smlum/scription

An editor for speech-to-text transcripts such as AWS Transcribe and Mozilla DeepSpeech

scription's Issues

Consolidate settings and media player into navbar

• move audio into navbar
• toggles into a settings dropdown
• file management into a file dropdown
• move annotations into a togglable sidebar
• move load, save and export into a dropdown

Update in-app documentation

Update instructions of how to use the app. Add instructions to the about button.

Reformat audio in the browser

HTML5 audio player has problems with large audio files. This causes delayed playback and syncing problems with the text.

The following was found to help with playback without compromising quality:

• Use a lossy file format (like mp3)
• Reduce bitrate and sample rates to 16k
• Reformat to a constant bitrate
• Convert to mono

It would be better to do all of this in the browser once a user uploads audio, e.g., using https://github.com/Kagami/ffmpeg.js

Make the interface responsive

The app doesn't need to be fully functional on mobile - but it should tell the user to use desktop at the very least. Or show an example of what the app can do.

Support videos as well as audio

Remove jQuery dependency

Add support for assembly.ai

Fix cookies for hosted site

Add support for rev.ai json

Allow audio and json load from url

Note: does loading large audio files from url cause the problems as uploading through file API?

Add notifications for common site errors

• warning for clearing transcript/ loading a new one
• audio size too large