Archive

Archive for the ‘Japanese’ Category

Protected: Example 4: Speech recognition tutorial and voice training in Japanese on Windows 7

2014/09/07 Enter your password to view comments.

This content is password protected. To view it please enter your password below:

Common commands in Speech Recognition for all languages supported

(I cut a corner and left out the language variants ZH-TW and EN-UK, sorry, we do not teach those here):

Faculty Workshop Spring 2014: "Mira, mamá! Sin manos!". Practice speaking L2 with automatic intelligent feedback by operating LRC PCs through speech recognition instead of keyboard/mouse

  1. When: March 28, 2:15-3:15, April 4, 2:00-3:00
  2. Where: LRCRoomCoed434
  3. What: Language learning speaking practice assignments with automatic intelligent feedback using Windows Speech Recognition
    1. As part of the foreign language tools we installed with Windows 7 this past Fall, we got speech recognition on the LRC PCs for 6 languages (English, Chinese, French, German, Japanese, Spanish ) representing over 85% of our enrolment.
    2. Unlike the speech recognition that comes with learning content packages like Auralog or Rosetta Stone
      1. which had to be purchased, for individual languages, but stopped functioning on the server on a long time ago,
      2. was limited to built-in content ,
      3. was restricted by a separate account system,
    3. Windows Speech Recognition is
      1. free (with the operating system), runs on the local lab pcs, and should be a bit more robust,
      2. content agnostic and hence can integrate flexibly with your curriculum and contribute meaningfully to your students’ progression,
      3. can be integrated with the existing user accounts.
    4. We combine Windows speech recognition with the new LRC screencast software, MS-Office and Moodle to offer a simple self-access assignment type that
      1. is available on all 45 LRC PCs (= scales even to large enrolment languages and 1st-year classes that cannot use the 24-seat Sanako for face-to-face speaking proficiency training)
      2. and blends the “artificial intelligence” of speech recognition with human intelligence to provide students with immediate automated feedback during pedagogically sound speaking practice, with minimal grading overhead for the teacher (= grade secure assignments by looking at the very end of a student-submitted screencast).
    5. This workshop will show actual speech recognition usage and assignment samples
      1. so far in English, French, German;
      2. if you want to bring your own samples to this workshop – there might still be time- , or to an upcoming faculty showcase, I can help you during my biweekly LRC clinics (see LRC main schedule, or schedule your own).
    6. We will step you through – hands-on, including tips&tricks – a sample voice training and assignment completion: Better than my made-up assignments would be if you could bring one or more concrete tasks to be solved using speech recognition that we could prepare assigning to your students. Here are some parameters for that:
      1. Speech recognition can replace mouse and keyboard when operating the computer. Voice commands are simpler than sentences, so this could be a beginner task, as long as you have students study the (limited) command vocabulary (which I will make available during the workshop).
      2. Speech recognition can replace any writing task with dictation. Suggestions for proficiency levels:
        1. I have dictated a web page assigned for reading comprehension in a textbook used in 1200 or even a as a false beginner.
        2. However, a one-time training helping the computer recognize an individual’s voice is required and comes sentences that vary in complexity between languages
          1. English: very easy, Beginner level;
          2. German, French: let’s have a look together, I’d say 1202 level;
          3. Japanese: 3000 level, I was told;
          4. Please test with me during the workshop: Spanish, Chinese.
  4. Download the SlideDeck (too big too embed)

Example 5: Watch how you can dictate to Windows speech-recognition (e.g. in English) and correct results in MS-Word

  1. Important: Listen carefully: I am not a native speaker, but have a reasonably low amount of errors, because it enunciate, speak clearly and slowly, and separate the words.
  2. Consider it part of the exercise that you will have to re-read and re-type some your output – use track changes in MS-Word:
    1. Make it a game: How good can you get?
    2. If you get really good at it, make a screencast like this one and include it in your Mahara ePortfolio  as authentic evidence of your foreign language proficiency.
  3. Overall, it’s like how I refer to cycling: Beats walking. Anytime. Smiley

A first look at the Google Dictionary extension for Chrome

  1. We
    1. have not pre-installed in the LRC (for that the extension would need to be more manageable by the teacher during face-to-face classes, which include exams),
    2. but can (with some reservations) recommend the Google Dictionary extension (even though it is only available for Chrome). Here is why:
  2. Google dictionary extension provides an interface to Google define and translate
    1. that is convenient (as quickly accessed like glosses) for reading activities in many languages (Q: is the privileged word sense displayed here intelligently chosen?)
    2. while (for some languages more than for others) providing access to additional word senses, usage examples and historical background information
  3. Interface 1: Tooltip,
    1. for English with audio image
    2. for other languages without audio (even though audio pronunciation may be available in Google translate for that language): image
    3. convenient access (I have been loving the tooltip interface since Google toolbar days)
    4. limited, but useful  information,
      1. a word sense – not that this is still not contextually intelligent (Cannot blame them here!) and hence more than one word sense should be offered (here I must blame them: Boo!!): E.g.  here “arch” should at show more than the most common word sense: imageimage
      2. including pronunciation (not IPA, but audio)
    5. Interface 2 (“more”)
      1. For English, a click on “more” leads to the Google “define”search operator (the related etymology search operator has been reviewed here before): image
      2. Interface 3: unfold the search results by clicking on the down arrow at the bottom to access additional information:image =
        1. additional word sense entries
        2. historical:
          1. etymology
          2. frequency data
        3. translation/dictionary entry:
          1. for our learners of languages other than English, the translation appears right in the tool tip, see above;
          2. for our ESL learners, this seems a few too many steps for accessing this information, although a monolingual dictionary is useful in many instances also.
    6. For languages other than English, a click on more leads to Google translate, which (should get its own article, but for what it is worth) can be
      1. more limiting than “define”: While you are given multiple word senses for
        1. Spanish: image
        2. and to a lesser extent, for
          1. Arabic: image
          2. Hindi: image
      2. for many languages the results are much more limiting:
        1. Even if you look up German or French, you revert back to the (pedagogically terrible) single word-sense original “translation” interface ) image
        2. For East Asian languages, you get Roman alphabet transcriptions
          1. e.g. Chinese with Pinyin: image
          2. e.g. Japanese: image
  4. Still no per-user tracking? Here it would make sense for the user.

Quizlet.com for Vocabulary learning practice

Neallt 2014 is featuring a presentation on “Using Quizlet.com to generate and share vocabulary activities” (William Price, University of Pittsburgh). I cannot attend, but the program inspired me to hold my own sneak preview:

Quizlet.com is yet another site that provides a variety of flashcard and quiz activities for a given wordlist. A nice example is the “Speller”activity – which proves a text-to-speech generated aural cue for dictation (not included in this video):quizlet-speller-german

Or step-by-step:

imageUnfortunately, the AI seems limited to only 1-1 L1-L2 relationships (which precludes how vocabulary seems to be learnt best: in phrasal contexts):

image

Feedback on “wrong” user input is color red, aural and visual presentation of the correct form: image

Then the application re-prompts for user input and allows user correction:

image

This is a “Test activity. Foreign language character input seems easy (but does beg the question since the inputs appear only when they are needed…)

imageimage

Mix-and-match is called Scatter:

image

Here is the activity overview: image

 

Wait, there is more: image

Quizlet supports many dozen languages, including non-western, including ancient, not differentiating between modern and ancient Greek, but the browsing capabilities – admittedly a hard task – are somewhat flat (search and language)):

image

And boasts 20 million sets (as of today – many consisting of 2 terms or few more). As so often, usefulness for class instruction hinges on the availability of textbook-aligned vocabulary lists. However, if you have them with your textbooks, Quizlet makes it automatic to generate uploaded materials into exercises.image

However, as said, you may not like how much you have to dumb it down.

Overview over speech recognition assignment possibilities in the LRC–from beginning to end

2014/01/21 1 comment
  1. UPDATE: now with more videos to watch.
  2. First switch languages, once you are in the right language:
  3. The speech recognition loads at startup with this window:
    image
  4. or can be started from the desktop shortcut:
    image
  5. To activate = make it start/stop listening, click on the microphone image
  6. create and save individualized voice training data (only the first time you use speech recognition in the LRC),
    1. Watch how to train (for English, short 5 minutes);
    2. Watch how to back up/restore;
    3. beginner voice training example (long, before/after effect in speech recognition),
  7. homework assignment exercise design,
    1. Watch how (for English, with correction in MS-Word)
    2. dictation example.