{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/rhr-net-a-residual-hourglass-recurrent-neural","title":"RHR-Net: A Residual Hourglass Recurrent Neural Network for Speech Enhancement","arxiv_id":"1904.07294","date":"2019-04-15","proceeding":null,"authors":["Jalal Abdulbaqi","Yue Gu","Ivan Marsic"],"abstract":"Most current speech enhancement models use spectrogram features that require\nan expensive transformation and result in phase information loss. Previous work\nhas overcome these issues by using convolutional networks to learn long-range\ntemporal correlations across high-resolution waveforms. These models, however,\nare limited by memory-intensive dilated convolution and aliasing artifacts from\nupsampling. We introduce an end-to-end fully-recurrent hourglass-shaped neural\nnetwork architecture with residual connections for waveform-based\nsingle-channel speech enhancement. Our model can efficiently capture long-range\ntemporal dependencies by reducing the features resolution without information\nloss. Experimental results show that our model outperforms state-of-the-art\napproaches in six evaluation metrics.","url_abs":"http://arxiv.org/abs/1904.07294v1","url_pdf":"http://arxiv.org/pdf/1904.07294v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"rhr-net-a-residual-hourglass-recurrent-neural","repo_url":"https://github.com/CODEJIN/RHRNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"rhr-net-a-residual-hourglass-recurrent-neural","repo_url":"https://github.com/Jalal-Abdulbaqi/AudioSamples","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"speech-enhancement","task_name":"Speech Enhancement"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dilated-convolution","method_name":"Dilated Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}