[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-124672-en":3,"doc-seo-124672-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},124672,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",8,"Research & Report","HRTF Generation for Data Demanding Machine Learning Algorithms - A Thesis","This thesis investigates the application of machine learning techniques to binaural audio research, addressing a central limitation in current work: the amount of available Head Related Transfer Function (HRTF) data needed to train modern neural network models. It presents a method to simulate effectively unlimited realistic HRTFs by generating heads from three-dimensional morphable models (3DMMs). The approach enables the HUMMNGBIRD database and supports PCA and VAE reconstruction models, offering insights for near-term research directions using synthetic HRTF datasets.","HRTF Generation for Data Demanding Machine Learning Algorithms  \nBenjamin Tsui  \nPhD  \nUniversity of York Electronic Engineering  \nJanuary 2023  \nAbstract  \nThis thesis investigates the application of Machine Learning (ML) techniques to binaural audio research. Whilst there is plenty of work done in this domain currently, much of it is limited by the amount of available Head Related Transfer Function (HRTF) data required to train modern neural network-based ML models, resulting in researchers using a less data-driven approach or finding some workaround with the limited data. This thesis focuses on the generation of enough data to unleash the power of a wide variety of modern ML algorithms. A novel method is presented that can simulate unlimited realistic HRTFs using heads generated from Three-dimensional Morphable Models (3DMMs) . The result has led to the creation of the HUman Morphable Modelbased Numerically Generated Binaural Impulse Response Database (HUMMNGBIRD) database, created with the first 5000 HRTF sets generated by this method. Principle Component Analysis (PCA) and Variational Auto-Encoder (VAE) reconstruction models were created to investigate the potential of such a large amount of data. The results provide valuable insights into the research directions that could make good use of these types of artificially generated databases in the near future.  \nContents  \nAbstract 2  \nContents 3  \nList of Figures 7  \nList of Tables 11  \nAcknowledgements 13  \nDeclaration 15  \nThesis 17  \n1 Introduction 19  \n1.1 Motivation .................................. 19  \n1.2 Objectives ................................... 20  \n1.3 Thesis Structure ............................... 23  \n2 Literature Review 25  \n2.1 Head Related Transfer Functions (HRTFs) ................. 25  \n2.1.1 Localisation Cues ........................... 26  \n2.1.2 Use of HRTFs ............................ 29  \n2.1.3 Obtaining HRTFs .......................... 30  \n2.1.4 HRTF Data .............................. 33  \n2.1.5 HRTF Interpolation ......................... 34  \n2.1.6 Personalised HRTFs ......................... 39  \n2.2 Machine Learning ............................... 43  \n2.2.1 Supervised, unsupervised and reinforcement learning ....... 43  \n2.2.2 Symbolic vs Connectionist ...................... 44  \n2.2.3 Traditional machine learning models ................ 45  \n2.2.4 Modern machine learning models .................. 51  \n2.2.5 Machine learning based HRTF research .............. 59  \n2.3 Discussion ................................... 63  \n2.3.1 HRTF Data format .......................... 63  \n2.3.2 The future of HRTF Personalisation ................ 64  \n2.3.3 HRTF Personalisation with Transformers ............. 64  \n2.4 Conclusion .................................. 65  \n3 A HRTF Consolidation Tool For High Variance Machine Learning Algorithms 67  \n3.1 Motivation and Challenges .......................... 67  \n3.2 Development of an HRTF Database Consolidation Tool ......... 68  \n3.2.1 Pipeline and File Structure ..................... 68  \n3.2.2 Details ................................. 70  \n3.2.3 Using the Toolbox .......................... 71  \n3.2.4 Extra Tool: Plot Measurement Angles ............... 72  \n3.3 Summary ................................... 72  \n4 Low-order Spherical Harmonic Interpolated HRTF Restoration using a Neural Network Approach 75  \n4.1 Motivation .................................. 75  \n4.2 Spherical Harmonic HRTF interpolation .................. 75  \n4.3 Machine learning HRTF Restoration .................... 76  \n4.3.1 Data pre-processing ......................... 77  \n4.3.2 Baseline Model ............................ 80  \n4.3.3 Model Enhancement ......................... 84  \n4.4 Evaluation ................................... 93  \n4.4.1 Perceptual Spectral Difference ................... 93  \n4.4.2 Localisation performance ...................... 95  \n4.5 Discussion ................................... 98  \n4.6 Conclusion ...........","cbCaigCxBzMRXkQm","https://ap.wps.com/l/cbCaigCxBzMRXkQm","pdf",23037464,1,185,"English","en",105,"# 1 Introduction\n## 1.1 Motivation\n## 1.2 Objectives\n## 1.3 Thesis Structure\n# 2 Literature Review\n## 2.1 Head Related Transfer Functions (HRTFs)\n## 2.2 Machine Learning\n## 2.3 Discussion\n# 3 A HRTF Consolidation Tool For High Variance Machine Learning Algorithms\n## 3.1 Motivation and Challenges\n## 3.2 Development of an HRTF Database Consolidation Tool\n## 3.3 Summary\n# 4 Low-order Spherical Harmonic Interpolated HRTF Restoration using a Neural Network Approach\n## 4.1 Motivation\n## 4.2 Spherical Harmonic HRTF interpolation\n## 4.3 Machine learning HRTF Restoration\n## 4.4 Evaluation\n## 4.5 Discussion\n## 4.6 Conclusion\n# 5 Generating HRTFs with a 3D morphable model of human heads\n## 5.1 Motivation and Challenges\n## 5.2 Method\n## 5.3 Evaluation\n## 5.4 Conclusion\n# 6 Preliminary Investigation into the Potential of Using HUMMNGBIRD HRTF Datasets in Machine Learning\n## 6.1 Motivation\n## 6.2 Setup\n## 6.3 Experiment results\n## 6.4 Discussions\n## 6.5 Conclusion\n# 7 Conclusion\n## 7.1 Restatement of Research Hypotheses\n## 7.2 Future Work","[{\"question\":\"Why is additional HRTF data important for modern machine learning models?\",\"answer\":\"Modern neural network-based ML models require large amounts of Head Related Transfer Function (HRTF) data. The limited availability of measured HRTFs restricts fully data-driven learning and motivates alternative strategies.\"},{\"question\":\"What new method does the thesis propose for generating HRTFs?\",\"answer\":\"The thesis introduces a method that simulates effectively unlimited realistic HRTFs by generating heads from three-dimensional morphable models (3DMMs).\"},{\"question\":\"How is the generated data evaluated and used in further modeling?\",\"answer\":\"The work creates PCA and variational auto-encoder (VAE) reconstruction models to analyze the generated database. Evaluation results highlight how these approaches can exploit artificially generated HRTF datasets for machine learning research directions.\"}]","HRTF Generation for Data Demanding Machine Learning Algorithms - A Thesis | PDF",1785893834,466,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"hrtf-generation-for-data-demanding-machine-learning-algorithms-a-thesis","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/hrtf-generation-for-data-demanding-machine-learning-algorithms-a-thesis/124672/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why is additional HRTF data important for modern machine learning models?","Question",{"text":75,"@type":76},"Modern neural network-based ML models require large amounts of Head Related Transfer Function (HRTF) data. The limited availability of measured HRTFs restricts fully data-driven learning and motivates alternative strategies.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What new method does the thesis propose for generating HRTFs?",{"text":80,"@type":76},"The thesis introduces a method that simulates effectively unlimited realistic HRTFs by generating heads from three-dimensional morphable models (3DMMs).",{"name":82,"@type":73,"acceptedAnswer":83},"How is the generated data evaluated and used in further modeling?",{"text":84,"@type":76},"The work creates PCA and variational auto-encoder (VAE) reconstruction models to analyze the generated database. Evaluation results highlight how these approaches can exploit artificially generated HRTF datasets for machine learning research directions.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]