[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-142654-105":59,"doc-detail-142654-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","a-comparison-of-human-and-computer-recognition-accuracy-for-childrens-speech","A Comparison of Human and Computer Recognition Accuracy for Children’s Speech","","Automatic speech recognition error rates tend to be worse for children than for adults, with word recognition errors increasing as age decreases. This work tests whether the same age and bandwidth effects seen in machine systems also appear in human listening results. Experiments measure human recognition performance using down-sampled British English children’s speech and compare it directly with machine recognition on identical data. Findings indicate similar dependencies on age and bandwidth, suggesting intrinsic properties of children’s speech drive reduced accuracy.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/a-comparison-of-human-and-computer-recognition-accuracy-for-childrens-speech/142654/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/a-comparison-of-human-and-computer-recognition-accuracy-for-childrens-speech/142654.png","ImageObject",300,407,{"name":92,"@type":93},"Aditya","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-18","2026-08-25",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",12,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What is the main goal of the study?","Question",{"text":113,"@type":114},"To compare human and machine recognition accuracy for children’s speech, using the same speech data, and to determine whether age and bandwidth effects occur for both.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"How is the children’s speech data prepared for the experiments?",{"text":118,"@type":114},"Speech from children aged 6 to 11 with an original 11 kHz bandwidth is down-sampled to 8, 6, 4, and 3 kHz, then recognized using an automatic system and by human listeners.",{"name":120,"@type":111,"acceptedAnswer":121},"What do the results suggest about recognition errors?",{"text":122,"@type":114},"Both human and automatic recognition show similar effects of age and bandwidth, implying the reduced accuracy is due to properties inherent to children’s speech rather than artifacts of the technology.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},142654,1787685528,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":19,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":130,"read_time":52,"preview_page_count":4},962085564549,"https://ap-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45","A Comparison of Human and Computer Recognition Accuracy for Children’s  \nSpeech  \nS D’Arcy and M.J Russell  \nDepartment of Electronic, Electrical and Computer Engineering, The University of Birmingham, Edgbaston,  \nBirmingham, B15 2TT United Kingdom  \n[SXD130@bham.ac.uk](SXD130@bham.ac.uk)  \n[m.j.russell@bham.ac.uk](m.j.russell@bham.ac.uk)  \n10.2 1437 /Interspeech .2005-697  \nAbstract  \nSeveral studies have shown that automatic speech recognition error rates are greater for children’s speech than for adult’s speech. Investigations have demonstrated that word recognition error rates increase as age decreases, and that recognition performance for children’s speech is more sensitive to bandwidth reduction, compared with adult speech. This paper presents the results of experiments to measure human recognition performance for children’s speech. The paper compares human and machine recognition performance on the same children’s speech data. It is shown that human recognition performance for children’s speech exhibits similar effects of age and bandwidth to those observed for automatic systems. The results suggest that effects of age and bandwidth on automatic speech recognition accuracy are due to properties of children’s speech rather than artifacts of the technology  \n1. Introduction  \nAutomatic recognition of children’s speech is a challenging task with potentially important applications in education and entertainment,[1] . Previous studies have shown that, compared with adults’ speech, word recognition error rates can be 100% worse for children’s speech [2], children’s speech is more variable, especially for very young children,[3], and performance degradation due to bandwidth reduction is more pronounced for children’s speech [4] . This paper asks whether two of these phenomena, namely the increase in word error rates for younger children and the effect of bandwidth reduction, are restricted to computer speech recognition, or whether they also occur in the results of human speech recognition experiments.  \nThis paper reports the results of experiments which compare the effects of age and bandwidth on computer and human recognition performance for the same children’s speech recognition tasks. The data used in the experiment is part of the British English children’s speech data component of the EU FP5 ‘PF STAR’ project’s children’s speech corpus. This corpus also includes recordings of native children’s speech in German and Swedish; non-native language children’s speech in English; and spontaneous and emotional children’s speech [5] . Speech data from children aged between 6 and 11, with an original bandwidth of 11 kHz, was down-sampled to 8, 6, 4 and 3 kHz bandwidths. This data was then recognised using an automatic speech recognition system trained on children’s speech data, and also by 40 human listeners.  \nThe results show that automatic speech recognition accuracy for this data is very poor. For example, for 11 year old  \nchildren and a full bandwidth of 11 kHz, the word error rate is 73.39% . Human recognition accuracy is also relatively poor, with an error rate averaged over all age groups of 3.6% at 11kHz bandwidth. This suggests that the data is inherently difficult to recognise and goes some way to explaining the poor performance of the automatic system. Although human performance is far superior to that of the automatic system, the results show that age and bandwidth induce similar effects in both experiments.  \nThe paper is organised as follows: The corpus used in the experiment is described in detail in section 2, the automatic speech recognition experiments are then described and results presented in section 3 . The experiment with human listeners are presented in section 4 . Finally, we present our conclusions.  \n2. The ‘PF STAR’ British English children’s speech corpus  \nThe corpus used for these experiments is a subset of the children’s speech corpus collected at the University of Birmingham as part of the EU FP5 ","cbCaiowdQBwy6Sxe","https://ap.wps.com/l/cbCaiowdQBwy6Sxe","pdf",107955,"English","# Abstract\n# 1. Introduction\n# 2. The ‘PF STAR’ British English children’s speech corpus\n## 2.1. Down-sampling speech the data\n## 2.2. Details of the automatic speech recognition system","[{\"question\":\"What is the main goal of the study?\",\"answer\":\"To compare human and machine recognition accuracy for children’s speech, using the same speech data, and to determine whether age and bandwidth effects occur for both.\"},{\"question\":\"How is the children’s speech data prepared for the experiments?\",\"answer\":\"Speech from children aged 6 to 11 with an original 11 kHz bandwidth is down-sampled to 8, 6, 4, and 3 kHz, then recognized using an automatic system and by human listeners.\"},{\"question\":\"What do the results suggest about recognition errors?\",\"answer\":\"Both human and automatic recognition show similar effects of age and bandwidth, implying the reduced accuracy is due to properties inherent to children’s speech rather than artifacts of the technology.\"}]","A Comparison of Human and Computer Recognition Accuracy for Children’s Speech | PDF"]