JAutosub is an Windows app that can be used to generate automatic transcription / automatic subtitles for audio/video files through a friendly graphical user interface. For the hard work of speech recognition it is made use of the Google Speech Recognition API.
This app is basically a friendly Java GUI for a pre-packaged, ready to use, Windows version of Autosub 0.3.12. All the hard work of processing the audio is done by Autosub.
The app by default outputs the subtitles as .srt and the transcribed audio on the user interface as well as .txt files.
Internet connection is REQUIRED because it uses the Google Cloud Speech Server for the job, in the same way as the Youtube Automatic Subtitles.
IMPORTANT: The accuracy of the result can vary a lot, depending on many factors, mainly the quality/clarity of the audio. Ideally the audio input should not have background noise, sound effects or music. If there is a single speaker and he speaks in a clear and slow speed seems that the recognition is much more accurate. Sometimes, under ideal conditions it is possible to get a accuracy result close to 95%.
- Migrate to Java 10 Modular Project (smaller app package)
- JAutosub License change from MIT to GPL v3 (Autosub is licensed under MIT License)
- When closing the app force to end any remaining autosub subprocess
- Fixed bug: not able to find autosub_app.exe (wrong path)
v1.0 - 29/08/2017
- To get autosub running on Windows followed this instructions.
- The packaged version of Autosub contained was generated by pyInstaller.
- The final autosub_modified.py with all the modifications needed for packaging on the pyinstaller and the respective patch for the autosub 0.3.12
- Apache Netbeans 9 or newer -
- JDK 9 or newer
- Clone/Download git repository
- Extract and open the netbeans project
- Run it