Skip to content

0.20.1-beta
Compare
Choose a tag to compare
@arianneorpilla arianneorpilla released this 11 Jun 06:45
· 813 commits to main since this release
0.20.1-beta
e858c03
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

This release implements a Romaji filter that can be enabled in the audio/subtitles menu, which strips and hides away subtitle lines that have Latin characters or English punctuation characters. This feature is useful for practicing reading comprehension for videos that include Romaji subtitles alongside Japanese subtitles or preventing accidental viewing of English subtitles.

If a line has exclusively Latin characters or English punctuation characters, the line will be hidden overall. Otherwise, if the Latin characters co-exist in a line beside Japanese characters, the characters will be replaced by maru (○) to indicate use of Latin characters.

Developer Note: The filter has been tested on a variety of Japanese and English subtitles. Should any significant special character artifacts exist in your subtitles when using the filter, please feel free to report it to the issue tracker so that I may improve the filter. I will need to know in particular which Unicode characters are involved.

Template Updates: The default templates have been updated with 0.20.1. Instructions on how to update the template can be reviewed here.

System Requirements: This application has been tested by myself and other users to work on a variety of real-world devices and emulators running Android 7 and above. If the installation fails, try clearing enough storage space.

0.20-beta

  • (New Feature) Filter Latin Characters - Users are now able to select a new option in the audio/subtitles player menu which will censor out most characters used in English to improve reading comprehension experience for subtitles that contain non-target language text
  • Typography used in the app has changed, fixing a critical issue where Han unification characters were shown in the app rather than using Japanese characters
  • The creator will now not automatically search a photo when opened via a share context (as a quality of life experiment for users who intend to use the app for novel and manga mining), if this significantly affects your workflow, please make an issue
  • If the image search term field is empty when the search button is pressed, the word will now be used as the search term as a quality of life improvement
  • Added a warning to inform a user when they watch a video from scoped storage rather than direct play
  • Fixed a significant player performance issue where the tap subtitle widget would constantly rebuild during playback
  • Fixed an issue where video history would not update once the maximum of 20 videos has been reached
  • (Hotfix) - Added tilde (~) to allowed characters to check only for line striping

0.20.1-beta

  • The default templates have been slightly changed. Instructions on how to update the template can be reviewed here.
  • A new reader template is used for shared text on the edge case that users desire to have a picture in the front for manga cards and in the back for novel sentence cards
  • The export button will now indicate if the standard creator or reader template is used
  • Reverted a change where a picture was not automatically searched for when sharing text into the app, and selecting a word will now also search a picture once again
  • Doubled the thickness of the pitch accent diagrams in the default templates

Notes on making a supposedly fairly functional filter

Developer Note: Every little effort to focus oneself to consuming more target language content rather than native language content is a significant improvement, so I desired to work on this feature. It was an interesting thing to implement, and it reminded me of the Scunthrope problem.

At first, I got rid of all the English relevant ASCII characters, which left me with a problem for non-English subtitles. I was fortunate to have some Spanish and French subtitles to test at hand. I then had to take an extended alphabet into account to include diacritics and accents. I then realized that punctuation was hit or miss - and sometimes that Japanese subtitles would sometimes use Western punctuation marks. I had to make it so that the punctuation marks only get censored if the rest of the subtitles were mostly those filtered Latin characters.

Then, I had to decide when to strip away the line completely, and when to just keep a partially censored line around. What I've decided is that if Latin characters existed alongside Japanese, maru or ○ would be shown instead to inform the user that there is censored text that may or may not be of interest to them. They may then turn the filter on or off to their own content.

It's a very complicated issue, and what results is a filter that I am mostly happy with and probably tailored to perfectly censor what videos I have tested with, but I am sure that some edge cases exist for some videos that simply do not conform to the standards I have observed in subtitles.

Which file should I download?

The files are split between the different device architectures to save user download bandwidth and storage space. Depending on which device you are using, download the appropriate APK by expanding the "Assets" drop-down menu below.

Test each APK available for each architecture. If you're not sure, the arm64 download should be suitable for most mobile devices.

Assets 5
Loading