Difference between revisions of "Proposal Languages Ecosystem"

From Audacity Wiki
Jump to: navigation, search
(add link to Moodle)
(Use Cases link.)
 
(3 intermediate revisions by one other user not shown)
Line 11: Line 11:
 
** Audacity interfaces to this in recording/comprehension practice.   
 
** Audacity interfaces to this in recording/comprehension practice.   
 
** Audacity also interfaces to this with Vowel Quadrilateral.
 
** Audacity also interfaces to this with Vowel Quadrilateral.
 +
*** Notice how much more informative this is than 'good/bad' tracking found in for example "101 Languages of the world" software.
 
** A separate app for Export/Import to learning-tracking systems like [http://moodle.org/ Moodle] and Khan Academy - these tend to track only at the level of a lesson, so some data would not transfer.
 
** A separate app for Export/Import to learning-tracking systems like [http://moodle.org/ Moodle] and Khan Academy - these tend to track only at the level of a lesson, so some data would not transfer.
 
* Voice recognition workbench (research level).  This is for display and work with internal data used in voice recognition, the intermediate forms in which audio data is represented when working out what was said.  In this proposal it's mostly about data formats, being able to import and display such data on specialised tracks.  The key to this format is ease of extensibility.  
 
* Voice recognition workbench (research level).  This is for display and work with internal data used in voice recognition, the intermediate forms in which audio data is represented when working out what was said.  In this proposal it's mostly about data formats, being able to import and display such data on specialised tracks.  The key to this format is ease of extensibility.  
Line 24: Line 25:
 
There are some changes which can perhaps be done quickly and easily, long before a full ecosystem is developed.
 
There are some changes which can perhaps be done quickly and easily, long before a full ecosystem is developed.
  
* Plug-in for Audio practice.  Has the notion of a 'current label'.  Button to play that label.  Button to move to next label.  Button to over-record a clip holding one's best take for that label.  Button to pack up all the clips into an mp3 to e-mail to teacher.
+
* Plugin for Audio practice.  Has the notion of a 'current label'.  Button to play that label.  Button to move to next label.  Button to over-record a clip holding one's best take for that label.  Button to pack up all the clips into an mp3 to email to teacher.
* Plug-in for Audio listening.  Has dial to increase pauses at ends of sentences.  Has button to back up 3 seconds (configurable) or replay from start of sentence. Has button to play an alternative track 'explanation' for a given sentence.
+
* Plugin for Audio listening.  Has dial to increase pauses at ends of sentences.  Has button to back up 3 seconds (configurable) or replay from start of sentence. Has button to play an alternative track 'explanation' for a given sentence.
* Plug-in for Dictation Exercise.  Plays a sentence.  Waits for you to type.  Continues when you type the '. and <return>' at the end of the sentence.
+
* Plugin for Dictation Exercise.  Plays a sentence.  Waits for you to type.  Continues when you type the '. and <return>' at the end of the sentence.
  
 
== Developer Backing ==
 
== Developer Backing ==
Line 35: Line 36:
 
== Use Cases ==
 
== Use Cases ==
  
 +
* See [[Use_Cases#languages|Use Cases for Language Learning]].
  
 
== Random Thoughts ==
 
== Random Thoughts ==
  
 +
* See also Praat for (open source) vocal model.
 +
* See also Google WordNet for methods for training deep networks for vocalisation (bisection exponentially reduces Neural-net training times)
 
   
 
   
 
==Interface==
 
==Interface==

Latest revision as of 21:10, 10 June 2020

Proposal pages help us get from feature requests into actual plans. This page is a proposal for an ecosystem of programs built around Audacity, for language learners. The key is shared formats that Audacity and related programs understand.
Proposal pages are used on an ongoing basis by the Audacity development team and are open to edits from visitors to the wiki. They are a good way to get community feedback on a proposal.


  • Note: Proposals for Google Summer of Code projects are significantly different in structure, are submitted via Google's web app and may or may not have a corresponding proposal page.


Proposed Features

This is for language learners.

  • Add Proposal Structured Audio to Audacity. Audacity becomes the editing tool for preparing structured audio lessons.
    • Includes import of .srt format.
  • Progress tracking format. This is tracking vocab e.g. extent to which a word has been learned, both from and to the foreign language (written) and comprehension/production (audio). Has hierarchical features, so can track at level of phonemes/words/phrases.
    • Audacity interfaces to this in recording/comprehension practice.
    • Audacity also interfaces to this with Vowel Quadrilateral.
      • Notice how much more informative this is than 'good/bad' tracking found in for example "101 Languages of the world" software.
    • A separate app for Export/Import to learning-tracking systems like Moodle and Khan Academy - these tend to track only at the level of a lesson, so some data would not transfer.
  • Voice recognition workbench (research level). This is for display and work with internal data used in voice recognition, the intermediate forms in which audio data is represented when working out what was said. In this proposal it's mostly about data formats, being able to import and display such data on specialised tracks. The key to this format is ease of extensibility.
  • Improved customisability and support for plug-ins. This would allow Audacity to be customised for use in the classroom, without it being a different executable.

Most of these features are 'format' features.

The ecosystem would include 'flash card' apps, extraction of vocab from wikipedia, Moodle extensions and so on. These are not necessarily developed by Audacity developers. We would 'own' the data formats.

The formats, as they evolve, will drive new features in Audacity. For example, the ability to work with and navigate in an audio corpus.

Low Hanging Fruit

There are some changes which can perhaps be done quickly and easily, long before a full ecosystem is developed.

  • Plugin for Audio practice. Has the notion of a 'current label'. Button to play that label. Button to move to next label. Button to over-record a clip holding one's best take for that label. Button to pack up all the clips into an mp3 to email to teacher.
  • Plugin for Audio listening. Has dial to increase pauses at ends of sentences. Has button to back up 3 seconds (configurable) or replay from start of sentence. Has button to play an alternative track 'explanation' for a given sentence.
  • Plugin for Dictation Exercise. Plays a sentence. Waits for you to type. Continues when you type the '. and <return>' at the end of the sentence.

Developer Backing

Use Cases

Random Thoughts

  • See also Praat for (open source) vocal model.
  • See also Google WordNet for methods for training deep networks for vocalisation (bisection exponentially reduces Neural-net training times)

Interface