weixin_39883260
weixin_39883260
2020-12-31 06:21

Idea - Audio Library for Word Pronunciations

This is probably an idea that would be more useful in the sciences, but could have interesting applications elsewhere too.

Text-to-speech does a very poor job with the pronunciation of complex scientific words. In our case, properly pronouncing drug/medicine names is important for our learners to communicate medications to patients, but it's always difficult to decipher how a word should be pronounced from how it's spelt.

http://howjsay.com/ has an online library of audio clips where the word is read out properly, which is where this idea came from.

Therefore my plan was to record short audio clips reading specific words out, upload them to Media - then have an option within the course (possibly in the accessibility settings) to switch on a 'medicine name audio library' to get the following effect when rolling your mouse over (or clicking) the word in your course: http://codepen.io/rickhumphries/pen/MKLYXP (try typing metronidazole into https://www.ivona.com/ and you'll see how different they can be!)

My thinking is that the instructor could create their own libraries of words - and link each word to an audio shortcode from Media. The student would have the option in the accessibility settings to switch on/off a specific library. If a library is switched on, the tool scans the book for all instances of this word and processes it in a way so that of you click the word, the audio file plays.

Just wondering if this is an interesting concept for anyone else - and to get implementation ideas too...

该提问来源于开源项目:elmsln/elmsln

  • 点赞
  • 写回答
  • 关注问题
  • 收藏
  • 复制链接分享
  • 邀请回答

14条回答

  • weixin_39637363 weixin_39637363 4月前

    Also, this could be a good opportunity to add HTML5 audio recording to ELMS. In the pronunciations content type, it would be used by admins to optionally record their audio straight from their browser. In other instances, it could be used as an evaluation tool, such as within a linguistics course.

    点赞 评论 复制链接分享
  • weixin_39954674 weixin_39954674 4月前

    Which we have a widget for. Search drupal.psu.edu for html5 recording widget!

    点赞 评论 复制链接分享
  • weixin_39883260 weixin_39883260 4月前

    Yeh great plan ! And I think the HTML5 audio recording straight from the browser would be fantastic - such a great way to reduce the barriers for instructors to add these in :) :) Let me what your group thinks

    点赞 评论 复制链接分享
  • weixin_39954674 weixin_39954674 4月前

    almost 2 years ago -- https://www.youtube.com/watch?v=kNUSmWpeWsk

    点赞 评论 复制链接分享
  • weixin_39954674 weixin_39954674 4月前

    love the idea, I was texting that he's thinking too much like me if he beat me to the punch on where we could take this idea. Our group may have less direct usage for this but I know our art history and world music courses could benefit from this concept. -wehr any applicability in the business realm?

    点赞 评论 复制链接分享
  • weixin_39830012 weixin_39830012 4月前

    There's definitely an opportunity here for audio recording in conjunction with slides or other media in a presentation style assignment. I'm thinking like VoiceThread except with a nicer interface (and free/open source of course). The application would be HUGE for folks learning second languages, or to help faculty create databases of commonly-used vocabulary for ESL students and offer it as a resource for those learners. Also, giving essay/written test responses orally is a common accommodation for students, I could see a tool like this being used in that capacity as well.

    点赞 评论 复制链接分享
  • weixin_39954674 weixin_39954674 4月前

    thoughts from a K-12 perspective? The multi-lingual / ESL support immediately comes to mind, good idea on that -wehr .

    点赞 评论 复制链接分享
  • weixin_39796116 weixin_39796116 4月前

    Adding to 's implementation, it might be useful to also have an entity reference which could point to related words/pronunciations. I was thinking of brand vs. generic names of medications or stars & the constellations in which they reside.

    点赞 评论 复制链接分享
  • weixin_39932611 weixin_39932611 4月前

    When thinking of a multi-cultural environment where English maybe for many a second or third language, this is a reasonable idea.

    However, thinking about the function / feature from a layman's perspective I am reminded of the famous Winston Churchill quote (in not so many words) that:

    " Never were two countries so divided by a common language"

    Which language are you proposing to use, English or American!?

    Then also consider that an English billion is not the same as an American billion ;)

    Kr, Chris

    点赞 评论 复制链接分享
  • weixin_39883260 weixin_39883260 4月前

    Yes, pronunciations can be different - but by creating your own audio then you can at least localise it to an extent.

    Alternatively - you can record a slightly longer audio and try to cover multiple common pronunciations, e.g. http://howjsay.com/mp3/capillary.mp3 if the course has a more global scope.

    点赞 评论 复制链接分享
  • weixin_39954674 weixin_39954674 4月前

    we also wouldn't be packaging audio, just the ability to provide it. multi-track might be good too for a simple multiple pronunciation option; that or for things like saying the word paired with saying the word in a sentence as two separate files.

    点赞 评论 复制链接分享
  • weixin_39932611 weixin_39932611 4月前

    Multiples would be needed for some words and syllables that have differing pronunciations for the same spelling depending on the context, e.g. noun vs. verb variations. I agree the sound track quality does not need to be high fidelity. Another though that comes to mind when thinking of the larger purpose embodied in this idea is that done by some modern web services in which key words in a text are automatically identified and linked to relevant information, such as a Wikipedia page or thesaurus. If this can be done in this language context or perhaps even combined in one system of dynamic content linking might be useful. I suspect a manual override might be needed though as otherwise this may cause a hyper-link overload of less relevant info and distract from links to info more directly relevant to the lesson.

    / Chris

    点赞 评论 复制链接分享
  • weixin_39883260 weixin_39883260 4月前

    http://soundcite.knightlab.com

    点赞 评论 复制链接分享
  • weixin_39954674 weixin_39954674 4月前

    woah. that's awesome!

    点赞 评论 复制链接分享

相关推荐