2023","journal-title":"Appl. Acoust."}],"container-title":["The Journal of the Acoustical Society of America"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.aip.org\/asa\/jasa\/article-pdf\/160\/3\/2272\/21138556\/2272_1_10.0046397.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.aip.org\/asa\/jasa\/article-pdf\/160\/3\/2272\/21138556\/2272_1_10.0046397.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T11:57:23Z","timestamp":1789387043000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.aip.org\/jasa\/article\/160\/3\/2272\/3404435\/A-dual-stream-spectral-network-for-versatile-full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,1]]},"references-count":61,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2026,9,1]]}},"URL":"https:\/\/doi.org\/10.1121\/10.0046397","relation":{},"ISSN":["1520-8524"],"issn-type":[{"value":"1520-8524","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,9]]},"published":{"date-parts":[[2026,9,1]]}}}